I

NASA TECHNICAL T R A N S L A T I O N
N h

N A S A T I F-292
c.(
AN r p 7 w : RFTL

N
I

eE
* 4 X * n
3 3

-0

= 4
i

LL

+ c
4

m
4

I 2

FUNDAMENTALS OF THE ANALYTICAL MECHANICS OF SHELLS

by N. A. Kilcheuskiy
Izdutelstuo Akademii Nauk Ukraimkoy SSR, Kiev, 1963.

NATIONAL AERONAUTICS AND SPACE ADMINISTRATION

WASH I N GTO N,

TECH LIBRARY KAFB,

NM

NASA TT F-292

00b8832

FUNDAMENTALS O F THE ANALYTICAL MECHANICS O F SHELLS By N. A. Kil'chevskiy

T r a n s l a t i o n of "Osnovy a n a l i t i c h e s k o y mekhaniki obolochek." Izdatel'stvo A k a d e m i i Nauk Ukrainskoy SSR, Kiev, 1963.

NATIONAL AERONAUTICS AND SPACE ADMINISTRATION

For sale by the Clearinghouse for Federal Scientific and Technical Information Springfield, Virginia 22151 Price $7.00

. .

. -

FlTNDAMENTALS OF THl3 ANALrrICAL MECHANICS OF SHELLS

By N. A. Kil'chevskiy

The book discusses a n a l y t i c methods of constructing e l a s t o s t a t i c and elastodynamic systems of d i f f e r e n t i a l and i n t e g r a l equations of s h e l l theory without r e q u i r i n g t h e use of a d d i t i o n a l assumptions on t h e deformation of t h e s h e l l s , and a l s o methods of solving t h e systems of i n t e g r a l equations by reducing them t o systems of ordinary different i a l and a l g e b r a i c equations. N o use i s made of t h e well-known assumptions t h a t cons t i t u t e t h e foundation of c l a s s i c a l s h e l l theory, but t h e author starts o u t from t h e general p r i n c i p l e s of e l a s t i c i t y theory and d e r i v e s more exact d i f f e r e n t i a l equations of t h e s h e l l theory, of higher order than those of t h e c l a s s i c a l theory
The book i s intended f o r s c i e n t i s t s , post-graduate students and t e c h n i c a l u n i v e r s i t y students s p e c i a l i z i n g i n t h e theory of e l a s t i c s h e l l s .

Numbers i n margin refer t o pagination i n f o r e i g n text.

iii

TABLE O F C O N T E N T S
Page PREFACE

...................................................... INTRODUCTION .................................................

........... Section 1 . Tensor General C h a r a c t e r i s t i c s of t h e Application of Analysis t o t h e Theory of S h e l l s ....... S e c t i o n 2 . Systems of Curvilinear Coordinates . Metrics of Space . The Symbol f o r Summation ........... S e c t i o n 3 . Metrics i n S h e l l s ............................. S e c t i o n 4 . S h e l l s of Revolution . S p e c i a l Cases of S h e l l s of Revolution . A r b i t r a r y Cylindrical Shells ........................................ 1 . The C i r c u l a r Cylindrical S h e l l ............ 2 . The Conical Cylindrical S h e l l ............. 3 . The S h e l l with t h e Base A r e a i n t h e Form of a Hyperboloid o f Revolution ............ S e c t i o n 5 . S c a l a r s . Vectors and t h e i r Contravariant and Covariant Components . The Mutual Coordinate Base .......................................... Section 6 . Tensors of Various Ranks and S t r u c t u r e s . The Metric Tensor of t h e S h e l l .................... Section 7 . Operations of Tensor Algebra .................. 1 . Addition .................................. . M u l t i p l i c a t i o n ............................... ............................ 3 . Contraction 4. 'PRaising'r and rFLowerhg*i of Indices ....... 5 . Permutation of Indices . Symnetrization and Alternation ........................... Section 8 . Various Applications of Tensor Algebra ........ 1 . Second Analytic D e f i n i t i o n of a Tensor .... 2 . Antisymmetric Tensor of Second Rank as a Vector i n Three-Dimensional Space ....... 3 . Vector Product of Two Vectors i n a n ............... A r b i t r a r y Coordinate System 4 . Pseudoscalars and Pseudovectors ........... S e c t i o n 9 . Absolute D i f f e r e n t i a l of a Tensor . Tensor F i e l d and t h e Absolute Derivative ............. 1 . Absolute D i f f e r e n t i a l of a Vector ......... 2 . Absolute D i f f e r e n t i a l of a Tensor of .............. A r b i t r a r y Rank and S t r u c t u r e 3 . Tensor F i e l d . The Absolute (Covariant) Derivative of a Tensor of A r b i t r a r y Rank and S t r u c t u r e .............................
2
V

x i
1

CHAPTER I

Elements of Tensor A n a l y s i s and t h e i r Application t o t h e D i f f e r e n t i a l Geometry of S h e l l s

6 6
6

7
10
11 12

13

13
11
20 22

22 22 23

2425 26 26 26 28 28 29 29

33

34

Page P a r a l l e l Displacement of Tensors i n t h e Sense of Levi-Civita. The Tensor of Curvature' 1 . P a r a l l e l Displacement 2. Tensor of Curvature (Riemann-Christof f e l Tensor) 3. Change of t h e Sequence of Operations i n Successive Absolute D i f f e r e n t i a t i o n 4. Geometric Construction o f t h e Covariant Derivative Section 1 1 . Operator of Parallel Displacement of Tensor Quantities on t h e Base Area of a S h e l l Section 12. Expansion of Tensor Functions i n Generalized Taylor S e r i e s 1. Analytical D e f i n i t i o n of t h e Radius Vector of a Point of Space i n Curvil i n e a r Coordinates 2. Expansion of Tensor Functions i n Generalized Taylor S e r i e s Section 10.

CHAPTER 1 1 .

.............................. ...... ................... ...................... ............... P r i n c i p a l Relations of t h e Nonlinear Theory ............ of E l a s t i c i t y i n t h e I n v a r i a n t Form .......................... ........................ ............ ....................... ............. ..... .. ...

......................~*..~.......* 35 ................... 35 ................................. 35 ..... 38

39
39

44 44
4 4
46
4.6
47 4.7

Section 1 . E u l e r and Iagrange Variables. Displacement Vector, Velocity Vector, and Acceleration Vector of a n Element of a Continuous Medium Section 2. Tensor of Small Deformations and Tensor of F i n i t e Deformations 1 . Tensor of Small Deformations and Vector of Small Rotation of an Element of a Continuous Medium 2. Tensor of F i n i t e Deformations 3. Concluding Remarks Section 3. Conditions of Compatibility Section 4. S t r e s s Tensor. Generalized Hookfs Law 1 . Linear Generalization of Hooke's Law. Physical and Geometric Nonlinearity of t h e Equations of t h e Theory of E l a s t i c i t y 2. The Nonlinear Hookefs Law 3 . Concluding Remarks Section 5. Equations of Motion of a n Element of a Continuous Medium. The Linear Lame' Equations 1. Equations of Motion of a n Element of a Continuous 3iedium i n an Arbitrary System of Lagrange Coordinates 2 . Linear Lam6 Equations a Section 6. Relationships between Covariant Derivatives i n Deformed and Undeformed Media 1 . Fundamental Determinant

...

49

50 50

51
51-

.. ................ ....................... .. ...........

53 54.

55 55 56
57 57

... ................ ............. ...................

vi

Page Covariant and Contravariant Components of t h e Metric Tensor of a Deformed Medium 3. C h r i s t o f f e l Symbol i n a Deformed Medium 4. Covariant Derivative i n a Deformed Medium 5. Conclusion Nonlinear Lame Equations Initial and Boundary Nonlinear Conditions. Conditions of Contact of Layers 1 . I n i t i a l Conditions 2. Nonlinear Boundary Conditions 3 . Conditions of Contact on Surfaces of Separation of Media with b t t e r of Differe n t Mechanical C h a r a c t e r i s t i c s 4 . General Characterization of t h e Formulat i o n of Monlinear Woblerns of t h e Theory of E l a s t i c i t y I n t e r n a l and External Nonlinear Problems Extension of t h e Kinematic Relations of t h e Kirchhoff-Clebsch Thin-Rod Theory t o S h e l l Theory P o t e n t i a l Energy of Deformation and Kinetic Energy of t h e E l a s t i c Body Work and Reciprocity Theorem i n Nonlinear E l a s t i c i t y Theory E l a s t i c Medium with I n i t i a l S t r e s s e s

2 .

Section 7. S e c t i o n 8.

................................... .................................. ..................................... ...............................

..................... .............. ....................... ............
........... ............................ .....

57 59
61 61
62

63 63 64.

67 67

S e c t i o n 9. S e c t i o n 10. Section 1 1 . S e c t i o n 12. S e c t i o n 13.

68

........................................ .................... ............................. ..........

70

74
75 77

CHAPTER 1 1 1 . Reduction of t h e Three-Dimensional Problems of t h e Mechanics of E l a s t i c Bodies t o t h e TwoDimensional Problems of t h e Theory of S h e l l s

S e c t i o n 1. General Characterization of t h e Problem S e c t i o n 2. Remarks on t h e Methods of Reduction given by Poisson, Cauchy, Kirchhoff, and Love S e c t i o n 3 . Preliminary C l a s s i f i c a t i o n of S h e l l s i n Connection with t h e Kirchhoff-Love Hypotheses. Linear and Nonlinear Problems S e c t i o n 4. Application of Tensor S e r i e s . Reduction of t h e Three-Dimensional Problem t o t h e Determinat i o n of a n I n f i n i t e Sequence of Functions of a Point of t h e Base Area of t h e S h e l l S e c t i o n 5. Reduction of t h e Three-Dimensional Problem t o t h e Determination of S i x Functions of a Point of t h e Base A r e a of t h e S h e l l S e c t i o n 6 . Application of t h e Symbolic Method S e c t i o n 7. Ebqressions f o r t h e Vormal Part of t h e S t r e s s Tensor. The Equations Defining t h e Fundamental

....... ....... .................

...

80
80

81

83

........... ........... ............

86 89
90

vi i

Section 8 . Section Section Section

Section Section Section Section Section Section

Section Section Section Section Section Section Section

Section Section

Functions Further Development of t h e C l a s s i f i c a t i o n of S h e l l s with Respect t o Dynamic Problems 9 . Method of Successive Approximations 10. Expansion of t h e Deformation Tensor i n t o a Tangential P a r t and a Normal P a r t 1 1 . Two Methods of S e t t i n g up t h e Equations of t h e Theory of S h e l l s , both Connected with t h e Method of Successive Approximations. F i r s t Version of Establishment of t h e Elastodynamic System of Equations 12. Approximate Expressions f o r t h e Components of t h e Displacement Vector and t h e Components of t h e S t r e s s Tensor 13. Boundary Conditions 1. F i r s t Boundary Problem 2. Second Boundary Problem l . 4 .I n i t i a l Conditions. General Remarks on t h e F i r s t Version of t h e S o l u t i o n of t h e Problem of Reduction 15. Application of t h e General Equations of Dynamics t o t h e S o l u t i o n of t h e Problem of Reduction 16. D i f f e r e n t i a l Equations of t h e O s c i l l a t i o n s of a S h e l l 17. Natural Boundary Conditions Derived from t h e Variational Equations (Bib1.15,16) 1. With Rigidly Attached Contour Surface 2. With Free Contour Surface 18. I n i t i a l Conditions 19. On Concentrated Forces 20. Second Version of t h e Solution of t h e Problem of Reduction 21. F i r s t Group of Elastodynamic Equations of t h e Theory of S h e l l s 22. Second Group o f Elastodynamic Equations of t h e Theory of S h e l l s 23. Boundary and I n i t i a l Conditions 24.. Generalized Conclusions and Further Development of t h e Analytic Mechanics of S h e l l s 1. Choice of Generalized Coordinates Corresponding t o t h e Optimum Quadratic Approximations 2 . One of t h e New Versions of t h e Choice of Generalized Coordinates 25. Application of Analytic Methods t o t h e Theory of O s c i l l a t i o n s of layered S h e l l s 26. Equations of O s c i l l a t i o n of a Twolayered S h e l l

................................... ... ......... ..... .....

92

95 98 1 0 1

............... ................ ......................... ................... .................. ................................ ................................... .................................. . .. .... ............... .......................... ...................... ........................

103
1 1 0 112 112 113

116
119

126 127 127 127 129

132 134. 135

........................ ........................ ............. ....

137
138
138

.............. .... ...............................

viii

..........................

14.7

150

Section 27. Section 28.

Section 29. Section 30.

............. .................................... .... Section 31. .......................... ........................ CHAPTER I V . Approximately Equivalent Systems ................ Section 1 . Introductory Remarks .......................... Section 2. F i r s t Method of Linear A p p r o d t i o n of t h e Components of t h e S t r e s s Tensor and t h e Finite-Deformation Tensor ..................... 1. O n t h e Construction of an I s o t r o p i c , ..... Approxhately Equivalent, E l a s t i c Body 2. Connection with t h e Theory of Optimum Systems .................................... 3 . Determination of t h e Parameter a ........... Section 3. Second Method of Linear Approximation o f t h e Components of t h e S t r e s s Tensor and t h e Finite-Deformation Tensor ..................... 1 . Preliminary S e l e c t i o n of t h e Region o f
Approximate Representation of t h e P o t e n t i a l Energy by t h e k e r g y 1 1 0 2. Preliminary Delimitation of t h e Region of Variation of t h e Q u a n t i t i e s nik

D i f f e r e n t i a l Equations of Motion of a Two-Layered S h e l l Natural Boundary Conditio1 . With t h e Contour Surface Kinematically not Free 2 . With t h e Contour Surface Free C l a s s i c Theory of S h e l l s 1 . Forces and Moments 2. Equations of Equilibrium and Motion B r i e f Survey of Recent Results of Reducing t h e Three-Dimensional Problem of t h e Theory of E l a s t i c i t y t o t h e Two-Dimensional Problem of t h e Theory of S h e l l s 1 . Reduction by t h e Use of S e r i e s . Applic a t i o n of t h e DtAlembert-Lagrange P r i n c i p l e 2. The tlSemi-Inverset' Method of Reduction 3 . Reduction by Determining t h e Coeffic i e n t s of t h e Expansion of t h e Displacement Vector Components i n S e r i e s , i n s p e c i a l Functions of t h e z Coordinate 4. Generalized Formulations of t h e Dynamic Problems of t h e Theory of P l a t e s and Shells Comparison of Various Methods of Reduction 1. Equations Obtained by Use of Expansions i n Tensor S e r i e s 2. Equations Resulting from t h e DtAlembertLagrange P r i n c i p l e

............................ .................. ................................. ............ ..................... ....................... ......

......................
....

158
160

160 160

162 162

165

167

..

168
169

170

171 171
172 172

173

173

l73
173 179 179
182

.................... ............

189
190

ix

Page

3. Determination of t h e Averaged Q u a n t i t i e s

of X4t and p* Section 3a. Further Development of t h e Method of Linear Approximation Linearization i n an Element of t h e S h e l l Section !+. Section 5. On t h e Relation between Linear Approximation of t h e Components of t h e FiniteDeformation Tensor and t h e Method o f Equival e n t Linearization and t h e Probability Methods. Further Stages of Successive Approximations n Axisymmetric Deformations and E l a s t i c Section 6. O S t a b i l i t y of a Circular Tube Subjected t o t h e Action of h n g i t u d i n a l Compressive Forces 1 . Evaluation of t h e E f f e c t o f t h e Component on t h e Stressed S t a t e of a Shell, Depending on t h e Ratios To, :T and h:R Section 7. Brief Conclusions 1. On t h e Mechanism of Development of a Local Equilibrium and Motion I n s t a b i E t y of a S h e l l 2. Role of Random Imperfections of Shape 3. Region o f S t a t i s t i c a l I n s t a b i l i t y Section 8. Construction of a Uniform I s o t r o p i c S h e l l Approximately Equivalent t o a Layered S h e l l 1 . Application of an Incompatible System of Algebraic Equations 2. Evaluation of t h e Weights cl 3. Application of t h e Weighted Quadratic Approximation 4. Application of Boundary-Problem Solut i o n s of t h e Dynamics of Homogeneous S h e l l s t o t h e Construction of a Homogeneous S h e l l Approximately Equivalent t o a layered S h e l l Section 9. Construction o f t h e Approximate Solution of Problems of t h e Dynamics of Layered Shells. Application of t h e Method of Pert u r b a t i o n s and Non-Removable Errors Section 10. Application of Optimum Quadratic Approximations t o t h e Problem of Reduction of t h e Three-Dimensional Problem of E l a s t i c i t y Theory t o t h e Two-Dimensional. Problem Section 1 1 . Approximate Expressions of t h e Displacement Vector Components and t h e Equations of Motion of a She1.l Section 12. Boundary Conditions. Various Versions of t h e Solution of t h e General Problem of t h e

.............................. ................................. .....

190

191
192

...............................

198

GI

....................................... ........ ............................

................................... ..... ......... ... .................... .............. .............................

200 208 224 224 225 225 220 227

238
240

........................

247

..........

253

........ .........................
X

255 258

Dynamics of S h e l l s . I n i t i a l Conditions 1.. Remarks on Boundary Conditions 2. O n t h e Ekistence and Uniqueness of Solutions of t h e Boundarg Problems Posed 3. Natural Boundary Conditions Section 13. Approximate Methods of I n v e s t i g a t i n g t h e Equilibrium and 0 s c i l . l a t i o n s of S h e l l s as Discrete-Continuum Systems Section 1.4. The Fundamental D i s c r e t e System of Un-

knowns

Section 15. Boundary Conditions and t h e Equations of Connectivity. I n i t i a l Conditions 1. F i r s t Boundary Problem 2. Second Boundary Problem 3. I n i t i a l Conditions Section 16. Equations of Motion of a S h e l l Section 17. Concluding Remarks CHAPTER V. Section 1. Section 2. I n t e g r a l and Integro-Differential of t h e Theory of S h e l l s

..... ........... ... .............. .................. ...................................... ............. ................... .................. ....................... .............. ..........................
....................... ...................
Equations

Page 270 270 271 272 275 276 279 280 281. 282 283 284
286

Section 3. Section

4.

Section 5.

Section 6.

Section 7. S e c t i o n 8. Section

9.

General C h a r a c t e r i s t i c s of t h e Contents of t h e Concluding Chapter Elementary S o l u t i o n s o f t h e ThreeDimensional Problems of t h e Theory of E l a s t i c i t y Containing Singular Points and Lines Integro-Differential and I n t e g r a l Equat i o n s of t h e S t a t i c s of S h e l l s w'ith Focusing Kerne1.s Methods of Approximate Solution of a System of I n t e g r a l Equations of S h e l l Theory Integro-Differential and I n t e g r a l Equat i o n s of t h e Dynamics of S h e l l s 1 . S t a t i o n a r y O s c i l l a t o r y Process 2. Nonstationary O s c i l l a t o r y Process Local. Systems of Integro-Differential Equations of t h e Dynamlcs of S h e l l s with Focusing Kernels and t h e i r Approximate Solution 1. S t a t i o n a r y O s c i l l a t o r y Processes. The Frequency Spectrum 2. Nonstationary Processes Application of t h e Discrete-Continuum Method Nonlinear Integro-Differential Equations of t h e Dynamics of S h e l l s On t h e Construction of Kernels of

286

................................... ..................................... ...................................... ............. ........... ........

.................................... ....................... .................. ...................................... ...................
xi

286
300

3 14
320 322 323

32k
325 3 27

3 29

334

Page Integro-Differential Equations with Focusing P r o p e r t i e s Section 10. Integro-Differential Equations Defining Contiguous Solutions o f t h e Boundary Problems of t h e S t a t i c s and Dynamics o f Shells 1 . Concluding Remarks on t h e IntegroSection 1 D i f f e r e n t i a l and I n t e g r a l Equations of t h e S t a t i c s and Dynamics of S h e l l s

.........................

343

...................................... ..............

345
347

350

SUBJECT INDEX

............................................

354 356

xii

PREFACE

Le

The s t a t i c s and dynamics of t h i n e l a s t i c s h e l l s have been comprehensively studied, b u t t h e problem o f developing accurate and e f f e c t i v e methods of calcul a t i n g s h e l l s s t i l l r e t a i n s a l l of i t s current i n t e r e s t . The o b j e c t of t h e present work i s t o study and systemize various new boundary problems of t h e s h e l l theory and t h e methods of t h e i r s o l u t i o n derived from t h e general equations of t h e s t a t i c s and dynamics of e l a s t i c bodies i n t h e three-dimensionally s t r e s s - s t r a i n state. This i n v e s t i g a t i o n permits an i n d i c a t i o n of e f f e c t i v e methods f o r solving t h e s e boundary problems i n both l i n e a r and nonlinear form. The i n t e r r e l a t i o n between t h e three-dimensional problems of t h e e l a s t i c i t y theory and t h e two-dimensional problems of t h e s h e l l theory i s a n a l y t i c a l l y established, without r e q u i r i n g t h e use of auxiliary kinematic hypotheses, such as t h e most familiar of a l l , namely t h a t of Kirchhoff and Love.
The general method of i n v e s t i g a t i o n i s r e f l e c t e d i n t h e arrangement of t h e Chapters and Sections. The b a s i c concepts of t h e book a r e linked with t h e a n a l y t i c a l i n v e s t i g a t i o n s of t h e problems of s h e l l theory which t h e author f i r s t took up i n 1937-1938.

It must not, however, be assumed t h a t t h i s work i s a mere r e c a p i t u l a t i o n of t h e r e s u l t s of twenty y e a r s of work. Those r e s u l t s a r e given only p a r t i a l l y here and must be regarded as preparatory s t a g e s i n t h e development of new methods of t h e a n a l y t i c theory of s h e l l s s e t f o r t h i n t h e main Chapters of t h i s book. I n turn, t h e book focuses t h e a t t e n t i o n of t h e reader on t h e s t a t u s o f t h e i n v e s t i g a t i o n s now completed by t h e author, and c o n s t i t u t e s another s t e p toward t h e next s t a g e i n t h e a n a l y t i c theory o f s h e l l s . T h i s i s why t h e book should be regarded a s a f i r s t p a r t of t h e study on s h e l l theory. The second p a r t w i l l be ready t o go t o p r e s s i n or about 1961,. 'de have included i n t h e book t h e p r i n c i p l e s of t h e a n a l y t i c theory of s h e l l s , confining ourselves t o an extended discussion of t h e new methods and of t h e r e s u l t i n g general formulations of t h e boundary problems of t h e s t a t i c s and dynamics of s h e l l s , i l l u s t r a t e d by a l i m i t e d number of examples. The second p a r t of t h e book c o n s i s t s of a p p l i c a t i o n s of t h e theory t o s p e c i f i c problems. These a p p l i c a t i o n s w i l l , i n turn, undoubtedly encourage t h e f u r t h e r development and enrichment of t h e theory. The reader i s assumed t o be acquainted with mathematical a n a l y s i s a t t h e u n i v e r s i t y l e v e l , t h e o r e t i c a l mechanics, theory of e l a s t i c i t y , and c l a s s i c a l theory of s h e l l s . The s p e c i a l methods of mathematical a n a l y s i s , p a r t i c u l a r l y t h e t e n s o r calculus, a l s o play a considerable r o l e i n t h i s work. This i s because of t h e f a c t t h a t t h e apparatus of modern higher geometry, namely t h e t e n s o r calculus, i s b e s t s u i t e d f o r t h e construction of t h e a n a l y t i c a l mechanics of s h e l l s . W e t h e r e f o r e deemed it expedient t o make use of this apparatus, which i s gradually

xiii

e give p e n e t r a t i n g o t h e r f i e l d s of t e c h n i c a l science. I n Chapters I and I1 w t h e b a s i c information on t e n s o r a n a l y s i s and nonlinear e l a s t i c i t y theory, nece s s a r y for t h e understanding of t h e theory of s h e l l s which i s developed there-

after.
The r e a d e r will note t h a t c e r t a i n o f o u r new results i n t h e nonlinear /10 theory of e l a s t i c i t y , which are included i n Chapter 1 1 , a r e n o t f u r t h e r mentioned i n t h e l a t e r Chapters concerning t h e theory of s h e l l s . These r e s u l t s , however, will b e applied i n t h e next p a r t of t h e book. Although t h e t i t l e of t h e present book i s "Principles of t h e Analytic M e chanics of Shells", i t does n o t contain an exhaustive treatment of t h e applicat i o n s of t h e mechanics of Lagrange, Ostrogradskiy, Hamilton, Gauss, Jacobi, Hertz, Chaplygin, and o t h e r s t o t h e problems of s h e l l theory, although t h e recent works on a n a l y t i c a l mechanics do i n d i c a t e ways of extending t h e s e methods t o t h e mechanics of a continuous medium. The development of t h e mechanics of s h e l l s i n t h i s d i r e c t i o n i s unquestionably of i n t e r e s t , b u t will r e q u i r e a cert a i n amount of t i m e . Finally, a word on our method of i n t e r n a l reference t o formulas t o be found elsewhere i n t h e book, i n t h e l i t e r a t u r e o r i n o t h e r sources. W e have adopted t h e following method of i n d i c a t i n g references. A n entry of t h e form (11, Sect.3) means a reference t o t h e content of Section 3 , Chap1 , while t h e symbol (111, 11.5) means a reference t o eq.(5) of Section 1 1 , ter 1 Chapter 1 1 1 , and so on. The l i t e r a t u r e references a r e divided i n t o two groups. The first includes t h e p r i n c i p a l sources, which a r e l i s t e d a t t h e end o f t h e book. References t o sources not included i n t h i s Bibliography, a r e given as footnotes on t h e pert a i n i n g page. The Bibliography a l s o includes c e r t a i n works t o which t h e text does not r e f e r b u t which helped i n formulating t h e i d e a s expressed i n i t s content. This Bibliography,of course, makes no claim a t completeness, and t h e d i v i s i o n i n t o two groups i s q u i t e a r b i t r a r y . The author expresses his thanks t o A.S.VolTmir, A.D.Kovalenko, and G.N.Savin f o r checking t h e manuscript and f o r valuable comments, and a l s o t o A.Kh.Konstantinov and G.L.Komissarova of t h e I n s t i t u t e of Mechanics, UkrSSR Academy of Sciences, f o r reading t h e manuscript, t o 2.I.Yasinchuk and L.A.Rudneva who p a r t i c i p a t e d i n i t s t e c h n i c a l preparation, and t o S.G.Shpakov who performed some of t h e computations. Kiev, September

1960 t o January 1962

K .Kilt chevskiy

xiv

INTRODUCTION

&

A s o l i d , bounded by two boundary s u r f a c e s and by a contour surface i n t e r s e c t i n g t h e boundary s u r f a c e s along t h e contour curves, i s c a l l e d a s h e l l . Between t h e boundary s u r f a c e s l i e s t h e b a s i c ( o r coordinate) surface, whose sel e c t i o n i s a r b i t r a r y and i s based on t h e conditions of t h e s p e c i f i c problem. The o b j e c t of this s e l e c t i o n i s t o simplify t h e system of equations of t h e theory of s h e l l s .
The l e n g t h of t h e segment of t h e normal t o t h e b a s i c s u r f a c e included between t h e boundary s u r f a c e s of t h e s h e l l i s c a l l e d t h e thickness of t h e s h e l l , which will h e r e a f t e r be denoted by 2h. The thickness of a s h e l l may be e i t h e r constant o r variable. The l o c u s of t h e midpoints of t h e segments of t h e norm a l s t h a t d e f i n e t h e t h i c k n e s s of t h e s h e l l will be a r b i t r a r i l y termed t h e midd l e surface of t h e s h e l l .
A c h a r a c t e r i s t i c f e a t u r e o f t h e s h e l l i s t h e smallness of t h e r a t i o 2h:a, where a i s a c e r t a i n parameter determining t h e dimensions of t h e s h e l l . For example, for coverings, a i s one of t h e dimensions d e f i n i n g t h e p r o j e c t i o n of t h e covering on a h o r i z o n t a l plane. Sometimes one of t h e p r i n c i p a l r a d i i of curvature of t h e middle surface i s s e l e c t e d a s t h e parameter a. The geometric a l c h a r a c t e r i s t i c s of s h e l l s will be discussed i n g r e a t e r d e t a i l i n Chapter

111.
S h e l l s are common elements of various machines and s t r u c t u r e s , because of t h e e x c e l l e n t s t r e n g t h c h a r a c t e r i s t i c s of designs with thin-walled elements of t h e s h e l l type. This book w a s w r i t t e n i n a period o f i n t e n s e development of t h e s t a t i c s and dynamics of t h i n s h e l l s . Research i s being pressed i n v a r i o u s d i r e c t i o n s , and it would be d i f f i c u l t today t o specify any one group o f works t h a t could with complete j u s t i f i c a t i o n be c a l l e d t h e b a s i s of t h e theory. For t h i s reason, t h e contents of t h i s book are t o some e x t e n t a r e f l e c t i o n of t h e narrow s c i e n t i f i c i n t e r e s t s of i t s author. The choice of t h e problems touched on i n t h e main p a r t of t h e book (Chapters 1 1 1 - I V ) has been determined by t h e contents o f t h e well-known monographs by S.A.Ambartsmyan, I.A.Birger, V.V.Bolotin, V.Z .Vlasov, A .S.Voltmir, A.L.Gol*denveyzer, Kh.M.Mushtart and K.Z .Galimv, /12 A.I.Lurfye, V.V.Novoehilov, and 0.D.Oniashvili. W e have attempted here t o cons i d e r and analyze those t r e n d s of research i n t o s h e l l theory t h a t have not been s u f f i c i e n t l y covered by t h e above-mentioned major works. It must b e emphasized t h a t complete attainment of t h i s aim would t a k e u s beyond t h e scope of t h e present volume, and we have t h u s had t o confine ourselves t o t h e construction of separate fragments of t h e o r y which, i n our opinion, a r e a b a s i s f o r f u r t h e r investigations. W e have thought it expedient t o concentrate a t t e n t i o n on t h e development of a n a l y t i c a l methods of i n v e s t i g a t i o n based, more p a r t i c u l a r l y , on t h e theory of i n v a r i a n t s of coordinate transformations and t h e a n a l y t i c d e f i n i t i o n of t h e b a s i c geometric operations performed on v e c t o r and t e n s o r f i e l d s t h a t d e f i n e

t h e physical state of a s h e l l . The working apparatus connected with t h e theory of i n v a r i a n t s i s t h e c a l c u l u s o f tensors, t o g e t h e r with t h e p r i n c i p l e s and vari o u s propositions of c l a s s i c a l a n a l y t i c a l dynamics. W e have o f t e n made use of a more popular method of approximate r e p r e s e n t a t i o n of functions, t h e method of c o n s t r u c t i n g approximation functions t h a t satisfy t h e requirement of t h e leastsquare d e v i a t i o n from t h e approximation function within t h e region of t h e approximate r e p r e s e n t a t i o n required. This method of i n v e s t i g a t i o n , in our opinion, a l s o belongs t o t h e amalyt i c a l mechanics of s h e l l s . Here we d e p a r t from t h e c l a s s i c a l concept of t h e f i e l d of a n a l y t i c mechanics, b u t t h i s formal d e v i a t i o n i s thoroughly j u s t i f i e d by t h e e s s e n t i a l n a t u r e of t h e method, t h e more so t h a t i t permits t o o b t a i n t h e general equations o f motion of a s h e l l and i s linked t o one of t h e fundamental v a r i a t i o n a l p r i n c i p l e s of mechanics, namely t h e Gauss principle.

To f a c i l i t a t e t h e reading of t h e book, t h e main content i s preceded by Chapters I and 1 1 , which give i n o u t l i n e t h e elements o f t e n s o r a n a l y s i s and d i f f e r e n t i a l geometry, i n d i c a t e t h e elementary geometric p r o p e r t i e s o f s h e l l s , and present t h e r e l a t i o n s o f t h e l i n e a r and nonlinear theory of e l a s t i c i t y t o which l a t e r reference will be made. Some of t h e s e r e l a t i o n s , a s mentioned i n t h e Preface, will be used i n t h e next Volume o f t h i s work.
Most of t h e s u b j e c t m a t t e r in t h e main Chapters ( 1 1 1 - V ) i s connected w i t h t h e theory of s m a l l displacements and deformations of s h e l l s , described by l i n e a r d i f f e r e n t i a l and i n t e g r o - d i f f e r e n t i a l equations. I n a number of cases, we consider problems of t h e nonlinear theory.
ill

The problems of s h e l l dynamics occupy t h e c e n t r a l p o s i t i o n i n t h e book. I n Chapter I11 we consider t h e l i t t l e - i n v e s t i g a t e d methods of reducing t h e three-dimensional problems of t h e dynamics of homogeneous and inhomogeneous (layered ) s h e l l s t o two-dimensional problems. W e analyze t h e boundary condit i o n s and i n i t i a l conditions. W e compare t h e v a r i o u s methods of s e t t i n g up t h e equations f o r d e f i n i n g t h e two-dimensional problems of s h e l l theory. Here we make no use of t h e simplifying assumptions i n h e r e n t i n t h e Kirchhoff-Love hypotheses. The refinements of t h e equations of t h e theory of s h e l l s , once considered by some s c i e n t i s t s t o b e of purely t h e o r e t i c a l i n t e r e s t , have now assumed profound s i g n i f i c a n c e i n connection w i t h t h e study of dynamic processes r a p i d l y proceeding in time. Quite meaningful in t h e s e cases i s t h e i n v e s t i g a t i o n of high-frequency o s c i l l a t i o n s , which a r e u s u a l l y damped more r a p i d l y than those of low frequency. I n studying t h e b r i e f dynamic processes caused by t h e short-time a c t i o n of forces, t h e e f f e c t of t h e d i s s i p a t i o n o f energy i s not o f d e c i s i v e importance, and we must study a broader segment of t h e frequency spectrum than i n studying slowly proceeding processes::.
-x- The concept of t h e speed of a process i s of course r e l a t i v e . The n a t u r a l measure of time i n this case, i n our opinion, i s t h e time i n t e r v a l required for t h e spread of dynamic disturbances over t h e e n t i r e region within t h e s h e l l .

The r e f i n e d d i f f e r e n t i a l equations of t h e s h e l l theory e s t a b l i s h systems In t h i s of higher o r d e r than t h e system of equations of t h e c l a s s i c a l theory. connection, t h e r e arises again t h e problem of t h e formulation of t h e boundary and i n i t i a l conditions completing t h e formulation of t h e dynamic boundary prob1 1 and I V . If we r e c a l l t h e lems. These questions a r e discussed i n Chapters I h i s t o r y of t h e development of modern s h e l l theory, we can c l e a r l y apprehend t h e fundamental d i f f i c u l t i e s involved i n t h e generalized formulation of t h e boundary conditions i n t h i s f i e l d of applied e l a s t i c i t y theory. The author d i s t i n c t l y d e p i c t s t h e s e d i f f i c u l t i e s and t h e c o n t r o v e r s i a l n a t u r e of a nunber of propositions advanced by him. Chapter I V considers v a r i o u s approximate methods of solving t h e problems of s h e l l mechanics - a l l methods governed by a s i n g l e common idea. Their essence r e s i d e s in t h e replacement of t h e s h e l l by an e l a s t i c system approximately equivalent t o i t i n r e s p e c t t o c e r t a i n general features. I n p a r t i c u l a r , we propose a new approximation method of solving t h e nonlinear problems of s h e l l theory, c l o s e l y resembling t h e method of equivalent l i n e a r i z a t i o n known from n o n l i n e a r mechanics of systems with one degree of freedom. The perspective of t h i s method i s t o some extent confirmed by t h e example f o r t h e s o l u t i o n of t h e problem of s t a b i l i t y of a closed c y l i n d r i c a l s h e l l , given i n Chapter IV, Sect.6.
W e a l s o d i s c u s s t h e construction of a homogeneous s h e l l approximately equivalent i n number of l a y e r s . This construction i s based on t h e approximat i o n o f t h e Lagrange f u n c t i o n of a layered s h e l l by t h e Lagrange function of a homogeneous s h e l l . To r e a l i z e t h i s approximation, i t w a s necessary t o s e l e c t a s p e c i a l system of v a r i a b l e s which, i n t h e modern l i t e r a t u r e on mechanics, /14 a r e termed v a r i a b l e f i e l d s .
A preliminary s u b s t i t u t i o n of a homogeneous s h e l l for a layered s h e l l perm i t s , a s shown i n Chapter I T , t o develop a method of approximate determination of f i e l d s of displacement, deformation, and s t r e s s i n layered s h e l l s .

Further, t h e method of c o n s t r u c t i n g t h e best-square approximations i s app l i c a b l e t o t h e establishment of a new system of equations of motion of a s h e l l element, modifying i n t h i s case a general p r i n c i p l e of a n a l y t i c a l dynami c s , namely t h e DfAlembert-Lagrange p r i n c i p l e . Obviously, t h i s involves t h e connection between t h e method of l e a s t squares and t h e p r i n c i p l e of l e a s t c o n s t r a i n t , s t a t e d by Gauss, which possesses t h e same degree of g e n e r a l i t y as t h e DTAlenibert-Lagrange p r i n c i p l e . The r e s u l t a n t system of equations has a number o f p r o p e r t i e s permitting i t s use as a means of solving new problems of s h e l l theory. The s o l u t i o n o f t h e l i n e a r problems of s h e l l theory, and p a r t i c u l a r l y t h e nonlinear problems, reduces i n t h e general case and a t t h e p r e s e n t l e v e l of development t o t h e numerical s o l u t i o n of systems of l i n e a r and nonlinear algeb r a i c equations. W e have t h e r e f o r e deemed i t advisable t o consider, i n Chapter I V , a method based on t h e a p p l i c a t i o n o f i n t e r p o l a t i o n fomulas,making it p o s s i b l e t o reduce t h e problem of t h e motion of s h e l l elements t o t h e solut i o n o f a f i n i t e s y s t e m of ordinary d i f f e r e n t i a l equations, which are t h e

Euler-Lagrange equations f o r t h e corresponding v a r i a t i o n a l problem. The select i o n of t h e v a r i a t i o n p r i n c i p l e , permitting t o set up t h e equations of motion, depends on t h e p r o p e r t i e s of t h e r e l a t i o n s r e s u l t i n g from t h e boundary condit i o n s of t h e problem when applying t h e method of reduction based on t h e expans i o n i n t e n s o r series of t h e required components of t h e displacement v e c t o r and I n t h e general case, t h e s e r e l a t i o n s a r e kinematic and bet h e stress tensor. long t o types t h a t have not been studied in c l a s s i c dynamics. I n Chapter IV, t h e question of t h e formulation of t h e i n i t i a l conditions i s again i n v e s t i gated. The mentioned method makes it p o s s i b l e t o lay t h e foundation f o r a numere term t h i s i c a l s o l u t i o n of s h e l l theory problems by t h e use of computers. W method t h e discrete-continuous method, following t h e terminology proposed by V.Z.Vlasov. The methods of s o l u t i o n o f t h e boundary problems of s h e l l theory, developed i n Chapter V, l i k e w i s e have t h e o b j e c t of l a y i n g t h e foundation f o r programming t h e numerical s o l u t i o n of t h e boundary problems of s h e l l theory. Here we i n d i c a t e systems of i n t e g r o - d i f f e r e n t i a l equations t h a t r e s u l t from t h e theorem of work and r e c i p r o c i t y i n both i t s conventional treatment and a s generalized by u s t o t h e case o f a n o n l i n e a r l y deformed a n i s o t r o p i c medium. The g e n e r a l i z a t i o n o f t h e theorem of r e c i p r o c a l work i s presented i n Chapter 1 1 . I n Chapter V a new method i s described f o r reducing t h e three-dimensional problems of t h e e l a s t i c i t y theory t o two-dimensional problems of t h e s h e l l theory. This method permits u s t o d e s c r i b e a f i e l d of displacements within /1-5 a s h e l l by systems of i n t e g r o - d i f f e r e n t i a l equations or, i n p a r t i c u l a r , of int e g r a l equations with k e r n e l s having p e c u l i a r p r o p e r t i e s , which w e have c a l l e d frfocusingfti n accordance with t h e term proposed by K.Lantsosh+t. The i n t e g r o - d i f f e r e n t i a l equations o f t h e s h e l l theory, and t h e i n t e g r a l equations with focusing kernels, a r e undoubtedly of considerable t h e o r e t i c a l and applied importance. By applying t h e discrete-continuous method t o systems of i n t e g r o - d i f f e r e n t L a l equations with focusing kernels, we approximately reduce t h e problems of s h e l l dynamics t o t h e s o l u t i o n of r e l a t i v e l y simple systems of l i n e a r ordinary d i f f e r e n t i a l equations o r , i n more general cases, of nonlinear types. I n problems of s t a t i c s , t h e s e systems degenerate t o systems of a l g e b r a i c equations

Thus, t h e s o l u t i o n s of t h e boundary problems of s h e l l s t a t i c s and dynamics may be found i f t h e c o e f f i c i e n t s of t h e s e ordinary d i f f e r e n t i a l equations or t h e a n a l y t i c a l expressions f o r t h e k e r n e l s with focusing p r o p e r t i e s a r e known. I n Chapter V w e i n d i c a t e two methods of constructing t h e s e kernels. The f i r s t method i s based on expansions of t h e functions, permitting a construction The i n v e s t i g a t i o n of t h e s e exof focusing kernels, i n Legendre polynomials. pansions i s connected with problems only a s t e p removed from t h e c l a s s i c a l

Cf.K.Lantsosh, P r a c t i c a l Methods of Applied Analysis. Physical and Mathem a t i c a l Publishing House, 1961. (Fizmatgiz)

4
11111I I I I 1 1 1 1 11 1 1 1 1 1 1 1 1 1 1 1 1I

1 -1 1 1 1 1 1 1 1 1 1 . 1 1 1 1 1 1

I I

11111111

problem of moinentSi5. The f i r s t method allows us t o f i n d k e r n e l s with s t r o n g e r focusing p r o p e r t i e s than does t h e second method. The second method i s simpler than t h e f i r s t and does n o t r e q u i r e complicated a d d i t i o n d mathematical invesIn the t i g a t i o n s , b u t l e a d s t o k e r n e l s with weakened focusing properties. second volume of t h i s work, we s h a l l i n d i c a t e s p e c i f i c a n a l y t i c expressions f o r focusing k e r n e l s f o r s p e c i a l t y p e s of s h e l l s and s h a l l p r e s e n t Tables f o r f i n d i n g t h e numerical v a l u e s f o r t h e c o e f f i c i e n t s of t h e approximate d i f f e r e n t i a l equations of s h e l l dynamics and t h e a l g e b r a i c equations of s h e l l s t a t i c s , t h a t follow from t h e i n t e g r o - d i f f e r e n t i a l equations with focusing kernels. At t h e same time, w e s h a l l continue t h e i n v e s t i g a t i o n of t h e a n a l y t i c p r o p e r t i e s of t h e i n t e g r o - d i f f e r e n t i a l equations of s h e l l theory with focusing kernels. These i n v e s t i g a t i o n s w i l l allow u s t o f i n d standard programs f o r t h e computat i o n o f s h e l l s of a r b i t r a r y form on modern computers. Such i s t h e plan f o r f u t u r e i n v e s t i g a t i o n s , with t h e o b j e c t of establ i s h i n g a general and e f f e c t i v e method f o r solving problems of s h e l l s t a t i c s and dynamics, i n both l i n e a r and nonlinear formulations.

3s

Cf.N.I.Akhiyeeer, The C l a s s i c a l Problem of Moments. matical Publishing House, 1961.

Physical and Mathe-

CHAPTER I
E;LEME;C\rTS OF TENSOR ANALYSIS AND THEIR APPLICATION TO THE DIFFERATTIAL GEDXE3'RY OF SHFLLS Section 1,General Description of t h e Applications o f Tensor Analysis i n S h e l l Theo_ry

/16

Tensor a n a l y s i s i s a modern mathematical apparatus permitting expression, i n t h e most general a n a l y t i c a l form, of t h e fundamental geometric operations performed on t h e q u a n t i t i e s encountered i n t h e i n v e s t i g a t i o n of various problems of geometry and physics. Among these o p e r a t i o n s we may note elementary operations, f o r i n s t a n c e t h e measu-rement o f d i s t a n c e s between p o i n t s of space, or measurement of angles between d i r e c t e d segments, and more complex operat i o n s , t o which reduces t h e mutual comparison of geometric and physical obj e c t s , of a given system o f nunibers or system of functions of c u r v i l i n e a r coord i n a t e s of p o i n t s i n space. P a r t i c u l a r l y important i s t h e problem of constructing q u a n t i t i e s independent of t h e choice o f t h e coordinate system. These q u a n t i t i e s a r e termed inTensor q u a n t i t i e s a r e t h e base f o r t h e v a r i a n t s of coordinate transformations. construction of i n v a r i a n t s . The s c a l a r s and vectors, which we know f r o x e l e mentary geometry and mechanics, a r e s p e c i a l cases o f t e n s o r q u a n t i t i e s . Most i n v a r i a n t s have a d e f i n i t e geometric or physical meaning. I n v a r i a n t s a r e t h e b a s i s f o r t h e general a n a l y t i c a l formulations of t h e l a w s o f physics, e s p e c i a l l y those of mechanics. The a p p l i c a t i o n s of t e n s o r a n a l y s i s t o t h e geometry o f s u r f a c e s a r e numerous, since here t e n s o r a n a l y s i s allows u s t o f i n d expressions of geometric theorems i n a simple and y e t general form. The kinematics and k i n e t i c s of s h e l l s i s p r e c i s e l y t h e t h a t i s i n t e r n a l l y linked t o t h e geometry of surfaces. The boundary problems of s h e l l theory r e q u i r e s t h e i n t r o d u c t i o n o r d i n a t e systems d e f i n i n g t h e p o s i t i o n of t h e p o i n t s of t h e branch of mechanics formulation o f t h e of c u r v i l i n e a r coshell.

To s e t up t h e kinematic and k i n e t i c equations of s h e l l theory without / 1 7 recourse t o t h e methods o f t e n s o r a n a l y s i s i s a cumbersome and complicated operwhat i s very important i t sometimes involves l o s s e s o f various a t i o n , and terms of t h e equations. Errors of t h i s kind a r e most f r e q u e n t l y encountered i n attempts t o s e t up t h e equations of s h e l l theory on t h e b a s i s o f elementary "visualized" concepts.

The apparatus of t e n s o r a n a l y s i s , as already remarked, was developed f o r t h e very purpose o f solving t h e problems of geometry and mechanics i n curvil i n e a r coordinate systems. This apparatus i s most s u i t a b l e for solving various problems of s h e l l mechanics. A l l t h e operations necessary f o r s e t t i n g up t h e kinematic and k i n e t i c equations of s h e l l theory receive, i n t h e framework o f

t e n s o r a n a l y s i s , a rigorous a n a l y t i c i n t e r p r e t a t i o n t h a t makes i t unnecessary t o appeal t o v i s u a l i z e d ttobvioustr ideas. While t e n s o r a n a l y s i s does allow u s t o set up t h e equations of s h e l l theory, it does not, of course, eliminate t h e d i f f i c u l t i e s o f solving t h e corresponding boundary problems. However, it p e n i t s u s t o mark new methods of solvi n g t h e dynamic boundary problems, based, f o r instance, on i n t r o d u c t i o n of t h e functions of k i n e t i c s t r e s s e s (Bibl.?), Section 2. Systems of C u r v i l i n e a r Coordinates. Metrics of Space. The Symbol f o r Summation. I n passing t o a d i s c u s s i o n of t h e mathematical p r i n c i p l e s of t h e theory of s h e l l s , we assume t h a t t h e r e a d e r i s f a m i l i a r with t h e r u l e s of operation of v e c t o r algebra and t h e elements of d i f f e r e n t i a l geometry. S h e l l theory i s based on t h e a p p l i c a t i o n of various c u r v i l i n e a r coordinate systems d e f i n i n g t h e p o s i t i o n of t h e p o i n t s of t h e s h e l l . W e s h a l l f i r s t cons i d e r t h e general p r o p e r t i e s of an a r b i t r a r y coordinate system i n threedimensional space, and s h a l l then i l l u s t r a t e t h e s e p r o p e r t i e s by examples from s h e l l theory.
A system of independent parameters, uniquely determining t h e p o s i t i o n of e shall t h e p o i n t s i n space, i s c a l l e d a system of c u r v i l i n e a r coordinates. W denote them by xi. I n three-dimensional space t h e number of a coordinate (index) may be 1, 2, 3. O n a c e r t a i n surface, e.g. i n two-dimensional space, t h e index i t a k e s t h e values 1 and 2.

Let u s s e l e c t a fixed p o i n t 0 i n space, and draw t h e r a d i u s v e c t o r OM = r ' t o t h e p o i n t M i n space. Since t h e p o s i t i o n of p o i n t PI i s determined by t h e coordinates xi ( i = l,2,3), t h e r a d i u s v e c t o r i s a function of xi :

- + +

W e s h a l l assume t h a t F(xi ) i s a single-valued continuous function, d i f - / 1 8 f e r e n t i a b l e a t least twice w i t h r e s p e c t t o any argument xi. If t w o coordin a t e s xi o u t of t h e t h r e e are fixed, then eq.(2,1) can be regarded as t h e equat i o n o f a c e r t a i n curve. This curve i s c a l l e d a coordinate l i n e . Three coorConsider t h e two p o i n t s M(xi) and d i n a t e l i n e s pass through t&e p o i n t M. W(xi + dxi ). The v e c t o r MplT i s defined as follows:

where C i s t h e s i g n o f summation.

The d e r i v a t i v e s

-=e,
axi

VI

(i-

1, 2, 3)

( 2 . 3 )
4

are v e c t o r s d i r e c t e d along tangents t o t h e coordinate l i n e s . The v e c t o r s e, form t h e l o c a l coordinGe base a t t h e p o i n t M. Equation (2.2) d e f i n e s t h e expansion of t h e v e c t o r d r on t h e axes of t h e l o c a l coordinate base. The index i i n eq.(2.2) i s c a l l e d a dummy index, s i n c e it t a k e s no d e f i n i t e value b u t runs through a l l values from 1 t o 3 . a u a t i o n (2.2) obviously remains unchanged i f t h e dummy index i i s replaced by any o t h e r l e t t e r . W e s h a l l f r e q u e n t l y make use, h e r e a f t e r , of t h i s r i g h t t o change t h e dummy indices. L e t u s f i n d t h e d i s t a n c e NiY.
W e have

or

where d s =

I I W I ; the
-#

c o e f f i c i e n t s g,

a r e expressed by

Equation (2.4), d e f i n i n g d 8 , i s c a l l e d t h e fundamental q u a d r a t i c form o f t h e q u a n t i t i e s dx' and t h e c o e f f i c i e n t s g1 are c a l l e d t h e c o e f f i c i e n t s o f t h e fundamental q u a d r a t i c form. Below, we w i l l give a d i f f e r e n t term f o r t h e set of q u a n t i t i e s g,

The determinant

/19

(2.6a)

i s c a l l e d t h e fundamental determinant,

It follows from eq.(2.5) and t h e theory of determinants t h a t t h e fundamental determinant i s equal t o t h e square of t h e volume V of t h e parallelepiped constructed on t h e v e c t o r s gi :

Here V i s u s u a l l y taken as a p o s i t i v e quantity. t h e question of i t s properties.

W e shall return later t o

The set of q u a n t i t i e s g,, permits u s t o determine t h e d i s t a n c e between two points. It i s easy t o show t h a t t h e s e t of t h e s e q u a n t i t i e s allows u s t o f i n d By making use r and t h e angle between t h e d i r e c t i o n s of t h e two v e c t o r s $ of expansions of t h e type of eq.(2.2) of t h e v e c t o r s d' r and 6'r and t h e propert i e s of a s c a l a r product, we o b t a i n

rr.

Thus t h e system of f u n c t i o n s of gieJ d e f i n e s t h e m e t r i c of space, e.g. t h e method of measuring, in a given c u m i l l n e a r coordinate s y s t e m , t h e d i s t a n c e s between i n f i n i t e l y near p o i n t s and t h e angles between t h e d i r e c t i o n s of two vectors. EQuations (2.2), (2.4), and (2.7) may be put i n t o a simpler form by making u s e of t h e a r b i t r a r y s u m a t i o n convention proposed by A. Einstein.

3
Hereafter, sums of t h e form

a s a,bi

, omitting

the sign

1=

t o i n d i c a t e it s p e c i f i c a l l y whenever expressions of t h e form a,b* a r e not sums b u t monomials. I f we use t h e simplified n o t a t i o n f o r summation, then, f o r i n stance, eq.(2,4) t a k e s t h e following form:

t.

a ,b' w i l l conventionally be w r i t t e n simply i=l I n t h i s case, of course, it will be necessary

The abbreviated n o t a t i o n f o r summation i s a l s o a p p l i c a b l e t o m u l t i p l e summation.

Another abbreviated n o t a t i o n w i l l be introduced here. W e s h a l l denote t h e operation of d i f f e r e n t i a t i o n with r e s p e c t t o t h e coordinate x* by t h e symbol a, :

/20

Thus,

( 2 . 1 0 )
Section 3 . Metrics i n S h e l l s
A c u r v i l i n e a r system of coordinates i n a s h e l l involves t h e preliminary i n t r o d u c t i o n o f an undeformed base surface on which a network of coordinate l i n e s 2 and x? i s drawn. Plost o f t e n t h e base surface i s taken t o coincide with t h e middle surface of t h e undeformed s h e l l . W e put

are orthonormals t o t h e undeformed b a s i c surface. Thus t h e n o d s where t h e t o t h e undeformed b a s i c surface of t h e s h e l l form a system o f coordinate l i n e s along which t h e coordinate 2 varies. The system of coordinates xi i s t h e Lagrangian system d e f i n i n g t h e p o s i t i o n of t h e p o i n t s of t h e deformable medium c o n s t i t u t i n g t h e s h e l l . Under deformations of t h e s h e l l , t h e coordinates xi o f a m a t e r i a l element of t h e s h e l l do not vary, b u t t h e coordinate l i n e s 2 devia t e from t h e norma,ls t o t h e deformed b a s i c surface. Let To ( 2 , 2 ) be t h e rad i u s vector o f a p o i n t of t h e b a s i c surface o f t h e undeformed s h e l l . Then t h e r a d i u s vector of an a r b i t r a r y point of t h e s h e l l will be expressed by

Tne v e c t o r s of t h e l o c a l coordinate base w i l l be expressed by

1c

If t h e coordinate l i n e s on the b a s i c surface are taken t o coincide with its l i n e s o f curvature, as i s usually,done i n Rodrigues formula+, w e f i n d
+

din =

--

kidiro.

where'the

denote t h e p r i n c i p a l curvatures of t h e b a s i c surface:

and the

&

are t h e p r i n c i p a l r a d i i of curvature.

Thus,

-.
(3.5 1

e, = (1 - k$)

airo

(i = 1, 2).

Hence t h e c o e f f i c i e n t s of t h e fundamental quadratic fords: are: (3.6a) (3.6b) where t h e (ai , ) are t h e c o e f f i c i e n t s of t h e b a s i c quadratic form for 2 = 0, i.e..% on t h e b a s i c surface. e s h a l l show l a t e r , t o find t h e Equations (3.6a) and (3.6b) allow us, as w m e t r i c s f o r EL s h e l l with an a r b i t r a r i l y assigned coordinate n e t on i t s undeformed basi.c surface.. Section 4 , Sk1l-so f Rexolutionc Special Cases of S h e l l s of Revolution, Arbitrary Cylindrical Shells. which t h e b a s i c surface i s a surface of revolution. I f Cansider a. she21 51 t h e a x i s OZ of a rectangular Cartesian coordinate system i s superposed on t h e a x i s of revolution o f t h e b a s i c surface, then t h e vector equation of t h e b a s i c mrfac;e may b e w r i t t e n i n t h e following form:

x- See,
V "

f o r Instance, W.Blaschke,

D i f f e r e n t i a l Geometry.

ONTI, 1936

7-r

H e r e and hereafter t h e exponents are w r i t t e n i n parentheses. we will devtate from tEs rule in cases- where it could n o t cause misunderstanding o f t h e notation,

-+

ro(xl, x 2 ) = F ( x l ) [i'cos cp (9) +;sin

cp ( x 2 ) ] + k z ( x 1 ) .

(4.1

where t h e .equations

determine t h e fohn of t h e meridional s e c t i o n of t h e s h e l l , and t h e angle CD i s a function o f t h e second coordinate 2 d e f i n i n g t h e p o s i t i o n of a p o i n t on a c i r c l e of l a t i t u d e . Thus,

dlro =
f

F'(x') [i cos cp (9) j sin cp ( x z ) ]

f

d2r0 =

F ( x l ) [-

+ + kz'(xI), i sin cp (9) + j cos cp (xz)]

cp'(x2),

where t h e prime denotes d i f f e r e n t i a t i o n with r e s p e c t t o t h e corresponding a r - e gument Since meridional s e c t i o n s and c i r c l e s of l a t i t u d e a r e l i n e s of curvature e obtain on a surface of revolution, l e t u s make use of eqs0(3.6a). W

The r a d i i of curvature Rl and R, a r e determined from eqs.(4.2). consider c e r t a i n s p e c i a l cases.

Let us

1, The C i r c u l a r C y l i n d r i c a l S h e l l Equations (4.2) here taKe t h e form:

x = a = const, 2 = z ( x ' ) .

where a i s t h e r a d i u s of t h e c y l i n d e r and xl i s t h e coordinate d e f i n i n g t h e pos i t i o n of a point on t h e g e n e r a t r i x o f t h e b a s i c surface. Consequently, k, = 0,

kB

= L

It follows from eqs.(4.4a)

6 1 ,

and (4,kb) t h a t

= z'(2)( x ' ) ; g22 = (a - x 3 ) ( 2 ) c p ' ( 2 )

(9).

12

11
~

then

2. Circular-Conical S h e l l Let t h e b a s i c surface of t h e s h e l l be of t h e form of a c i r c u l a r truncated cone. L e t 2 be t h e distance of a point of the basic surface of the s h e l l , , be measured along the generatrix, t o t h e base of greater radius. Let rl and r L e t H be t h e r a d i i of the bases of t h e truncated cone. Assume t h a t rl > r,. t h e a l t i t u d e of the cone, and y t h e angle between t h e generatrix and the axis o f revolution. Let t h e axes OX and OY lie in t h e plane of t h e base of radius rl. Then eqs.(4.2) take t h e form:

(4.7)
From eqs.(lb.4a)
and (4.4b),

remembering t h a t kl = 0, w e find

where

/23

3. The S h e l l with t h e Base Area i n the Form of a

Hyperboloid of Revolution Consider a s h e l l whose base a r e a i s the hyperboloid formed by the revolut i o n of t h e hyperbola

about t h e axis OZ.

Equations (4.2) take t h e form of

13

. _

Then,

The q u a n t i t i e s gll,

e2 a r e

defined 3 ~ y eqs.(4,@)

and (4,Lb).

Consider, f i n a l l y a s h e l l with a c y l i n d r i c a l base area and an arbitrary directrix.

L e t t h e coordinate 2 define t h e d i s t a n c e of a point of t h e b a s i c surface, measured along t h e generatrix, from one of t h e f a c e sections, and t h e coordin a t e 2 be equal t o t h e l e n g t h of t h e a r c of a s e c t i o n o f t h e b a s i c "?ace by a plane normal t o t h e generatrix, measured from one -of t h e generakrices, Then t h e element of t h e a r c o f t h e b a s i c surfaee will be

Consequently,

and

Section 5. Scalars, Vectors and T h e i r Contravariant and Covariant Components. The Reciprocal Coordinate Base Without dwelling on t h e p r o p e r t i e s and examples o f s c a l a r and v e c t o r quant i t i e s , familiar from physics and geometry, we will proceed t o t h e i p arzalytical characterization.

W e will apply t h e term a b s o l u t e s c a l a r t o a quantity, determined by a /24 function of t h e coordinates o f t h e p o i n t s in space, whose value a t a f h e d point i n space does not depend on t h e choice o f t h e coordinate system, I n what f o l l o w s we will a l s o term such a q u a n t i t y an i n v a r i a n t of coordinate t r a n s f o r mations. The space in which a function d e f i n i n g an a b s o l u t e scalar 5s assigned

i s c a l l e d a s c a l a r f i e l d . I n a d d i t i o n t o a b s o l u t e s c a l a r s t h e r e a r e a l s o scal a r q u a n t i t i e s t h a t do depend on t h e choice of t h e coordinate s y s t e m . The proj e c t i o n s of d i r e c t e d segments on t h e coordinate axes are examples of such quant i t i e s . When t h e term "scalar" i s h e r e a f t e r used, it will refer only t o absol u t e scalars.
Consider now a c e r t a i n v e c t o r of a c u r v i l i n e a r coordinate s y s t e m .

z,Arse fiesr rgenerally e d t o t h e l o c a l coordinate 2ase si known, t h e v e c t o r a may be

n = e#.

represented by t h e expansion (B i b l . '7 )

-.

The q u a n t i t i e s a' are c a l l e d t h e contravariant components of t h e v e c t o r 2. The meaning of this term will be explained below. I n t h e general case t h e quantit i e s ai a r e f u n c t i o n s of t h e coordinates of t h e p o i n t s i n space. The space i n which t h e functions ai are assigned i s c a l l e d t h e f i e l d of t h e v e c t o r 3.

To e s t a b l i s h t h e a n a l y t i c d e f i n i t i o n of t h e v e c t o r 2 , consider t h e p o i n t transformation of t h e coordinates xi and t h e change i n t h e q u a n t i t i e s a' assoLet t h e formulas of t r a n s i t i o n from t h e coorc i a t e d w i t h t h i s transformation. d i n a t e s xi t o t h e new coordinates y' and from t h e new coordinates t o t h e o l d be of t h e following form:

Then t h e r a d i u s v e c t o r of an a r b i t r a r y point M ( y ' ) may b e regarded as a complex function o f t h e xi B y v i r t u e of eq. (2.3 ) w e obtain

where t h e e t J a r e t h e v e c t o r s o f t h e new coordinate base. Equations (5.3a) are t h e formulas of transformation of t h e coordinate base. The formulas f o r t h e j25 inverse transformation can be s i m i l a r l y found:

15

(5.4)
where

R p a t i o n s (5.5) a r e t h e formulas of transformation o f t h e q u a n t i t i e s a*. I n e x a c t l y t h e same way w e may f i n d t h e formulas of t h e transformation i n v e r s e we conclude t h a t t h e Comparing t h e r e l a t i o n s (5.3a) and (5.5), t o eq.(5.5). formulas of transformation of t h e q u a n t i t i e s ai are, i n v e r s e i n sense t o t h e f o m u l a s f o r t h e transformation of t h e v e c t o r s of t h e coordinate base. Hence t h e term frcontravariantft. Vectors can a l s o be defined by a system of "generalized projections" onto t h e axes of t h e l o c a l coordinate base. Consider t h e t h r e e q u a n t i t i e s
+ *

a, = a . e i .
These q u a n t i t i e s , likewise, a n a l y t i c a l l y determine t h e vector 2. To convince ourselves t h a t this i s so, it i s s u f f i c i e n t t o exprezs, i n terms of t h e q u a n t i t i e s +, t h e c o n t r a v a r i a n t components of t h e v e c t o r a. By v i r t u e of eq~~(5.1) and (2.5),
* *

we obtain*
(i, k = l , 2, 3).
(5.7)

a.=e..e 1 r k a k = g . I & aR

ak we f i n d t h a t
'

Considering t h e s e equations as a system of ' l i n e a r a l g e b r a i c equations i n

ak---g

ik ai

(i, I< = 1, 2, 3).

(5.8)
where

* Here

and h e r e a f t e r we make use of t h e r i g h t t o denote t h e dummy indexes by

any d e s i r e d l e t t e r .

16

In orthogonal systems of coordinates:

After i n t r o d u c t i o n o f t h e q u a n t i t i e s

,$k,

eq.(5.1)

assumes t h e form

a = e,gika,.
W e now introduce t h e n o t a t i o n

- -

( 5 . 1 0 )
Then eq.(a) t a k e s t h e form

cz= &a,.

-.-

(5.11)

The v e c t o r s

zi

form a coordinate base r e c i p r o c a l t o t h e o r i g i n a l b a s i s .

Let u s consider a few r e l a t i o n s , necessary f o r t h e f u r t h e r discussion, between t h e q u a n t i t i e s introduced here. Compare t h e systems o f equations (5.7) and (5.8). W e have

Since eqs. (b,) a r e v a l i d a t a r b i t r a r y values of t h e q u a n t i t i e s ak (k, j = 1,2,3), t h e following i d e n t i t i e s hold:

g'fig,

ak. 5

1 0

(k = j ) ,

(k # I ) .

(5.12)

where 6k i s t h e Kronecker symbol. we f i n d

Further, from eqs.(5.10),

(2.5),

and ( 5 . 1 2 ) ,

(5.13)
FQuation (5.13) permits us t o c a l l t h e q u a n t i t i e s g " c o e f f i c i e n t s of t h e

17
I .

fundamental q u a d r a t i c form on a r e c i p r o c a l coordinate base. same way we o b t a i n

rn exactly the

Obviously,

gf, ==gi:

..

Equations(5.14) permit u s t o f i n d simple expressions of t h e v e c t o r s of a r e c i p r o c a l coordinate base. W e obtain

where

i s t h e sign of t h e v e c t o r product.

The i n d i c e s i, j, k form a c y c l i c permutation o f t h e numbers 1, 2, 3. Similarly,

ei = V ( d

E').

(5.1%) from eqs.(5.l&} i s l e f t t o t h e m

The d e r i v a t i o n of eqs.(5.15a) reader.

+.

and (5.1%)

W e r e t u r n now t o eq. (5.1). On s c a l a r m u l t i p l i c a t i o n of this equation by eJ and bearing eq.(5.1&) in mind, w e obtain

(5.16)
Consider t h e formulas of transformation of t h e q u a n t i t i e s e obtain b a s i s of eqs.(5.3a) and (5.6) w

a,. OB

the

(5.17)

The formulas of t h e transformation inverse t o eqs. (5.17) may be s i m i l a r l y found. These formull-as allow u s t o c a l l t h e q u a n t i t i e s a, t h e covariant compone n t s of t h e v e c t o r a, s i n c e they coincide with t h e formulas of transformation of t h e coonlinate vectors.

18

Conclusions
A s c a l a r i s a geometrical o r physical q u a n t i t y t h a t i s determined by a s i n g l e function of a p o i n t i n space and does n o t change i t s value a t a fixed p o i n t under transformation of coordinates. A v e c t o r i s a geometrical o r physical q u a n t i t y determined by a s y s t e m of These f u n c t i o n s obey t h e t h r e e functions according t o eqs.(5.l) and (5.11). A c h a r a c t e r i s t i c f e a t u r e of t h e s e transformation formulas (5.5) and (5.17). formulas i s t h e i r l i n e a r i t y and homogeneity r e l a t i v e t o t h e c o e f f i c i e n t s o f .transformation (eq,.5,2a) :
a?=-.

ayj
axi

'

(5.18)

The transformation formulas corresponding t o eqs.(5,2b), as i s r e a d i l y demonstrated, a r e l i n e a r and homogeneous r e l a t i v e t o t h e c o e f f i c i e n t s of t h e transf orma.tiou;t

The transformation formulas are a l s o homogeneous with r e s p e c t t o t h e vect o r components. For this reason, a v e c t o r equal t o zero i n one s y s t e m o f coord i n a t e s i s equal t o zero i n a l l coordinate systems. I n concluding t h i s Section, l e t u s consider t h e p r o j e c t i o n s of a v e c t o r onto t h e axes of t h e l o c a l coordinate base. These p r o j e c t i o n s a r e sometimes c a l l e d "physical components of a vector" (Bibl.8).
4

/28

Denoting t h e modulus of t h e v e c t o r e, by e,, we have

I n orthogonal coordinate s y s t e m s :
n.,-I . = 0.'

lfG

( 5 . 2 1 )

(do not s u m over i 4 )

~

: T he expressions f o r t h e transformation c o e f f i c i e n t s adopted here are inverse t o those used by us i n an earlier book (Bibl.7).

19

Section

6. Tensors of Various&&
Tensor o f t h e S h e_ l- l

and St-ructure. The Metric

The concept of t h e t e n s o r i s a n a t u r a l g e n e r a l i z a t i o n of t h e concepts of s c a l a r and v e c t o r discussed above. The b a s i s of this g e n e r a l i z a t i o n i s given by t h e formulas f o r transformation of v e c t o r components, eq~~(5.5) and (5.17). To f i n d t h e d i r e c t i o n of t h e g e n e r a l i z a t i o n s , l e t u s consider t h e formulas of bransfomation of t h e q u a n t i t i e s gik, g and g ! ; On t h e b a s i s of t h e d e f i n i t i o n s of t h e s e q u a n t i t i e s and o f t h e transformation fonrmlas f o r t h e v e c t o r s of t h e p r i n c i p a l and r e c i p r o c a l coordinate bases, we have

or, finally,

Similarly,
g l i k = ai ak uj!. J 1s

A comparison of e q ~ ~ ( 6 . 1 and ) (6.2) with e q ~ ~ ( 5 . 5and ) (5.17) l e a d s t o t h e wanted generalization.

A comparison of eqs.(6.1) and (6.2) with t h e v e c t o r component transformat i o n f o d a s (5.5) and (5.17) and A t h - t h e p r o p e r t i e s of s c a l a r s permits t h e i n c l u s i o n of s c a l a r s , vectors, and of t h e q u a n t i t i e s g i k , g*k-, gf; among t h e t e n s o r q u a n t i t i e s (tensors);;-.

The s t r u c t u r e of t h e r e l a t i o n s (6.1) and (6.2) shows t h a t we must d i s t i n guish t e n s o r s w i t h covariant, contravAriant, and mixed components. The d i f f e r ence between t h e s e components i s t h a t t h e formulas of t r a n s i t i o n from t h e old covariant components t o t h e new contain o n l y t h e c o e f f i c i e n t s S i , t h e f o m a s of t r a n s i t i o n from t h e o l d c o n t r a v a r i a n t components t o t h e new contain o n l y t h e c o e f f i c i e n t s ai, while t h e formulas of t r a n s i t i o n from t h e old mixed compon- /29 e n t s t o t h e new contain t h e transformation c o e f f i c i e n t s cy: and 8:

Further comparison of eqs.(6.1) and (6.2) with t h e r e l a t i o n s (5.5) and (5.17) permits i n t r o d u c t i o n of t h e concept of t h e rank of a tensor. The rank of a t e n s o r i s equal to t h e dimensionality of t h e right-hand s i d e s of t h e t r a n s formational equations f o r i t s components r e l a t i v e t o t h e c o e f f i c i e n t s of t r a n s formati o n

x- The term t*tensorlt apparently o r i g i n a t e d i n connection w i t h t h e f a c t t h a t t h e s t r e s s e s in t h e neighborhood of a c e r t a i n p o i n t of a continuous medium a r e components of t h e s t r e s s tensor. It i s connected with t h e Latin word tendere, t o pull, t o s t r e t c h .

20

The number of components of a t e n s o r depends OF i t s rank. components of a t e n s o r o f rank n i s expressed by t h e formula

The number N of

N = 3*.
Thus s c a l a r s a r e t e n s o r s of zero rank, v e c t o r s a r e t e n s o r s of f i r s t rank, and t h e q u a n t i t i e s gik, g i k , &; a r e components of t e n s o r s of second rank. Tens o r s possessing t h e components &k, g 1 k and gfi a r e c a e d metric, s i n c e they detelrnine t h e measurement of d i s t a n c e s between p o i n t s of space and t h e measurement of a n g l e s between d i r e c t e d segments, i.e., t h e m e t r i c of space.

( 5 . 1 7 ) , (6.1) and (6.2), we s e t up a transformaGeneralizing eqs.(5.5), t i o n forrrmla f o r t h e components of a t e n s o r of a r b i t r a r y s t r u c t u r e and rank. These formulas are of t h e following form:

W e advise t h e r"eader t o s e t up t h e formula of t h e transformation i n v e r s e t o eq.(6.3), as an exercise. E p a t i o n s (6.3) express t h e fundamental property of t e n s o r components, t h e law of t h e i r transformation on passage from one coordinate system t o another. This l a w i s t h e same f o r a l l tensors, r e g a r d l e s s of t h e i r geometrical meaning o r physical nature. For t h i s reason, t o prove t h a t any q u a n t i t y has t e n s o r p r o p e r t i e s , it i s necessary and s u f f i c i e n t t o prove t h a t t h e transformation formulas (6-3) a r e s a t i s f i e d . It follows from eqs.(6.3), in p a r t i c u l a r , t h a t a t e n s o r with components equal t o zero i n a c e r t a i n system of coordinates will In general, every have components equal t o zero i n a l l systems of coordinates. t e n s o r equation t h a t i s v a l i d i n one coordinate system w i l l be s a t i s f i e d i n a l l o t h e r systems, i.e., such an equation w i l l be i n v a r i a n t under transformation of coordinates

If t h e components of a t e n s o r i n one system of coordinates a r e known, then permit u s t o f i n d i t s components i n any o t h e r system. In t h i s e q ~ ~ ( 6 . 3will ) case, t h e formulas of coordinate transformation (5.2a) and (5.2b) o r t h e coeff i c i e n t s of transformation CY^ and f3: m u s t be assigned.
Thus, f o r example, we found t h e expressions (3.6a) and (3.6b) f o r t h e components of t h e metric t e n s o r i n t h e s h e l l , under t h e assumption t h a t t h e coord i n a t e l i n e s on t h e base surface coincide with i t s l i n e s of curvature. These expressions permit a determination of t h e components of t h e metric t e n s o r i n /30 an a r b i t r a r y system of coordinates of t h e b a s i c surface. Let t h e formulas of coordinate transformation be of t h e following form:
x ~ = s i ( y i ) (i, x3 = y3.

i=1,

2),

Then, from eqsO(3.6a), (3.6b) and (6.l),

we o b t a i n

2 1

I n e x a c t l y t h e same way we could i n d i c a t e t h e formulas o f transformation of t h e components of t h e metric tensor, corresponding t o an e n t i r e l y arbitrary choice of t h e system o f coordinates i n t h e s h e l l . However, we will n o t present them here

Section

7. Operations of Tensor Algebra

Tensor algebra considers only those operations on t e n s o r s which r e % l t again i n a tensor. It goes without saying t h a t t h e s e operations do not include operations connected with d i f f e r e n t i a t i o n or i n t e g r a t i o n .

1 . Addition
The operation of a d d i t i o n can be performed only on t e n s o r s o f t h e same rank and s t r u c t u r e . The s w of t e n s o r s i s t h e t e n s o r determined by components equal t o t h e sums of t h e components of t h e t e n s o r s being added:

Indeed, i f t h e q u a n t i t i e s A!: j : Bt'f j : Ct'f j : are t e n s o r components, i.e., if they obey t h e transformation formulas f6.3), then, obviously, t h e q u a n t i t i e s This demonstrates t h a t t h e Ti. k 4* . * a l s o obey t h e transformation formulas (6.3). operations defined by eq. (7.1) belong t o t e n s o r algebra.

2. M u l t i p l i c a t i o n
The operation of m u l t i p l i c a t i o n ma7 be applied t o t e n s o r s of arbitrary rank and s t r u c t u r e . The product of t e n s o r s i s t h e t e n s o r with components equal t o t h e prod- /31 u c t s of t h e components of t h e t e n s o r s being multiplied. The rank of t h e product
22

equals t h e sum of t h e ranlts o f t h e f a c t o r s ,

For example:

If eqs.(6,3) are s a t i s f i e d for* t h e q u a n t i t i e s A l k , BJ then it i s obThis demonv i o u s t h a t t h e y w i l l a l s o be satisfied f o r t h e q u a n t i t i e s T ! J - strates t h a t t h e operations defined by t h e r e l a t i o n (7.2) belong t o t h e operat i o n s of t e n s o r algebra.

,..., :.

An example o f t h e a p p l i c a t i o n of t e n s o r m u l t i p l i c a t i o n i s t h e construction o f elementary, so-called m u l t i p l i c a t i v e tensors. Assume, f o r instance, t h a t we have assigned t h e v e c t o r s a t 9 b l , c , ' . ? . The products of t h e s e q u a r h i t i e s a r e t h e components of t h e mixed m u l t i p l i c a t i v e t e n s o r of t h i r d rank:
Tij. ..k-- (L' WC,.

(7.3

3. Contraction
The operation of c o n t r a c t i o n can b e performed o n l y on mixed tensors.

To perform t h i s operation on the.mixed t e n s o r T!tj:

ties

w e s e t up t h e quanti-

\4e s h a l l prove t h a t t h e q u a n t i t i e s Tf::* a r e components of a t e n s o r having a rank two u n i t s lower than t h e rank of t h e o r i g i n a l tensor, T t t j

Consider t h e transformation formula (6.3). t h e right-hand s i d e of t h e equation, t h e s u m

Put i n t h i s formula k = j. I n

will be eliminated.

Thus, we obtain

Conseqcently, t h e operation o f contraction l e a d s t o a t e n s o r of rank two units lower t h a n t h e rank of t h e o r i g i n a l tensor.

%

Here and h e r e a f t e r t h e set of components of a t e n s o r is, f o r b r e v i t y , i t s e l f c a l l e d a tensor.

23

Ekample. Consider t h e mixed m u l t i p l i c a t i v e t e n s o r ai$. By performing c o n t r a c t i o n on it, we o b t a i n t h e s c a l a r - s c a l a r product of t h e t h e oper2tion v e c t o r s a and b:

zf

4.

ttF&ising" and "Lowering" of Indices

W e have already encountered s p e c i a l cases of this operation i n d i s c u s s i n g Let u s extend it t o t e n s o r s of any rank and t h e r e l a t i o n s (5.7) and ( 5 . 8 ) . structure

W e s h a l l first show t h a t an a r b i t r a r y t e n s o r can be represented as t h e sum o f m u l t i p l i c a t i v e tensors. It i s s u f f i c i e n t t o demonstrate this i n any spec i a l l y s e l e c t e d coordinate system. Consider, t o b e d e f i n i t e , t h e third-rank t e n s o r T i j o . Let u s set up, corresponding t o each component of this tensor, a
.k

system of t h r e e v e c t o r s apbqc,. Let u s s e l e c t t h e v e c t o r s of this system, f o r instance, as follows:

(IP = a p k ,
0'1

[i dkl

a? , cr === 6; .

Then t h e t e n s o r T:!;

may be represented by t h e sum

Bearing eq.(5.7)

i n mind, l e t u s consider t h e equation

Thus, (7.7) W e have "lowered" t h e first contravariant index, by converting it i n t o a covariant index. A covariant index may s i m i l a r l y be "raised"

Consequently, any system of t e n s o r components may be determined i n a space with a given metric tensor, i f a system of components of any s t r u c t u r e i s
kn0Wl-I.

24

1
I I

W e note in conclusion t h a t t h e representation (b) of an a r b i t r a r y t e n s o r For example, by a sum of m u l t i p l i c a t i v e t e n s o r s has a number of applications. t h i s r e p r e s e n t a t i o n permits us d i r e c t l y t o f i n d t h e '?physical components" of an a r b i t r a r y t e n s o r by making u s e of eqs.(5.20) and (5.21). For this it i s suffi-

Pq4
c i e n t t o s u b s t i t u t e f o r t h e components of t h e v e c t o r s a*bJ4 t h e i r p r o j e c t i o n s onto t h e axes of t h e l o c a l coordinate b a s i s .

5. Permutation of Indices. Symmetrization and Alternation

The interchange of any p a i r of i n d i c e s i n t h e components of t h e t e n s o r T t T j transforms this t e n s o r back i n t o a tensor. If, on interchange of a p a i r of indices, t h e t e n s o r remains unchanged, it i s c a l l e d symmetric with respect t o this p a i r of indices. For example, on s a t i s f y i n g t h e condition

/33

(7.9)
t h e t e n s o r T t l j i s c a l l e d symmetric with respect t o t h e i n d i c e s k and j. If, on interchange of a p a i r of indices, t h e components of t h e t e n s o r change t h e i r signs, then t h e t e n s o r i s c a l l e d antisymmetric with r e s p e c t t o this p a i r of i n d i c e s . For example, on s a t i s f y i n g t h e condition

t h e t e n s o r TtT;

i s antisymmetric with respect t o t h e i n d i c e s k and j.

Making use o f t h e transformation f o m l a s (6.3), it i s easy t o show t h a t t h e p r o p e r t i e s of symmetry and antisymmetry are i n v a r i a n t under coordinate transformations (Bibl.7). The proof i s l e f t t o t h e reader as an exercise.
A symmetric t e n s o r of second rank has six s u b s t a n t i a l l y d i f f e r e n t compone n t s i n three-dimensional space, while an antisymmetric t e n s o r of second rank has only three. Indeed, w e have, i d e n t i c a l l y ,

The formation of a doubled symmetric p a r t of a t e n s o r i s c a l l e d symmetrization, and t h a t of a doubled antisymmetric p a r t i s c a l l e d a l t e r n a t i o n .

25

Section 8. Various Applications of -~ T ensor Algebra -

1 . The Second Analytic D e f i n i t i o n of t h e Tensor --__.

~

,W e s h a l l prove t h e following theorem: Given t h e system of q u a n t i t i e s Tft j I f t h e sum T:rj a, 4 c j i s an i n v a r i a n t a n d ,t h e a r b i t r a r y v e c t o r s a, 4, cj under transformation o f coordinates ( t h a t is, an absolute s c a l a r ) , then t h e , q u a n t i t i e s Tf ; are components o f a mixed t e n s o r of t h i n l rank,

L e t u s make use of t h e v e c t o r component transformation formulas t h a t r e s u l t f r o m eqs.(6.3). W e have

S u b s t i t u t i n g t h e s e r e l a t i o n s i n t o eq.(a) and transposing all terms t o t h e l e f t s i d e o f t h e equ-ation, we obtain

cfr. ish.

Equation ( c ) hold-s f o r a r b i t r a r y values o f t h e q u a n t i t i e s a t p , b f , and This i s p o s s i b l e only i f a l l c o e f f i c i e n t s of t h e products af,bf,cTrvanW e then ob%&

W e have again a r r i v e d a t a r e l a t i o n of t h e form o f eqs.(6,3), T h i s proof may obviously be extended t o a t e n s o r of a r b i t r a q rank and structure.

2. The Antisymmetric Tensor of- Rank Two as a Vector i n Three-Dimensional Space

W e have already noted t h a t an antisymmetric t e n s o r of rank two i n threedimensional space has t h r e e s u b s t a n t i a l l y d i f f e r e n t components. W e s h a l l now show t h a t t h e r e exists a v e c t o r equivalent t o this tensor,

L e t us first consider t h e transformation formulas f o r t h e v e c t o r s o f t h e (5.3b) and (6.2), w e find r e c i p r o c a l coordinate base, From eqs..(5.10),
eri = a i eq.
4

( 8 . 1 )

3 y using t h e transformation formulas

(5.3b) we o b t a i n

Comparing t h e r e l a t i o n s (8.1),

( e ) and

( f ) ,

we obt4.11

where t h e i n d i c e s j , i, k and q, r, s t a k e t h e values 1, 2, 3 i n t h e o r d e r of/35 a p o s i t i v e c y c l i c permutation. Consider t h e transformation formulas f o r t h e components of an antisymIt follows from eqs.(6.3) and t h e a n t i m e t r i c covariant t e n s o r of rank two. symmetry of t h e t e n s o r t h a t

or, from eq.(8.2),

The sign of summation i n t h e right-hand s i d e extends over a l l pairwise combinat i o n s of t h e t h r e e numbers 1, 2, 3, corresponding t o t h e i n d i c e s r and S .
W e introduce t h e notation:

where t h e symbols j , i, k form t h e p o s i t i v e c y c l i c permutation of t h e numbers 1,

27

2, 3 .

Remembering eq.(2,6b),

we o b t a i n from ( h )

This r e l a t i o n shows - t h a t eqs. ( 8 . 3 ) determine t h e components of t h e contravaria y a n t v e c t o r equivalent t o t h e covariant antisymmetric t e n s o r of rank two. It m be shown similarly t h a t a contravariant a n t i s m e t r i c t e n s o r o f rank two i s equivalent t o a covariant v e c t o r with t h e components

3. The Vector Product of Two Vectors i n an __ A r b i t r a r y

Coordinate Svstem Consider t h e m u l t i p l i c a t i v e t e n s o r of rank two:

Performing t h e operation of a l t e r n a t i o n , w e obtain

The t e n s o r tor

%,,,

according t o eqs.(8.3),

i s equivalent t o t h e contravariant vec-

The vector cJ dete+pjnes+the contravariant components of t h e v e c t o r /36 product of t h e v e c t o r s a and b i n an a r b i t r a r y coordinate system. Similarly, from eqs. (8.4), we obtain:

4. Pseudoscalars and-Pseudovectors
Let us r e v e r t t o eqm(2.6b). The volume V of t h e parallelepiped constructed on t h e v e c t o r s of t h e coordinate basis i s a scalar. But i t i s impossible t o f i n d this s c a l a r as an absolute. Under coordinate transforanation, t h e volume V varies. I n p a r t i c u l a r , on passage of t h e local coordinate b a s i s from a r i g h t hand system of coordinate v e c t p s t o a left-hand s y s t e m , under preservation of t h e q u a n t i t i e s of t h e v e c t o r s e l , t h e volume V changes sign. The volume V i s The q u a n t i t i e s R i k , defined by e q s . ( l ) have t h e r e f o r e c a l l e d a pseudoscalar. 28

similar p r o p e r t i e s ( f o r detai.ls see Bibl.7). with t h e components expressed by eqs.(8.5) and (8.6), h a s a The vector dual meaning. If t h e s i g n f i i s fixed, then t h e components of t h e vector c will change signs on passage of t h e l o c a l coordinate system from a-right-hand t o a left-hand s y s t e m . I n t h i s case t h e componen4s of t h e vector c do not obey t h e transformation formulas ( 6 . 3 ) and t h e vector c i s c a l l e d a pseudovector.
If, however, .E i s regarded as a pseudoscalar, then eqs. (8.5) and (8.6) determine a polar vector, i.e. a vector t h a t does obey t h e transformation l a w (6.3 )*

W e note i n conclusion t h a t a vector product e x i s t s a s a vector only i n I n rrmlti-dimensional space it i s considered an a n t i three-dimensional space, syrmnetric tensor of rank two instead of a vector.

Section 9. The Absolute D i f f e r e n t i a l of a Tensor. The Tensor Field and t h e Absolute Derivative

1 . The Absolute D i f f e r e n t i a l of a Tensor

4

Consider t h e v a r i a b l e vector a w i t h contravariant components:

Assuming t h a t t h e components of %he vector and t h e p o i n t s of i t s application vary, w e find t h e d i f f e r e n t i a l da:

dil = eidai

/37

-. -.

-. + aide,.

Let u s f i n d t h e contravariant components of t h e d i f f e r e n t i a l da. eqs.(5.16) we have

f

From

(da)i = daj

- del, +aiel
a

Further,
d e 1 -- --'dxk.
aXk

de.

Using eqs.(2.3)

we now f i n d

W e introduce t h e notation 29

The q u a n t i t i e s r i k are c a l l e d C h r i s t o f f e l symbols of t h e second kind. C h r i s t o f f e l symbols are symmetric with r e s p e c t t o t h e i n d i c e s i, k:

Equation ( c ) now t a k e s t h e form:

Quation 19.3) determines+the cantravariant components o f t h e a b s o l u t e d i f f e r e n t i a l da of t h e v e c t o r a. The term ttabsolute d i f f e r e n t i a l " e v i d e n t l y arose in connection w i t h t h e i d e a s o f absolute motion, which are familiar from kinematics. Consider t h e covariant components o f t h e absolute d i f f e r e n t i a l . eq. (5.11) we have

From

-f

L e t . u s f i n d t h e covariant components of t h e v e c t o r da:

I t follows from eqs.(5.14)

that:

Making use of t h e r e l a t i o n s (d), (9.1) and ( h ) , we o b t a i n from eq.(g):

/3Ef

This r e l a t i o n d.e>ermines t h e covariant components of t h e absolute d i f f e r e n t i d l of t h e v e c t o r a.

bde will now show t h a t t h e C h r i s t o f f e l symbols a r e defined i n terns of t h e components of t h e metric tensor, and i n d i c a t e t h e formulas f o r t h e i r t r a n s f o r matior,. From eqs.(5.10) and (9.1) w e have

(9.5)

The q u a n t i t i e s

are c a l l e d C h r i s t o f f e l symbols of t h e f i r s t kind. and (5.12) t h a t

It follows from e q ~ ~ ( 9 . 5 )

The following formula of transformation of t h e C h r i s t o f f e l symbols of t h e first kind results from eq.(i):

where t h e

are t h e coordinates of t h e new system.

From t h e transformation formula (9.7a) and eq.(9.6) it i s easy t o d e r i v e t h e transformation formula f o r C h r i s t o f f e l symbols of t h e second kind:

Thus, t h e C h r i s t o f f e l symbols a r e not components of a t e n s c r , s i n c e t h e y do not obey t h e transformation formulas (6.3). Making use of t h e r e l a t i o n (e), we f i n d

or

31

Again making use of t h e r e l a t i o n ( e ) and eq.(2.5),

we o b t a i n

L29.
(9.8 1

It follows from eq.(9.8)

that

The absolute d i f f e r e n t i a l of a v e c t o r i s t h u s completely determined i f t h e metric of t h e space i s known. I n conclusion we note t h e existence of a d i r e c t r e l a t i o n between t h e absol u t e 2 i f f e r e n t i a l of a v e c t o r and t h e absolute d e r i v a t i v e of t h e v e c t o r funct i o n a ( t ) , which we know from t h e p r i n c i p l e s of t h e kinematics of a r i g i d body:-:

where

dfi dt is

t h e r e l a t i v e ( l o c a l ) d e r i v a t i v e of t h e v e c t o r

2.

This i n t e r r e l a t i o n results-from eqs.(9.1) and from t h e d e f i n i t i o n of t h e instantaneous angular v e l o c i t y w of t h e body.. I n this special. case we f i n d t h a t between t h e C h r i s t o f f e l symbols and t h e instantaneous angular v e l o c i t y of an a b s o l u t e l y r i g i d body t h e r e exists t h e relation:

where t h e w !; are t h e components of t h e antisymmetric t z n s o r of instantaneous angular v e l o c i t y of t h e body, equivalent t o t h e v e c t o r w (8.2)

A d i f f e r e n t i n t e r p r e t a t i o n of t h e meaning of t h e C h r i s t o f f e l symbols i s It follows from r e l a t i o n ( 9 . 9 ~ ) t h a t t h e product a l s o possible.

32

determines t h e generalized r e l a t i v e angle of r o t a t i o n of t h e coordinate b a s i s under displacement from t h e p o i n t M ( x ' ) t o t h e neighboring p o i n t MT (xi + dxi ).

2. Absolute D i f f e r e n t i a l o f a Tensor of A r b i t r a r y Rank and S t r u c t u r e

Consider t h e i n v a r i a n t

Differentiating the invariant

CD,

we o b t a i n

O n t h e b a s i s of eqs.(9.3)

(9.4) we r e p r e s e n t eq.(d) i n t h e following form:

Here w e have changed t h e dunmy i n d i c e s necessary f o r t h e transformation of e q . (.e>. Considering eq,(m), w e note t h a t i t s l e f t s i d e and t h e f i r s t summands in i t s r i g h t s i d e a r e s c a l a r s , Consequently, t h e l a s t term in i t s r i g h t s i d e i s a l s o a s c a l a r . But t h e q u a n t i t i e s q , h, cj a r e components o f a r b i t r a l y vect o r s . Consequently, according t o t h e second a n a l y t i c d e f i n i t i o n of a t e n s o r (Sect.8), t h e expressions i n parentheses a r e components o f a mixed t e n s o r :

The t e n s o r DfT" determined by eqs.(9.10) e n t i a 3 of t h e t e n s o r T *t .k .j:

'

i s c d l e d t h e absolute d i f f e r -

33

3* Tensor Field. The Absolute (Covariant) Derivative of a Tensor

o f A r b i t r a r y Rank and S t r u c t u r e

A t e n s o r f i e l d i s a region o f v a r i a t i o n of t h e coordinates x ' , such t h a t t o each p o i n t o f t h e region t h e r e correspond values of t h e components of some tensor. W e s h a l l assume, with infrequent exceptions, t h a t t h e components of t h e t e n s o r are single-valued functions of t h e coordinates of t h e p o i n t s of t h e e s h a l l a l s o assume t h a t t h e s e f u n c t i o n s have a n a l y t i c s i n g u l a r i t i e s field. Y a t i s o l a t e d p o i n t s of t h e f i e l d . A t all o t h e r p o i n t s of t h e f i e l d t h e t e n s o r components a r e continuous and d i f f e r e n t i a b l e functions of t h e coordinates xi. Then,

Equation (9.10) now t a k e s t h e form:

i s c a l l e d t h e a b s o l u t e or covariant d e r i v a t i v e o f t h e t e n s o r T f ; ; : . The geom e t r i c a l meaning of absolute d i f f e r e n t i a t i o n will be discussed i n t h e following Section.

Let u s r e t u r n t o eq.(9.9a). lowing form: This r e l a t i o n may be represented i n t h e fol-

Equation (9.13) expresses t h e theorem of Ricci: The absolute d e r i v a t i v e of t h e metric t e n s o r vanishes. This a s s e r t i o n a l s o a p p l i e s t o t h e contravariant and mixed components of t h e metric t e n s o r (Bibl.7). Consequently, i n covariant d i f f e r e n t i a t i o n t h e components of t h e metric t e n s o r must be regarded as constant q u a n t i t i e s . W e suggest t h a t t h e reader convince himself t h a t t h e well-known-rules f o r d i f f e r e n t i a t i o n of t h e sum o r produ c t of s c a l a r functions apply t o t h e absolute d i f f e r e n t i a t i o n of t e n s o r functions.

3L.

S e c t i o n 10. P a r a l l e l Displacement of-Tensors i n t h e Sense of Levi-Civita. The Tensor of Curvature,

1 . P a r a l l e__ l Displacement I n t h e i n v e s t i g a t i o n of various v e c t o r s o f geometry and mechanics, it is necessary t o compare t e n s o r q u a n t i t i e s a n a l y t i c a l l y assigned a t d i f f e r e n t p o i n t s of space. This comparison can b e accomplished after reduction of t h e q u a n t i t i e s t o be compared t o a s i n g l e point, W e meet such reductions, i n p a r t i c u l a r , i n t h e kinematics and s t a t i c s of an a b s o l u t e l y r i g i d body, where t h e system of s l i d i n g v e c t o r s i s displaced p a r a l l e l t o t h e i r i n i t i a l r e c t i l i n e ; z r . b a s e s t o t h e c e n t e r of reduction, I n parallel displacement of a vector, n e i t h e r i t s magnitude(m0dulus) nor i t s d i r e c t i o n are changed. Consequently, i n p a r a l l e l displacement of a v e c t o r from t h e p o i n t X(xi ) t o t h e neighboring p o i n t 'Elf (2 + dx' ) , t h e &sol u t e d i f f e r e n t i a l of t h e v e c t o r mst vanish. Let us adopt t h e above statement as a general d e f i n i t i o n of p a r a l l e l displacement of t e n s o r q u a n t i t i e s . This d e f i n i t i o n o f p a r a l l e l displacement coincides i n essence with t h e d e f i n i t i o n given by Levi-Civita (Bibl.8). Let 6 q : j : be t h e change i n t h e components of a t e n s o r under p a r a l l e l d i s placement from the p o i n t N(xl ) t o t h e neighboring point MT(x' + dx' ). Then, on t h e basis of eqs,(9.10), ) . t can s e t up t h e system of d i - f f e r e n t i a l equations of p a r a l l e l displa-t. This system h a s t h e following form: /lr2

( 1 0 . 1 )
I n p a r t i c u l a r , f o r a contravariant v e c t o r we find

and f o r a covariant v e c t o r

Ne note i n conclusion t h a t , i n a v e c t o r displacement t h a t i s p a r a l l e l i n t h e Levi-Civita sense, t h e s c a l a r product of t h e v e c t o r s remains unchanged and each of t h e v e c t o r s e n t e r i n g into t h e product may be independently displaced, and then t h e s c a l a r product of t h e v e c t o r s so displaced can be constructed. The proof i s l e f t t o t h e reader.

2. Tensor of Cgrvature (Riemann-Christoffel Tensor)

(10.3) a r e not i n general t o t a l l y d i f f e r e n t i a l equaEquations (10.2) t i o n s , The result o f a p a r a l l e l displacement o f v e c t o r s i n t h e Levi-Civita sense, therefore, depends on t h e shape and p o s i t i o n of t h e curve along which 35

C h i s displacement i s accomplished. P a r a l l e l displacement in Euclidean space i s here an exception. I n this case it i s always p o s s i b l e t o choose a c o o d i n a t e system such t h a t t h e components of t h e m e t r i c t e n s o r are constant and, consequently, t h e C h r i s t o f f e l symbols vanish.
Consider t h e r e s u l t of a displacement of t h e c o n t r a v a r i a n t v e c t o r a' from t h e point M(x1) t o t h e p o i n t W(xl + dxl + 6x1) on noncoinciding curves passing x 1 ). L e t u s c a l c u l a t e t h e componthrough t h e point M t (xl + dx* ) and Ml (2 + 6 e n t s of t h e displaced vector, u s i n g eq.(10.2). Let t h e components of t h e d i s placed v e c t o r a t p o i n t M be a'. Then, t h e components o f t h e p a r a l l e l - d i s p l a c e d v e c t o r a t point 149 w i l l he

The symbol M, here and h e r e a f t e r , denotes t h e v a l u e s of t h e f u n c t i o n s a t point M .

O n f u r t h e r motion t o p o i n t M" we must b e a r i n mind t h e change i n t h e Christ o f f e l symbols, which are f u n c t i o n s of t h e coordinates o f t h e p o i n t s i n space. W e have, a t p o i n t MI?

where t h e Ala1 a r e t h e changes in t h e components of t h e v e c t o r 2 , on passage t o t h e point M" along t h e curve MMfM'I. W e s h a l l neglect third-order i n f i n i t e s - & imals. Consider now t h e r e s u l t of p a r a l l e l displacement o f t h e v e c t o r 2 t o t h e point W 1 along t h e curve MMIMtt. The components of t h e p a r a l l e l - d i s p l a c e d vect o r a t point Ml will be expressed as follows:

where t h e &a1 a r e t h e changes i n t h e v e c t o r components ai on passage t o t h e P along t h e curve MMIMtl. point I

36

If t h e p a r a l l e l displacement of t h e v e c t o r i s accomplished i n a space without i n t e r n a l curvature (in Euclidean space), then ( b ) and ( d ) a r e i d e n t i c a l l y equal. Under p a r a l l e l displacement of t h e v e c t o r i n a non-Euclidean space, f o r i n s t a n c e on a nonplaner surface, ( b ) and ( d ) w i l l not coincide. Cons i d e r t h e vector

S u b t r a c t i n g ( d ) from (e), and making t h e necessary changes i n t h e dummy i n d i c e s , we obtain:

L e t us consider this equation. Noting t h a t i t s l e f t s i d e contains contrav a r i a n t components of t h e vector, we conclude, based on t h e second a n a l y t i c defi n i t i o n o f a t e n s o r (Secte8), t h a t t h e expressions

are mixed components of a t e n s o r of rank four. This t e n s o r i s c a l l e d t h e curv a t u r e tensor, o r t h e Riemann-Christoffel tensor.
I n Euclidean space, t h e curvature t e n s o r i d e n t i c a l l y vanishes. I n fact, i n Euclidean space w e can introduce a Cartesian coordinate system, i n which a l l t h e C h r i s t o f f e l symbols vanish, as t h e components of t h e curvature t e n s o r will then a l s o vanish. However, a t e n s o r t h a t vanishes i n one coordinate system w i l l a l s o vanish i n a l l t h e o t h e r s (Sect-6). Consequently, i n Euclidean space, t h e result of t h e p a r a l l e l displacement of a vector, a n a l y t i c a l l y determined i n a c u r v i l i n e a r coordinate system, i s independent of t h e choice of t h e curve along which t h e p o i n t of a p p l i c a t i o n of t h e vector i s displaced. The vanishing of t h e t e n s o r R;;: i s a condition of i n t e g r a b i l i t y of t h e equations of p a r a l l e l displacement.

i.

The above d i s c u s s i o n a p p l i e s t o t h e p a r a l l e l displacement of t e n s o r s of a r b i t r a r y rank and s t r u c t u r e .

/lr4

Consider t h e elementary p r o p e r t i e s of t h e curvature tensor. It will b e c l e a r from eq.(10.5) t h a t it is antisymmetric i n t h e i n d i c e s k and r. L e t u s f i n d i t s covariant components. W e have

L e t us transform ( f ) .

From eq.(9.9a),

we o b t a i n

37

Performing t h e operation of a l t e r n a t i o n with r e s p e c t to t h e i n d i c e s r and j,we obtain

( 1 0 . 6 )
and, making use of t h e expressions (9.8) for t h e C h r i s t o f f e l symbols, w e get
4

T h i s formula g i v e s u s t h e fundamental p r o p e r t i e s o f t h e curvature tensor, i t s antisymmetry i n t h e i n d i c e s r, j and i , k and i t s symmetry i n t h e indexp a i r s r j and ik. Hence follows, more s p e c i f i c a l l y , t h a t i n a three-dimensional space t h e curvature t e n s o r has only six s u b s t a n t i a l l y d i f f e r e n t components, and i n two-dimensional space (on a nonplanar surface), one.

I n Euclidean space, a s already noted, t h e curvature t e n s o r vanishes. I t s vanishing i s a necessary and s u f f i c i e n t condition f o r t h e p o s s i b i l i t y of i n t r o ducing i n t o a space a system of coordinates with t h e Euclidean metric, i n which t h e components of t h e metric t e n s o r are expressed by t h e equations:

( 1 0 . 8)
W e s h a l l not dwell here on t h e proof o f t h i s a s s e r t i o n , nor on t h e study of t h e various p r o p e r t i e s o f t h e curvature tensor, and r e f e r t h e reader t o t h e specialized manuals+t.

3. Change of t h e Sequence of Operations i n Successive

Absolute D i f f e r e n t i a t i o n

I t can be shown t h a t a change i n t h e sequence of operations of covariant d i f f e r e n t i a t i o n s u b s t a n t i a l l y changes t h e r e s u l t i n cases where t h e curvature t e n s o r does not vanish.
The following e q u a l i t y can be proved by d i r e c t c a l c u l a t i o n :
-:t

Cf f o r instance, Rashevskiy,P.K., Gostekhizdat, 1953.

.,

Riemannian Geometry and Tensor Analysis.

38

( 1 0 . 9 )
Consequently, repeated covariant d i f f e r e n t i a t i o n i s c o m t a t i v e i n Euclidean space,

L. Geometric Construction of a Covariant Derivative

~~~

W e can now convince ourselves t h a t t h e absolute d e r i v a t i v e determines t h e major p a r t of t h e increment of a t e n s o r function, l i k e t h e d e r i v a t i v e t h a t determines t h e rnajor part of t h e increment of .a s c a l a r function. Consider, for instance, t h e components of t h e increment of a contravariant vector correspondi n g t o t h e d i f f e r e n c e between t h e coordinates of p o i n t M ( x i ) and W(xi + d 2 ). To construct t h e v e c t o r increment, having v e c t o r p r o p e r t i e s , a t p o i n t M or e must use t h e operation of p a r a l l e l displacement. Ne have p o i n t N, w

This e q u a l i t y compels a t t e n t i o f f t o the d u a l i t y i n t h e meaning of t h e res u l t : t h e constructed q u a n t i t i e s (Aa); do have t h e p r o p e r t i e s o f a v e c t o r a t p o i n t N, b u t a r e expressed i n terms o f t e n s o r q u a n t i t i e s determined a t point N . Section 1 1 . Operator of P a r a l l e l Displacement of Tensor Q u a n t i t i e s on t h e Base Area of a S h e l l .Nost s t u d i e s on s h e l l theory a r e based on t h e reduction of t h e threedimensional problems o f t h e theory of e l a s t i c i t y and p l a s t i c i t y t o twodimensional problems, by means of t h e a n a l y t i c determination of t h e q u a n t i t i e s sought i n t h e coordinates and metric of t h e base surface,
W e s h a l l t h e r e f o r e now d i s c u s s t h e problem of t h e p a r a l l e l displacement of t e n s o r quaxitities from an a r b i t r a r y p o i n t on a given s h e l l t o t h e base sur- & f face. T h i s displacement may be accomFlished by i n t e g r a t i n g eqs.(lG,l).

Ne s h a l l here consider t h e i n t e g r a t i o n of t h e simpler formulas [eqs.(l0.2) (10,3)], which permit t h e p a r a l l e l displacement of both contravariant and coe s h a l l use t h e method of successive approximation employed v a r i a n t vectors. W by u s elsewhere (Bib1.23b) f o r t h i s purpose.

L e t u s consider again e q ~ ~ ( l 0 . 2f) o r t h e p a r a l l e l displacement of a cont r a v a r i a n t vector, and r e p r e s e n t them i n t h e following form:

Let

US

Given t h e components of t h e v e c t o r a' a t some point of t h e s h e l l . denote t h e s e components by 4. Required, t o f i n d t h e components ai after

N(4)

39

t h e p a r a l l e l displacement of t h e p o i n t of a p p l i c a t i o n of t h e v e c t o r along an a r b i t r a q curve t o t h e base surface. Assume t h a t t h e equations of t h e displacement curve are of t h e form
.vi = .v-(u),

( 1 1 . 1 )

and t h a t

Then, on t h e curve s e l e c t e d by us, t h e C h r i s t o f f e l symbols will be asThen, eqs. ( a ) w i l l t a k e t h e form signed f u n c t i o n s of t h e parameter U.

(11.2)
where

(11.3)
where t h e d o t i n d i c a t e s d i f f e r e n t i a t i o n wit'n r e s p e c t t o U. Following my earl i e r work (Bib1.23b), we replace t h e system o f equations (11.2) by t h e system o f equivalent i n t e g r a l equations:

S u b s t i t u t i n g t h e i n i t i a l values of t h e components o f t h e v e c t o r ai i n t h e expression under t h e s i g n of i n t e g r a t i o n , we o b t a i n t h e first approximation:

S u b s t i t u t i n g , again, t h e first approximation ( c ) i n t o t h e expression under t h e i n t e g r a l sign i n t h e r i g h t s i d e of eq.(11.4), we f i n d

Continuing this process, we obtain, a f t e r s e v e r a l permutations of t h e indices,

where

u u

( 1 1 . 6 )

Equation (11.6) d e f i n e s t h e resolvent of t h e system of i n t e g r a l equations The displacement of a covariant v e c t o r may be s i m i l a r l y considered. W e have, from eqs. ( 1 0 . 3 ) ,

( l l . 4 ) .

da, = N i a, du,

(11.7)

where N;(u)= From eq. (11.7) we f i n d

I'iS(u)Xs(U).

(11.8)

a,(u,[ i o= )

K O

(u) a, (u) du.

(11.9 j

This system of i n t e g r a l equations, l i k e t h e system ( l l . l + ) , i s solved by t h e method o f successive approximation. I t s s o l u t i o n i s of t h e form:

The resolvent Y i (u, u,) i s expressed a s follows:

U
I1

The proof f o r t h e convergence of t h e s e expansions i s known from t h e t h e o r y of V o l t e r r a ' s i n t e g r a l equations of t h e second kind. With i n s u b s t a n t i a l re-

II

11111 III ~ 1 1 1 1 1 1 1 1 1 1 1 1

s t r i c t i o n s I have a l s o presented t h e proof of convergence i n t h e work axready c i t e d (Bib1.23b). (u, u, ) and Ylf (u, u ,), l i k e t h e C h r i s t o f f e l symbols., are The r e s o l v e n t s n o t t e n s o r q u a n t i t i e s . These o p e r a t o r s permit t h e displacement of tensor quane s h a l l c a l l them o p e r a t o r s o f p a r a l l e l d i s t i t i e s over a f i n i t e d i s t a n c e . W placement

The formulas ( 1 1 . 5 ) and (11.10) can b e put i n t o a d i f f e r e n t form. setting

By

, & &

( 1 1 E12 )
we g e t

The r e l a t i o n s (11.5) and ( 1 1 . 1 0 ) w r i t t e n i n this farm are analoggus t o t h e W e therefore vector-component transformation formulas derived from eqs. (6.3). extend eqs.(ll.l3) t o a t e n s o r of a r b i t r a r y rank and s t r u c t u r e . By analogy t o eqs. (6.3) we o b t a i n

Let u s consider, as an example, t h e construction o f t h e o p e r a t o r s $I:(u,~, ) and Yif (u, u, ). L e t t h e metric of t h e s h e l l b e expressed by eqs.(3.6a)-(3.6b), The metric defined by t h e s e equations i s encountered i n an &deformed shell, 0;r we choose a new coordinate i n a deformed s h e l l i f - a f t e r i t s deformation system with t h e coordinate l i n e s coinciding with t h e l i n e s o f curvature on t h e deformed base surface and with t h e normals t o it.

Assume, f o r simplicity, t h a t t h e displacement t a k e s place along a normdl t o t h e base surface. Then, e q s . ( l l . l ) can be p u t i n t o t h e form

L e t u s a l s o put:

I n t h i s case:

then

. .

! i

3 e a r i n g formulas (5.9b), (9.5) and (9.8) i n mind, and c a l c u l a t i n g t h e C h r i s t o f f e l symbols rf3, we f i n d t h a t o n l y t h e synhols ri, do not vanish i n t h e e have selected. W e obtain sy.stem o f coordinates w
1 IT;, . = - - - . d,gii = d 3 In (1 2K;r

k,.r3)

(i = 1: 2),

( 1 1 017 )

or, kn v i e w of eqs.(ll.l5),
1 i3 (u)= ar1 In (1 -.i

kill).

(g)

(11.6),

Making use of eqs,(f) and ( g ) and t h e expression f o r t h e resolvent, we f i n d

L!Q

or, in t h e n o t a t i o n of eqs.(ll.l5)

(11.16),
( i = 1, 2).

-z) '1);i ( x 3 ,z ) = -k , (x3 . 1 -- k i X 3

The remai.ning o p e r a t o r s @* vanish. For x? = 0, eq.(11.17) y i e l d s t h e ope r a t o r of p a r a l l e l displacemeng of a contravariant v e c t o r t o t h e base surface (Bib1.23b ):
i

(9;( 0 , z )

- ki2.

(11.18)

L e t u s now determine t h e operator Yi (u, u , ). Using eqs. ( f ) and ( g ) and t h e expression f o r t h e resolvent, eq.(ll..ll), w e find

or

43

The o t h e r o p e r a t o r s Y ; vanish. For x?= 0, eq.(11.19) g i v e s t h e o p e r a t o r of p a r a l l e l displacement of a covariant v e c t o r t o t h e base surface:

Wi(O, x ) =

I 2 1 2

1 -kiZ

'

(11 2 0 )

Equations (11.17) and ( l l . 1 9 ) can be d i r e c t l y obtained from t h e system of equations (11.2) and ( 1 1 . 7 ) , s i n c e when t h e r e l a t i o n s ( f ) and (g) a r e satisf i e d , t h e system of equations of p a r a l l e l displacement breaks down i n t o i n d i v i dual equations. Section 12. Expansion o f Tensor Functions i n Generalized Taylor S e r i e s

1 . Analytical D e f i n i t i o n of t h e Radius Vector of a Point of S m c e i n C u r v i l i n e a r Coordinates

I n a n a l y t i c geometry t h e term r a d i u s v e c t o r i s customarily applied t o a d i r e c t e d segment drawn from a fixed p o i n t ( t h e o r i g i n of coordinates) t o a point i n space. I n a Cartesian coordinate system, t h e c o n t r a v a r i a n t components of /To t h e r a d i u s v e c t o r of a p o i n t a r e equal t o t h e components of i t s terminus o r t o d i f f e r e n c e s between t h e coordinates of t h e terminus of t h e r a d i u s v e c t o r and those of i t s fixed o r i g i n . Thus, t h e r a d i u s v e c t o r i s a geometrical o b j e c t connected w i t h two p o i n t s i n space, and t h e r e f o r e it i s n o t a v e c t o r a t t a c h ment, defined i n s t e a d by t h e coordinates of i t s point of application. This causes t h e t r o u b l e i n attempts at a n a l y t i c d e f i n i t i o n of t h e r a d i u s .vector i n c u r v i l i n e a r systems of coordinates, s i n c e t h e transformation formula (6.3) rel a t e s t o a f i x e d p o i n t i n space. To avoid misunderstandings, we s h a l l introduce t h e radius-vector i n t o syse shall f i r s t define tems of c u r v i l i n e a r coordinates by means of d e f i n i t i o n . W e t h e r a d i u s v e c t o r i n t h e Cartesian system of coordinates, as j u s t indicated. W s h a l l then d e f i n e i t s c o n t r a v a r i a n t components i n an a r b i t r a r y c u r v i l i n e a r sysW e shall a t tem of coordinates, applying t h e transformation formulas (6.3). t h e same time a l s o d e f i n e t h e transformation c o e f f i c i e n t s a t t h e fixed o r i g i n of t h e r a d i u s vector. Obviously a r a d i u s v e c t o r can e x i s t o n l y i n a space t h a t It does n o t e x i s t in t h e permits i n t r o d u c t i o n of t h e Cartesian coordinates. i n t e r n a l geometry of nonplanar surfaces. Here we can introduce only small r a d i u s v e c t o r s w i t h e r r o r s of t h e second o r d e r of smallness.

2. m a n s i o n of Tensor Functions i n t o Generalized

Taylor S e r i e s The three-dimensional problems of t h e theory of e l a s t i c i t y and p l a s t i c i t y e must a r e reduced t o two-dimensional problems by v a r i o u s methods, among which w mention t h e method given by Cauchy and Poisson i n t h e t h e o r y of plates. This method, based on t h e expansion of t h e required q u a n t i t i e s i n t o Taylor s e r i e s , w i l l be discussed i n Chapter 1 1 1 . Here we s h a l l dwell only on t h e general prop-

44

e r t i e s of such expansions i n t h e space within a s h e l l , r e f e r r e d t o c u r v i l i n e a r coordinates. Let u s f i r s t consider t h e tenso;? of rank n, r e f e r r e d t o t h e Cartesian system of coordinates. Expanding t h e components of this t e n s o r in Taylor ser i e s i n t h e neighborhood of some fixed point M, i n powers of t h e coordinate i n crements, and r e t u r n i n g again t o t h e non-coordinate r e p r e s e n t a t i o n of t e n s o r q u a n t i t i e s , we f i n d (12.1) where t h e l e t t e r s 14 and TJ denote q u a n t i t i e s determined a t t h e f i x e d points M and N . I n t h e expanded form, i n t h e Cartesian coordinate system, eq.(12.1) tne form:
f

has

1 5 1

(12.2)
4

where A r i s t h e r a d i u s v e c t o r with i t s o r i g i n a t t h e p o i n t M and i t s terminus a t t h e point N. O n passage t o c u r v i l i n e a r coordinates, t h e d e r i v a t i v e s e find be replaced by t h e a b s o l u t e d e r i v a t i v e s ye,. W

a,

i n eq. (12.2)must

This equation d e f i n e s t h e expansion of t h e t e n s o r T of p a r a l l e l d i s p l a c e 1 . ment from point M t o g s i n t $ I n o t h e r words, this expansion d e f i n e s t h e components of t h e t e n s o r T a t p o i n t N i n terms of t h e v a l u e s of t h e s e components and t h e i r d e r i v a t i v e s a t point M and i n t h e metric of space a t point M. The proof of eq. (12.3) follows from two propositions: a ) On passage t o a Cartesian system o f coordinates, eq.(12.3) i s t r a n s formed i n t o eq.(12.2), which r e s u l t s from t h e c l a s s i c a l Taylor expansion. b ) A t e n s o r equation v a l i d i n any system of coordinates i s v a l i d in all o t h e r systems.

n-

45

I II IIII ~111111111 I

CI-IAF'TW. II' PRINCIPAL RELATIONS OF THE NONLIPJEAR THEOFU' OF ELASTICITY I N THE INVARIANT F O R M
Section 1 . N e r and Lagrange Variables. Displacement Vecbor, Velocity Vector and Acceleration Vector of an Element of a Continuous Medium

A n a r b i t r a r y system of c u r v i l i n e a r coordinates, determining t h e p o s i t i o n of p o i n t s of a continuous medium, b u t not connected with t h e medium, i s c a l l e d a system of N e r v a r i a b l e s . The M e r v a r i a b l e s of t h e p o i n t s of a continuous medium vary on i t s motion.
A system o f coordinates determining t h e p o s i t i o n of p o i n t s of a medium and m a t e r i a l l y connected with t h a t m e d i u m i s c a l l e d a system o f Lagrange variables, The Lagrange v a r i a b l e s of t h e p o i n t s of a medium do n o t vary on i t s motion.
Let us assume for s i m p l i c i t y t h a t t h e Euler v a r i a b l e s a r e t h e Cartesian coordinates F,, while t h e Lagrange v a r i a b l e s are t h e a r b i t r a r y c u r v i l i n e a r coordinates x , . The q u a n t i t i e s x i likewise determine a c e r t a i n N e r i a n coordin a t e system. This w i l l be discussed l a t e r i n Sections 2 and 3 .

Let us i n t r o d u c e in t h e space of Eulerian coordinates a r a d i u s v e c t o r determining t h e p o s i t i o n of t h e p o i n t s of t h e medium. When t h e p o s i t i o n of t h e p o i n t s of t h e medium v a r i e s , t h e r a d i u s v e c t o r of a c e r t a i n point M(xi ) w i l l a l s o vary. W e have

The increment u ( t , x1 ) o f t h e r a d i u s v e c t o r ;(O, x1 ), determining t h e i n i t i a l p o s i t i o n s of a p o i n t of t h e continsous m e d i u m , i s c a l l e d t h e displacement ( x i ). The vector u ( t , xi ) i s a function of t h e Lagrangian v e c t o r of t h e p o i n t M coordinates x i and t h e time t. Determining t h e components of t h e r a d i u s v e c t o r r ( t , ordinates, we o b t a i n
+

x*) in N e r i a n co-

where t h e are t h e "physical components'' of t h e v e c t o r <(t,2 1 i n M e r - /53 i a n coordinates. Equations (1.2) may be considered as formulas o f t r a n s i t i o n with t h e parameter t, connecting t h e Lagrangian and Eulerian coordinates.

B y d i f f e r e n t i a t i n g eq.(l.l) w i t h r e s p e c t t o t, w e find the v e l o c i w vector and t h e a c c e l e r a t i o n v e c t o r of an element of t h e continuous medium:

Section 2. Tensor of Small Deformations and Tensor of F i n i t e Deformations

1 . Tensor of Small Deformations and Vector of Small Rotation of an Element of a Continuous Medium
This equation permits t h e i n t r o d u c t i o n of t h e Let us r e t u r n t o eq.(l.l). fundamental q u a n t i t i e s d e s c r i b i n g t h e v a r i a t i o n o f t h e i n t e r n a l geometrical p r o p e r t i e s of a space i n v a r i a b l y bound t o t h e deformable medium. Such quantit i e s are t h e t e n s o r o f small deformations and t h e t e n s o r of f i n i t e deformations. Let u s consider f i r s t t h e t e n s o r of s m a l l deformations. D i f f e r e n t i a t i n g eq.(l.l) with respect t o t h e coordinates x i r we f i n d

Further,

dr (0, x i ) = d,r ( 0 , x i ) d x k = ekodxk.

-.

where dium.

go a r e

t h e v e c t o r s of t h e l o c a l coordinate b a s i s i n t h e undeformed meUsing eqs.(I,

Let u s continue t h e transformation of eq.(a).

9.3) and (I,

9 . 1 1 ) , we o b t a i n

where t h e covariant d e r i v a t i v e i s determined i n t h e metric of t h e undeformed medium, Consequently,

Denoting t h e contravariant components of t h e v e c t o r d r by dx~', we g e t

Equations (2.2) show t h a t t h e deformation of a continuous medium may b e regarded t o b e a result of l o c a l transformations of coordinates in t h e neighborhoods of t h e p o i n t s of t h e medium.

47

The transformation c o e f f i c i e n t s

(5k

( 2 . 3 )
are components of a mixed, t e n s o r of rank two i n t h e m e t r i c of t h e undeformed medium. Let us consider t h e t e n s o r

(i)i!= -8 ; Y k d =Y , $ .

(2.4)

The tensor2$ i s c a l l e d t h e d i f f e r e n t i a l expans&zn of t h e v e c t o r G(Bibl.7). W e introduce t h e covariant Components of t h e t e n s o r @ and expand this t e n s o r W e find i n t o i t s symmetric and antisymmetric p a r t s (I, 7.4).
Oki
= - - - ( T klLi

= 1

1 + Ti u,;)-t2

(CklLi - VilLk).

The symmetric t e n s o r

i s c a l l e d t h e t e n s o r of small deformations of an element o f t h e continuous medium. The meaning of t h i s term w i l l be explained below.
The antisymmetric t e n s o r (2.7)

leads, on t h e b a s i s of (I, 8 . 2 ) t o t h e v e c t o r

, 2, 3. The @dices j , k, i are a c y c l i c permut3tion of t h e numbers 1 t o r 9 i s c a l l e d t h e curl of t h e v e c t o r u:

0

The vec-

__ -- -curl
.

fL.

(2.9)

It i s w e l l known t h a t t h e v e c t o r R approximately determines t h e absolute r o t a r y displacement of t h e p a r t i c l e s of t h e medium (Bibl.?)-X-.

* The

i d e n t i f i c a t i o n of t h e v e c t o r R with t h e mean angle of r o t a t i o n i s possi-, b l e o n l y i n t h e l i n e a r theory ( B i b l J l b ) .

L L 8

O n t h e b a s i s o f (I, 9.9d) we note t h a t t h e generalized r e l a t i v e angle of r o t a t i o n of adjacent elements of t h e deformed medium i s expressed i n terms of t h e C h r i s t o f f e l symbol in t h i s medium. To o b t a i n a complete idea of t h e kinematics of a medium a f t e r deformation, one m u s t t u r n t o t h e i n v e s t i g a t i o n of its metric. 2. Tensor of F i n i t e Deformations

& .

To f i n d t h e kinematic q u a n t i t y c h a r a c t e r i z i n g t h e change of t h e d i s t a n c e between two p o i n t s of a continuous medium under $ e f o m t & o n , and t h e change i n t h e angle between t h e d i r e c t i o n of two v e c t o r s dr, and br, o r i g i n a t i n g a t t h e a r b i t r a r y p o i n t M(x1) of an undeformed m e d i 5 under deformation, l e t us consid e r t h e change i n t h e s c a l a r product d?o 6ro caused by deformation. W e have, on t h e b a s i s of eq. (b ):

and

dr, = eiodxi; 6ro= ek,6xk

-. -.

-. -.

where t h e components of t h e metric t e n s o r r e l a t e t o t h e undeformed s t a t e of t h e medium. Further, by t h e a i d of eq.(2.1), w e obtain

d r 81- = (eio

-- - . +ejo - vi
+

uj) (eko

-+- eroVkur) dxi6xk =

f

( 2 . 1 0 )

The expressions i n parentheses are t h e covariant components of t h e symmet r i c t e n s o r of rank two:

Equations (2.11) determine t h e t e n s o r of f i n i t e deformations of t h e continuous medium. From a comparison of eqs.(2.11) and (2.6) follows t h e following relation:
2 0 , = 2"ik

+@!;
49

@kj.

(f)

1111

IIII " I

III II I 111

For small values of t h e t e n s o r components2$, t h e t e n s o r D , , imately coincide with t h e t e n s o r of s d l deformations e l k .

It w i l l b e seen from eq.(2.10) determined by t h e equations

will approx-

t h a t t h e metric i n t h e deformed medim i s (2.12)

Hence, from eqs.(I.5.9. ) we may f i n d t h e contravariant and mixed components of t h e metric tensor, and then t h e C h r i s t o f f e l symbols and t h e operation of abso'lute d i f f e r e n t i a t i o n i n t h e metric of t h e deformed medium.

3 . Concluding Remarks

/56

The reader has probably noted a c e r t a i n arbitrariness i n t h e construction of t h e t e n s o r of small deformations and t h a t of t h e t e n s o r of f i n i t e deformae d i d i n fact determine t h e increment of t h e displacement v e c t o r in tions. W t h e metric of t h e undeformed medium. It would have been possible, however, t o use t h e metric of t h e deformed medium. The a r b i t r a r i n e s s i n t h e choice of t h e m e t r i c i s n o t f o r t u i t o u s . This randomness i s due t o t h e f a c t t h a t i n a general study of t h e i n t e r n a l geometry of manifolds of coordinates xi, t h e metric i s introduced by d e f i n i t i o n and cannot b e connected in advance with t h e p r o p e r t i e s of t h e manifold. These i d e a s are w e l l known from modern d i f f e r e n t i a l geometry (Bibl.6). W e have chosen t h e simp l e s t method of d e f i n i n g t h e m e t r i c of a deformed medium and a t t h e same time have defined t h e t e n s o r s of s m a l l and f i n i t e deformations. A d i f f e r e n t j u s t i f i :ation of t h e r e l a t i o n s obtained i s also possible. One could a s s e r t t h a t t h e i i n t h e undeformed medium simultaneously d e f i n e two systems of coordinates x coordinates, t h e Ehlerian and Lagrangian. The expressions found f o r t h e t e n s o r components of s m a l l and f i n i t e deformations a r e connected w i t h t h e N e r i a n coo r d i n a t e system. Section 3 . Conditions of Compatibility Equations (2.11) determine t h e f i n i t e d e f o r m a t i o n t e n s o r components i f we know t h e components of t h e v e c t o r of displacement of an element of t h e contin-' uous medium.

It i s n a t u r a l t o pose t h e inverse problem; t o f i n d t h e displacement v e c t o r from t h e components of t h e finite-deformation tensor. This problem i s solved by i n t e g r a t i n g a system of six nonlinear equations (2.11) with t h r e e unknown functions, t h e v e c t o r components q. Obviously, t h e p o s s i b i l i t y of a singlevalued determination of t h e functions u, from t h e system of equations (2.11) must be assured by s a t i s f a c t i o n of a d d i t i o n a l conditions imposed on t h e compone n t s of t h e s t r a i n tensor. It i s simplest here t o stazt out from general geom e t r i c a l considerations. The existence of t h e v e c t o r u i s equivalent t o t h e existence o f t h e coordinate transformation formulas (1.2), and of transformat i o n formulas i n v e r s e t o eqs.(l.2), permitting us t o pass from t h e metric i n t h e deformed medium t o t h e i n i t i a l metric. B u t t h e i n i t i a l m e t r i c i s t h e

50

m e t r i c of Euclidean space. I n t h i s metric, t h e curvature t e n s o r vanishes idenConsequently, a l s o i n t h e deformed medium t h e curvature t e n s o r w i l l tically. vanish i f t h e r e " x i s 2 t h e transformation formulas (1.2) o r i f t h e r e exists a displacement v e c t o r u as a single-valued function of t h e coordinates x i at a fixed time t. The f a c t t h a t t h e components of t h e curvature t e n s o r vanish i s t h e Wanted condition, which must b e s a t i s f i e d by t h e s t r a i n t e n s o r components i n order t h a t t h e displacement v e c t o r determined from eqs.(2.11) be i n existence. M a k m i n g u s e of ( 1 , 1 0 . 7 ) , we f i n d :

a r e C h r i s t o f f e l symbols of t h e f i r s t kind expressed i n terms of t h e where ri:)Jk S u b s t i t u t i n g eqs. (2.11) metric t e n s o r components of t h e deformed medium, Gr 6 . i n t o eqs.(3.1), we f i n d t h e required compatibility conditions o f eqs.(2.11), or A s p e c i a l case of eqs.(3.1), f o r small deformat h e i n t e g r a b i l i t y conditions. t i o n s , i s given by t h e well-known Saint-Venant condition&*. Section 4. S t r e s s Tensor. Generalized Hookets Law

1 . Linea-r Generalization of Hookers Law. Physical and Geometric N_onlinearity of t h e Quations of t h e Theory of E l a s t i c i t y

The second t e n s o r determining t h e state of t h e deformed medium i s called t h e stress tensor. I t s p r o p e r t i e s are w e l l known from t h e p r i n c i p l e s of t h e mechanics of a continuous medium, and they w i l l n o t be discussed here. The stress t e n s o r and t h e s t r a i n t e n s o r a r e c o r r e l a t e d by a system of rel a t i o n s r e s u l t i n g from t h e generalized Hookers law. This connection i s u s u a l l y considered as l i n e a r and i s accomplished by means of t h e e l a s t i c i t y t e n s o r elk, r I From e n e r g e t i c considerations it follows t h a t t h e components of t h e e l a s t i c i t y t e n s o r cikD r B are qynunetric i n t h e l a b e l s i and k, r and s, and t h e p a i r o f i n d i c e s i k , rs. Thus i n t h e most general case of anisotropy of t h e mat e r i a l , t h e t e n s o r C i k t r * has only 21 independent components.

The generalized Hookets l a w i n t h e i n v a r i a n t form i s expressed as follows:

$5

Cf.E.Trefftz,

Mathematical Theory o f E l a s t i c i t y and a l s o (Bibl.7).

ONTI, 1934

51

where

d k

are t h e contravariant components of t h e stress tensor.

The expressions f o r t h e components of t h e deformation t e n s o r in terms of t h e . s t r e s s t e n s o r a r e of t h e following form:

'Eo f i n d t h e s e where t h e q u a n t i t i e s y r r , l k a r e expressed i n terms of C i k , expressions, i t i s s u f f i c i e n t t o perform t h e i n v e r s i o n of eqs.(b.la), solving@ t h e system o f l i n e a r equations (,!+.la) with r e s p e c t t o Drs

L e t u s consider an i s o t r o p i c medium. I n t h e case o f an i s o t r o p i c medium t h e e l a s t i c i t y t e n s o r has only two s u b s t a n t i a l l y d i f f e r e n t components. A l l t h e components of t h e e l a s t i c i t y t e n s o r can b e expressed i n terms of two independ e n t q u a n t i t i e s , . w h i c h a r e constants i n a homogeneous medium.
W e now introduce ,the Lam4 constants h and p :

EV

(1 - 2 v ) ( l + v )

-,

[ . L E

E 2(1+

V j

The inverse r e l a t i o n s a r e of t h e following form:

I n eqs.(4.2a) - (4.2b), E i s Youngfs modulus, ahd u i s Poisson's constant. Equations (,!+.la) can then be represented in t h e following form: (4.3
oilc

= )\h oik

0 f- 2pgirg""Drs.
(4.4)

where
0 = grsDrs

i s t h e l i n e a r i n v a r i a n t of t h e s t r a i n tensor.
The q u a n t i t i e s g p q are t h e contravariant somponents of t h e metric t e n s o r of t h e undefomed medium. The i n t r o d u c t i o n of t h e metric t e n s o r o f t h e deformed medium would here be superfluous, s i n c e it would l e a d t o a nonlinear rel a t i o n between t h e components of t h e s t r e s s t e n s o r and those of t h e s t r a i n tensor, which would be contradictory t o eq.(l+.la).

52

W e s h a l l n o t dwell on t h e problem of j u s t i f i c a t i o n of t h e a n a l y t i c express i o n of t h e generalized Hookers l a w defined by eqs.(.!+.la) ( 4 . 3 ) , b u t shall adopt these equations as t h e d i r e c t consequences of experimental data, which are v a l i d i n a c e r t a i n region o f v a r i a t i o n of t h e stress t e n s o r and t h e s t r a i n tensor.

It follows from eqs.(4.3)

(4.4) t h a t

Passing t o t h e covariant components of t h e stress tensor, we find:

A comparison of eqs.(4,5a) and (4.la) l e a d s t o t h e following expression f o r t h e components of t h e e l a s t i c tensor:

Equations (,!+.la) - (4.lb) and (4.5a) express t h e linear Hookers law, s i n c e t h e components of t h e stress t e n s o r and t h e strain t e n s o r e n t e r l i n e a r l y A t t h e same time, it mst be emphaszzed t h a t t h e teni n t o these relations. s o r D,, contains n o n l i n e a r terms i n t h e v e c t o r components u and t h e i r derivai . In this connection, we d i s t i n g u i s h t i v e s with r e s p e c t t o t h e coordinates x between t h e physical n o n l i n e a r i t y and t h e geometrical n o n l i n e a r i t y of t h e equat i o n s of t h e e l a s t i c i t y theory or t h e equations of t h e mechanics of a continuous medium with p r o p e r t i e s more general than those of an e l a s t i c b o d p . The nonlinear terms e n t e r i n g i n t o t h e composition of t h e t e n s o r of finite' deformation determine t h e geometrical n o n l i n e a r i t y of t h e equations. Physical n o n l i n e a r i t y depends on t h e form of f u n c t i o n a l connection between t h e compone n t s of t h e s t r e s s t e n s o r and those of t h e s t r a i n tensor.

2. The Nonlinear Hookers Law For an a n i s o t r o p i c body, on i n t r o d u c t i o n of terms containing products and squares of t h e components of t h e s t r a i n tensor, we o b t a i n

(4.7)
Here we meet two e l a s t i c tensors: t h e t e n s o r Cik* ,', discussed above, and V.V.Novozhilov i n h i s monograph ( B i b L l l b ) g i v e s clear-cut d e f i n i t i o n s of t h e s e forms of nonlinearity.

53

, t h e t e n s o r of rank six g k pk ~ , ' I , This t e n s o r i s symmetric i n t h e l a b e l s i, k p, q , r, s and t h e p a i r s of corresponding indices. A n elementary c a l c u l a t i o n shows t h a t t h i s t e n s o r has 79 s u b s t a n t i a l l y d i f f e r e n t components. Consequently, t h e r e are 100 s u b s t a n t i a l l y d i f f e r e n t components of t h e t e n s o r s C i k , r s and C i k $ P P I I , taken together.

It i s d i f f i c u l t t o o b t a i n t h e expressions determining t h e s t r a i n t e n s o r components i n terms of t h e s t r e s s t e n s o r components by inversion of eqs,(4.7), s i n c e such inversion l e a d s t o t h e s o l u t i o n of a system of six q u a d r a t i c equat i o n s , i.e., t o t h e s o l u t i o n of an a l g e b r a i c equation of t w e l f t h degree. Such an equation i n t h e general case cannot be solved i n r a d i c a l s ,
A l l this i n d i c a t e s t h e g r e a t d i f f i c u l t i e s t h a t a r i s e i n t h e study of probl e m s of t h e mechanics of a n i s o t r o p i c e l a s t i c bodies with trphysical nonlinearity".
Consider now an i s o t r o p i c body. I n o r d e r t o s e t up t h e i n v a r i a n t express i o n of t h e generalized Hookefs law, including terms of t h e form D p q D r , , it i s s u f f i c i e n t t o consider t h e components of a t e n s o r of rank six, constructed from t h e contravariant components of t h e metric tensor, permitting us, as a r e s u l t of m u l t i p l i c a t i o n and contraction, t o f i n d a d d i t i o n a l l i n e a r l y independent terms e n t e r i n g i n t o t h e composition of t h e components o f t h e s t r e s s t e n s o r /60 These components of t h e required t e n s o r of rank six, a s can e a s i l y be aik. v e r i f i e d , a r e expressed by t h r e e combinations: g i k g p q g s, gPqgi 2 s, gip gik@s. rhus, t h e generalized nonlinear Hookefs l a w may be represented by t h e following i n v a r i a n t equation:

Equation (4.8) contains f i v e parameters determining t h e e l a s t i c p r o p e r t i e s a m 6 constants h and )J. and t h e a d d i t i o n a l c o e f f i c i e n t s cl, of t h e medium: The L %, %. These c o e f f i c i e n t s are constants i n a homogeneous body. A l l t h e above-enumerated c o e f f i c i e n t s a r e experimentally determined. Both eqs. (4.5a)(4.6) and eq. (4.8) w i l l contain contravariant components of t h e metric t e n s o r of an undeformed medium. Indeed, t h e a p p l i c a t i o n of t h e metric t e n s o r compone n t s of t h e deformed medium would lead, as i s c l e a r from eq.(2.12), t o t h e introduction i n t o eq,(LL.$) of a d d i t i o n a l terms of t h e t h i r d dimension with reThis would c o n t r a d i c t eqs.(4.7) spect t o t h e components of t h e s t r a i n tensor. by which, i n advance, we r e s t r i c t e d t h e accuracy o f t h e wanted r e l a t i o n . muat i o n (4,8)in essence coincides with t h e "Voigt-14urnaghan law". A c r i t i c a l analy s i s of c e r t a i n conseqcences t h a t result f r o n r e l a t i o n s a n a l o g o u s t o eq.(L5.t3) i s given elsewhere ( B i b L l l b ) .

3 . Concluding Remarks
I n most works on t h e mechanics of deformable bodies, t h e construction of generalized formulations of Hookefs l a w i s based on e n e r g e t i c considerations. 54

A d e t a i l e d exposition of e n e r g e t i c p r i n c i p l e s would be o u t s i d e t h e scope of this book, and we have t h e r e f o r e employed an outwardly formal method, postulat i n g t h e invariance of t h e l a w sought and using t h e propositions o f t e n s o r a l gebra f o r t h e construction of i t s i n v a r i a n t formulation. The f o m l i s m of t h i s method 5 s i l l u s o F j . It i s w e l l known from modern physics t h a t t h e requirement of t h e invariance of t h e mathematical formulation of t h e l a w s of nature r e s u l t s from generalized p r i n c i p l e s , which are t h e expanded e n e r g e t i c considerations t o which we have r e f e r r e d above.

Section 5. Equations of Motion of an Element of a Continuous Medium. The Linear Lame' Equations

1 . Equations of Motion of an ELement of a Continuous Medium i n an A r b i t r a r y System o f L a g r a n g e Coordinates

The equations t o be considered determine t h e motion of an element of a deformed continuous medium. For this reason, i n determining t h e components of t h e metric tensor, t h e C h r i s t o f f e l symbol, and t h e fundamental determinant, which are necessary f o r s e t t i n g up t h e equations of motion, w e must base our-& Here t h e components of t h e s t r a i n t e n s o r e n t e r i n t o t h e s e l v e s on eqs.(2.12). fundamental determinant and t h e C h r i s t o f f e l symbols.
VarThe fundamental determinant i n t h e deformed medium will be c a l l e d G. i o u s q u a n t i t i e s connected with t h e deformed medium w i l l be i n d i c a t e d by t h e i n The C h r i s t o f f e l symbols i n t h e metric of t h e deformed medium will be dex (D). i n d i c a t e d by b r a c k e t s and braces:

(5.1) The covariant d e r i v a t i v e i n t h e deformed medium w i l l be i n d i c a t e d by

9).

I n t h i s n o t a t i o n t h e equations of motion of an element of a deformed medium a d , i n p a r t i c u l a r , of an element of an e l a s t i c or p l a s t i c body, a r e o f t h e following form [cf., f o r i n s t a n c e (Bibl.7, 8, l l b ) ] :

where p i s t h e d e n s i t y 3 f t h e material o f t h e body, a d p a r e t h e contravariant components of t h e m a s s forces. Making use o f (I, 9.12), we represent eq.(5,2a) i n t h e form

W e now transform t h e sum of t h e C h r i s t o f f e l symbols

ii;}.

O n t h e b a s i s of

55

eq,(I, find

5.9a) and (I, 9.8),

after t h e necessary changes i n t h e d w q y indices, we

Consequently,

(5.3

S u b s t i t u t i n g eq.

( 5 . 3 ) i n t o eq. (5.2a>, we o b t a i n

To t h e systems of equations (5.2a) or

(5.4) we must a s s o c i a t e t h e equation

expressing t h e l a w of t h e conservation of mass. Equation ( a ) i s u s u a l l y called t h e equation o f c o n t i n u i t y [cf., f o r i n s t a n c e (Bibl.7)]. 2. Linear ~ a m e Equations

W e 3hal.l assume that t h e components of t h e t e n s o r of t h e d i f f e i e n t i a l expansions $ (Section 2 ) are small q u a n t i t i e s . Accordingly, in t h e generalized l i n e a r Hookefs l a w (4.5a), we will replace t h e components of t h e f i n i t e deformation t e n s o r by t h e components of t h e t e n s o r of small deformations c i k . W e exclude t h e s t r a i n t e n s o r components of t h e equations of motion (5.2a) from t h e components of t h e m e t r i c t e n s o r and t h u s a l s o from t h e C h r i s t o f f e l symbols. Then, t h e covariant d e r i v a t i v e V y J i s transformed i n t o t h e covariant derivat i v e Vj in t h e undeformed medium. S u b s t i t u t i n g t h e expressions f o r t h e s t r a i n t e n s o r components (k.5a) i n t o eq. (5.2a), we obtain, a f t e r simple transformat i o n s , t h e well-hown Lam6 equations i n an a r b i t r a r y c u r v i l i n e a r coordinate syst e m (Bibl.7):

56

Multiplying eq0(5.5a) by g l r , performing t h e operation of contraction, and making use of t h e R i c c i theorem, we o b t a i n

(I-, k , s-

1, 2, 3).

Before s e t t i n g up t h e nonlinear L a " auxiliaqy propositions.

equations, l e t us consider s e v e r a l

Section 6. Relationships between Covariant Derivatives i n Deformed and Undeformed Nedia

1 . Fundamental Determinant

Consider t h e expressions f o r t h e q u a n t i t i e s determining t h e metric and p a r a l l e l displacement i n t h e space of t h e Lagrange coordinates of a deformed m e d i d ; , s e p a r a t i n g from t h e s e q u a n t i t i e s t h e p a r t s l i n e a r l y connected with t h e s t r a i n t e n s o r components, Consider f i r s t t h e fundamental determinant. and eqs. (2.12), we f i n d : Applying t h e Taylor formula

2. Covariant and Contravariant Components of t h e Metric Tensor of a Deformed Medium

/63

The covariant components of t h e metric t e n s o r i n t h e deformed medium a r e expressed by eqs.(2.12).. Consider t h e contravariant components of t h e m e t r i c tensor. W e introduce a s y s t e m of generalized Kronecker d e l t a 6 i k J which have t h e following p r o p e r t i e s : i f t h e i n d i c e s i, k, j form a p o s i t i v e c y c l i c permut a t i o n of t h e numbers 1 , 2, 3 , then t h e q u a n t i t i e s 6 i k J w i l l be equal t o +1, whereas i f t h e s u p e r s c r i p t s i, k, j form a negative c y c l i c permutation of t h e 1 ; and i f two i d e n t i c a l numbers 1, 2, 3 , then t h e q u a n t i t i e s G i k J a r e equal t o numbers are p r e s e n t i n t h e i n d i c e s i , k , j , then t h e generalized Kronecker d e l t a 6i k vanishX+.
3s
33'i

For brevity, we s h a l l speak h e r e a f t e r of "in t h e deformed medium" e-tc.,

Detailed information on t h e p r o p e r t i e s of t h e q u a n t i t i e s t o which t h e j u s t 3 belong, will b e found in O.Veblen* s introduced generalized Kronecker d e l t a book "Invariants of d i f f e r e n t i a l quadric forms", Chapter 1 .,IL, 1948.

57

The fundamental determinant g may b e represented by t h e following formula:

g = W j gr .lgk2gj 3 *
Similarly, b e a r i n g i n mind eq.

(6.2a)

(6.1),

This equation may be simplified by making u s e of t h e i d e n t i t y

Applying t h e Taylor f o m l a and eqs.(2.12),

we find

or, on t h e b a s i s of (I, 5.9a) and eq.(6.2a),

The expressions i n parentheses a r e t h e components o f a t e n s o r o f rank

fmr.
W e will consider only t h a t portion of t h i s t e n s o r which i s symmetric i n t h e l a b e l s r and s, since t h e tensor D , , i s symmetric i n t h e s e indices.

L e t us put

58

The t e n s o r A I k , r s i s symmetric i n t h e s u p e r s c r i p t s i and k, r and s and t h e p a i r s o f t h e s e indices. Thus, we f i n d

Gik
=e g i k

+A f k , 1s Q S + .

.. . .

(6.6~~)

Fiereafter w e s h a l l retain, i n t h e equations of ntotion o f tQ5continuous Under t h i c: medium, only terms which are qu-adratic i n t h e t e n s o r components Cp. condition, t h e small-deTormation t e n s o r must b e m b s t i t u t e d f o r t h e f i n i t e deformation t e n s o r i n eq. (6.6a). W e obtain
G'k

=g ' k

+AikSrs + ...

(6.6b)

It is here assumed t h a t t h e q u a n t i t i e s e , , are nonvanishing. The case where any component of e r s vanishes r e q u i r e s s p e c i a l i n v e s t i g a t i o n ( s e e also Sect. 8.h).
The d i f f e r e n c e s

(6.7)
;nay be regarded a s components of t h e contravariant s t r a i n t ; e n ~ o r . It i s c l e a r

e musz d i s t i n g u i s h from eq.(6.6a) t h a t , i n t h e nonlinear theory of e l a s t i c i t y , w t h e components of t h e c o n t r a v a r i a n t s t r a i n t e n s o r from t h e contravariant components o f t h e covariant s t r a i n t e n s o r defined by t h e equations

W e will not f u r t h e r go i n t o t h e s e questions.

3. C h r i s t o f f e l Symbols i n a Deformed Nedium

Calculating t h e C h r i s t o f f e l symbols of t h e f i r s t kind in t h e metric of t h e deformed medium, w e obtain

(6.9)
where

Retaining i n t h e equations of motion o f an elemenJ40f a continuous medium only terms t h a t are q u a d r a t i c i n t h e t e n s o r components @ i n s t e a d o f t h e quant i t y yJ, i k w e must consider t h e q u a n t i t i e s rrJ, f k :

/65 O n t h e b a s i s of (I, 9.8) it i s easy t o e s t a b l i s h t h a t t h e q u a n t i t i e s y J , l k a r e C h r i s t o f f e l symbols of t h e f i r s t kind i n a space whose metric i s determined by t h e components of t h e s t r a i n t e n s o r D , The q u a n t i t i e s rrj, i k a r e C h r i s t o f f e l syrribols of t h e f i r s t kind in a space with t h e metric t e n s o r elk.
L e t u s f i n d t h e C h r i s t o f f e l symbols of t h e second kind i n t h e metric of t h e deformed medium. W e have

Retaining i n t h e l e f t s i d e of eq.(a) t h e terms l i n e a r l y dependent on t h e s t r a i n t e n s o r components and t h e i r d e r i v a t i v e s , w e find

(6.11)
T t J e now introduce t h e n o t a t i o n

Since we s h a l l r e t a i p + i n our equations only n&inear terms wQ&ch a r e q u a d r a t i c i n t e n s o r components aj, we replace t h e t e n s o r P by t h e t e n s o r K with components expressed as follows:

or, i n accordance with eq.(6.1%),

{ !k)
5

+Nii!

(6.13b)

de s h a l l show t h a t t h e q u a n t i t i e s Pl;: a r e mixed components of a t e n s o r of t h i r d rank. This proof i s a l s o extended t o t h e q u a n t i t i e s W;;?

60

R Ii
i

Now, s e t t i n g up t h e transformation formulas f o r t h e C h r i s t o f f e l symbols rll, by eqs.(I, 9.7b), we f i n d t h a t t h e i r d i f f e r e n c e Pi;? (or, approximately, Nir? ) obeys t h e transformation formulas f o r t h e mixed components of a . t e n s o r of t h i r d rank. T h i s t e n s o r i s symmetric i n t h e i n d i c e s i and k

{ i { ] and

4. Covariant Derivative i n a Deformed Medium

Consider again (I, 9.12).
W e have

(6.U)

I n particular, w e f i n d f o r t h e contravariant v e c t o r

/66
(6.1%

and f o r t h e covariant v e c t o r

10.9) and eq.(6.1%),

Let u s consider t h e commutativity of t h e o p e r a t o r s V y ) and V, we f i n d

From (I,

Equation ( 6 . 1 6 ) i s s i m p l i f i e d i f t h e space f i l l e d by t h e medium i s Euclidean. It i s p r e c i s e l y t h i s case t h a t w e s h a l l consider hereafter. But even i n t h i s case, t h e o p e r a t o r s V , and 0 : ) are n o n c o m t a t i v e , which considerably comp l i c a t e s t h e nonlinear equations o f motion of an element of an e l a s t i c or plast i c body.

5. Conclusion
The r e l a t i o n s (6.1) (6.16) found by u s permit d e r i v i n g approximate nonl i n e a r equations of motion of an element of an e l a s t i c or p l a s t i c body in t h e m e t r i c of t h e undeformed medium. These r e l a t i o n s p e r n i t an inversion: a l l t h e q u a n t i t i e s required f o r this construction can b e expressed i n t h e m e t r i c of t h e deformed medium.

To summarize, we may assert t h a t we have constructed a fundamental system of q u a n t i t i e s which permit us t o set up t h e nonlinear approximate equations o f motion of an element of a continuous medium in an a r b i t r a r y c u r v i l i n e a r system

61

o f coordinates, i.e., i n t h e i n v a r i a n t form, i n one of two metrics: e i t h e r i n t h e m e t r i c of t h e undeformed m e d i u m or i n t h e m e t r i c o f t h e deformed medium. Section 7. Nonlinear L a 6 Q u a t i o n + I n order t o s e t up a system o f e q u a t i o n s p e r m i t t i n g i n v e s t i g a t i o n of t h e motion of t h e p a r t i c l e s o f e l a s t i c bodies under f i n i t e deformations, we must make use of t h e equations of motion,eqs. (5.&),and t h e nonlinear HookeTs law (BIurnaghanTs law) expressed by eqs.(4.7) - ( 4 . 8 ) . Here we must know t h e COFSince t h e components of t h e e l a s t i c t e n s o r ponents of t h e e l a s t i c tensor. have been l i t t l e investigated, even i n t h e case of an i s o t r o p i c body, we s h a l l confine ourselves t o t h e l i n e a r TIooke's l a w (L5a).

c,

Ye r e c a l l t h a t eqs.(4,5a) contain terms nonlinear i n t h e t e n s o r components so t h a t , without considering physical n o n l i n e a r i t y , we s h a l l preserve geom e t r i c a l nonlinearity.

' 6

fi7_

Let u s b e a r i n mind eqs.(6.1) following form:

and (6.13b).

W e r e p r e s e n t eq.(2.11)

i n the

Xe s h a l l denote by obk t h e s t r e s s t e n s o r components expressed on t h e b a s i s of eqsm(k.5a) i n terms o f t h e components of t h e t e n s o r of small deformations E,,

Then,

The equations o f motion o f a r l element of a continuous medium [eqs.(r.L)] may be represented i n t h e form

where

3:-

These equations were f i r s t considered by u s i n Reference 23b, P a r t 1 1 , Sect i o n 4. 62

The q u a n t i t i e s @i determine t h e influence o f g e o n e t r i c a l n o n l i n e a r i t y on t h e motion of an element o f an e l a s t i c body. These q u a n t i t i e s a r e eqtiivalent t o a d d i t i o n a l body forces. Thus, t o s e t up t h e nonlinear Lame' equati,ons, it i s s u f f i c i e n t t o i n t r o duce t h e a d d i t i o n a l body f o r c e s i n t o t h e l e f t s i d e of t h e l i n e a r Lame' equaW e obtain t i o n s (5.5a) o r (5.5b).

These i n v a r i a n t equations may obviously b e represented i n v e c t o r form:

where

i s t h e La.place operator i n an a r b i t r a r y c u r v i l i n e a r system of coordinates, and

Section 8. L n i t i a l q d Eoundary IJonlinear Conditions. Conditions of Contact of Layers

1 . I n i t i a l - Conditions

/hs

The statement o f problems of t h e mechanics of deformable bodies includes,

as a necessary element, t h e assignment of a system of i n i t i a l and boundary cond i t i o n s . Since t h e d i f f e r e n t i a l equations o f motion o f an element of a continuous mediun are equations of t h e second order w i t h respect t o t h e time t, t h e c l a s s i c a l i n i t i a l conditions become applicable: A t some time t = to, a r b i t r a r i l y called t h e i n i t i a l t i m e , t h e p o s i t i o n s and v e l o c i t i e s of t h e elements of t h e deformable medium must be assigned. In connection with t h e Tact t h a t t h e p o s i t i o n s and v e l o c i t i e s of t h e elements o f t h e medium are expressed i n terms of t h e displacement v e c t o r components and t h e i r time d e r i v a t i v e s , we have t h e following i n i t i a l conditions :

(sola.)
(8.lb)

63
I

where t h e d o t denotes d i f f e r e n t i a t i o n with r e s p e c t t o time.

2. Monlinear Boundary Conditions

W e s h a l l consider below o n l y t h e case where t h e continuous medium i s a s o l i d body and, i n p a r t i c u l a r , an e l a s t i c body.

A f e a t u r e of boundary problems of t h e mechanics o f s o l i d deformable bodies under f i n i t e displacements and deformations of t h e i r elements i s t h e assignment of boundary conditions on t h e deformed s u r f a c e of t h e body whose shape i s t o b e n t h e deformed surface of t h e body may be assigned: a ) t h e comdetermined. O ponents of t h e displacement vector, b ) t h e components of t h e s t r e s s vector, and c ) t h e mixed boundary conditions. The expressions of t h e boundary conditions, l i k e t h e equations of motion of an element of a continuous medium considered zbove, r e s u l t from t h e general equation o f dynamics. The d e r i v a t i o n of t h e s e conditions will n o t be discussed here, and t h e reader i s r e f e r r e d t o general Handbooks on e l a s t i c i t y theory (cf., f o r example, Bibl.1lb).
W e will confine ourselves here t o more elementary considerations. Assume t h a t i n t h e case o f t h e boundary problem a), t h e displacement v e c t o r components are assigned as f u n c t i o n s of t h e Lagrangian coordinates of t h e p o i n t s of t h e body surrace. Problem b ) w i l l be discussed i n g r e a t e r d e t a i l .
The components cf t h e s t r e s s v e c t o r f are expressed by t h e equations
oil;llk

= fi

(8.2a)

or
sialtk= f,

(i, k -

1, 2, 3).

whers rq, a r e t h e covariant components o f t h e u n i t v e c t o r of an e x t e r n a l nor- /6p m a l n t o t h e deformed surface of t h e body. Since t h e shape of t h e deformed surface of t h e body i s t o be determined, l e t u s express t h e components of t h e u n i t v e c t o r of t h e e x t e r n a l n c p " in terms of t h e components of t h e u n i t vector of t h e e x t e r n a l normal n , t o t h e undeformed surface. Let us make use of eq.(l.l) and assune t h a t t h e equations o f t h e surface of t h e body i n parametric form read a s f o l l o w s :
xi - x i ( E ' ,

5')

(i = 1, 2, 3).

(8.3)

where 5@ (a = 1,2) a r e t h e Gaussian coordinates o f t h e p o i n t s of t h e surface of t h e body. W e note t h a t eqs.(8.3) remain unchanged under deformation of t h e body.

The coordinate v e c t o r s of t h e l o c a l coordinate b a s i s on t h e surface of t h e body a r e determined by t h e equations

64

e. = ----_ -

dr

dr

ar

ax' - el7 ax;

at=
d E

( i = l , 2, 3; a = 1 ,

2).

axi

The unit v e c t o r of t h e e x t e r n a l normal t o t h e deformed surface of t h e body i s determined by (I, 3.1):

-= n

e, -x .. I x I
e1
e1
e2

- -

It i s here assumed t h a t t h e choice of t h e parameters Sg will ensure t h e It follows from eqs.(8.1L) and (8.5) direction selected f o r t h e u n i t vector that

z.

Making use o f eq.(I,

8.6),

w e find

These where t j i k r a r e q u a n t i t i e s analogous t o those considered i n Sect.6.2. q u a n t i t i e s a r e equal t o +1 if t h e l a b e l s i, k, r form a p o s i t i v e c y c l i c permu, 2, 3, while they are equal t o 1 i f this permutation t a t i o n of t h e numbers 1 i s negative and vanish i n a l l o t h e r cases. Then, t h e covariant components of the vector determined by eq.(8.6), w i l l b e expressed as follows:

z,

where

65

and where t h e q u a n t i t i e s Cj and B,, body.

do n o t depend on t h e deformation o f the a r e determined from

The contravariant components of t h e m e t r i c t e n s o r Plaking use of t h e s e equations, we g e t eqs0(5.6a).

(B,,Grs)-~ = [B,, (gr" A rs ,p q Dpq

+ .. .)J-i
=

Confining ourselves t o t h e l i n e a r approximation, w e find

J!Joting t h a t t h e covariant components of t h e u n i t v e c t o r equations

a r e expressed by t h e

noj=Cj (B,, grs)-i 7

we obtain

(8.11)

(8.12)

t h e condition (8.2a) f o r t h e deformed surface O n t h e basis of eq.(8.12), of t h e body t a k e s t h e following form:

(8.13 )

Equations (8.13) approximately express t h e nonlinear boundary conditions of t h e problem b). After t h e s e remarks, t h e a n a l y t i c a l statement of problem c )

66

111

i s now obvious.

3. Conditions of Contact on Surfaces of Separation between Media

w i t h l r a t t e r of D i f f e r e n t Mechanical C h a r a c t e r i s t i c s I n t h e s h e l l theory we have t o do with layered aggregates, The theory of The conditions f o r t h e layered s h e l l s i s developed i n a monograph (Bib1.1). i n t e r f a c e s between l a y e r s can b e very varied and depend on t h e method of con-&, s t r u c t i o n o f t h e layered s h e l l wKch i s e s s e n t i a l l y a system of s h e l l s .

Assume, for instance, t h a t t h e design o f a layered s h e l l ensures t h e conThe stress t e n s o r components on t h e int i n u i t y of t h e f i e l d of displacements. t e r f a c e between t h e l a y e r s must obey t h e conditions r e s u l t i n g from Newtonfs Third Law. Thus, on t h e i n t e r f a c e of media l a b e l e d k and k + 1 , t h e conditions

must be s a t i s f i e d .

4. e n e r a 1 Characterization o f t h e Formulation o f Konlinear

Problems of t h e Theoxy o f E l a s t i c i t y The problems of t h e nonlinear theory of e l a s t i c i t y belong t o two classes. The f i r s t c l a s s c o n s i s t s o f weakly nonlinear problems and t h e second, o f strongThe problems of t h e first c l a s s a r e charac&rized by l y nonlinear problems. t h e f a c t t h a t t h e absolute v a l u e s of t h e components of t h e t e n s o r $ I a r e proper f r a c t i o n s . This permits u s t o neglect, i n t h e equations of motion of t h e elements of an e l a s t i c body, t h e nonlinear terms w i t h an index2_qf homogeneity g r e a t e r than two, r e l a t i v e t o t h e components o f t h e t e n s o r 3.
A l l t h e remaining cases a r e strongly nonlinear. t i o n i s a r b i t r a q and i n c e r t a i n cases inapplicable. cases l a t e r i n t h e t e x t .

Ilowever, t h i s c l a s s i f i c a W e w i l l d i s c u s s these

The r e l a t i o n s o f t h e above nonlinear theory r e l a t e t o weakly nonlinear problems. The question n a t u r a l l y arises as t o t h e limits of a p p l i c a b i l i t y of weakly nonlinear and strongly nonlinear theory. These l i d t s obviously depend on two groups o f a c t o r s .

The f i r s t group of f a c t o r s limits t h e a p p l i c a b i l i t y of t h e equations of t h e theory of e l a s t i c i t y by t h e physical p r o p e r t i e s of t h e m a t e r i a l : The deformation m u s t be so small t h a t Hookefs l a w (L.5a) or t h e Voigt-Murnaghan l a w (4.8) are s a t i s f i e d . These considerations emphasize t h e d e s i r a b i l i t y of maki n g use o f t h e equations of t h e weakly nonlinear theory and narrow t h e limits of a p p l i c a t i o n of t h a s t r o n g l y nonlinear theory. The second group of f a c t o r s has a kinematic meaning and r e s t r i c t s t h e app l i c a b i l i t y of t h e equations of t h e weakly nonlinear theory or, more exactly, f o r c e s u s t o v e r i f y t h e i r a p p l i c a b i l i t y i n solving s p e c i f i c problems. I n f a c t , t h e s e equations were formally s e t up on t h e b a s i s o f expansions according t o

t h e degree of homogeneity of t h e terms containing components of t h e t e n s o r Qi. I n t h e equations were r e t a i n e d only t h e terms which, on appli22tion of Hookers l a w (4.5a), w i l l contain terms l i n e a r i n t h e t e n s o r components Q, o r t h e i r der i v a t i v e s and squares. LE
A simple example w i l l show t h a t this procedure does n o t always l e a d t o success. L e t u s consider t h e equation which i s familiar from courses on t h e s t r e n g t h of materials:

2-4

E/ y" (1 +y")a

=M-

(a)

Applying t h e above procedure, we f i n d

EJy" = M ,
i.e., here we might reach t h e erroneous conclusion t h a t t h e equations w e r e weakly nonlinear and coincided with t h e l i n e a r theory.

(b 1

These e r r o r s might have2keen avoided i f we had noted t h a t yr2 i s a component of t h e s t r a i n t e n s o r D i n which, i n this case, only t h e term containing y f 2 does not vanish. Thus, t h e expansion considered by u s may b e simplified by replacing t h e tensor D , , by t h e t e n s o r E l k , i f t h e components o f c i k are nonvanishing. If some t e n s o r components cik in some s p e c i f i c problem do vanish, then, i n t h e corresponding t e n s o r components D i k , we must r e t a i n t h e terms of higher o r d e r and eliminate them from t h e equations only a f t e r an a d d i t i o n a l a n a l y s i s making allowance f o r t h e s p e c i a l f e a t u r e s of t h e problem.

All above statements l e a d t o t h e conclusion t h a t it i s expedient separa t e l y t o consider t h e q u a n t i t i e s c h a r a c t e r i z i n g nonlinear deformations "as a wholeW$.
Section 9. I n t e r n a l and Ecternal Nonlinear Problems The mechanics of deformable bodies comprises two fundamental problems which we s h a l l c a l l f t i n t e r n d f fand "externaltt. The i n t e r n a l problem i s t o determine t h e s t r e s s e d and s t r a i n e d states of elements of a moving body. The e x t e r n a l problem i s t o d e s c r i b e t h e motion of t h e set of elements-of a-body or of t h e body ?'as a whole" r e l a t i v e t o a system of N e r i a n coordinates. The i n t e r n a l state of t h e elements of a body i s determined by t h e stress,
++ C f (BibLllb

pp.8 0-96

. .

) and a l s o I .Gekkeler, S t a t i c s of an E l a s t i c Body, ONTI, 1934,

68

s t r a i n , and e l a s t i c tensors. It i s n a t u r a l t o apply here t h e Lagrangian coordinates. The external problem i s solved a f t e r determination of t h e ' d i s p l a c e i , which on t h e b a s i s of eqs.(l.l) (1.2) permits e s t a b l i s h i n g t h e ment v e c t o r i configuration of t h e deformed body a t a r b i t r a r y time.

I n V.V .Novozhilovt s monograph ( B i b L l l b ), f o u r groups of problems of nonlinear e l a s t i c i t y 'theory a r e defined. I n accordance with h i s conclusions, we will h e r e a f t e r focus our a t t e n t i o n on problems physically l i n e a r b u t geometr i c a l l y nonlinear, s i n c e it i s p r e c i s e l y this group of problems t h a t i s c l o s e s t t o t h e nonlinear problems of t h e theory of e l a s t i c shells:;.
The d i v i s i o n o f t h e general problem of t h e mechanics of e l a s t i c deformable bodies i n t o an e x t e r n a l and an i n t e r n a l problem w a s q u i t e f u l l y accomplished by Kirchhoff and Clebsch i n t,he s t a t i c s of t h i n rods. They found t h a t t h e i n t e r n a l problem of t h e s t a t i c s o f t h i n rods i s l i n e a r . The s t r e s s e d and s t r a i n e d state of t h e elements o f a t h i n rod w a s described, perhaps i n first approximation b u t with s u f f i c i e n t accuracy, by t h e s o l u t i o n of t h e well-known Saint-Venant problem with indeterminate parameters, depending on t h e s o l u t i o n of t h e e x t e r n a l problem. The s o l u t i o n of t h e extern a l problem required t h e i n t e g r a t i o n of systems of nonlinear d i f f e r e n t i a l equat i o n s analogous t o those known from t h e dynamics o f a s o l i d body. The above-given equations of nonlinear e l a s t i c i t y theory do n o t permit a complete separation of t h e e x t e r n a l and i n t e r n a l problems, although t h e use of t h e coordinates xi i n s t e a d of W e r i a n Cartesian coordinates makes it p o s s i b l e t o advance considerably i n t h i s d i r e c t i o n . Considering t h e q u a i - c i t i e s e n t e r i n g i n t o t h e equation of t h e theory of e l a s t i c i t y , we note t h a t t h e i n t e r n a l deformed'.state of t h e elements o f a body must, according t o Kirchhoff and Clebsch, be described by t h e t e n s o r o f s m a l l deformations e i k . Among t h e q u a n t i t i e s d e f i n i n g t h e s t a t e of t h e body "as a whole?', t h e components o f t h e antisymmetric t e n s o r R i k , expressed by eqS.(2.7) and t h e q u a n t i t i e s I!,] must be included. A s noted i n (I, Sect.9), t h e C h r i s t o f f e l symbols have a nature simil a r t o t h a t of t h e t e n s o r of instantaneous angular v e l o c i t y o f a s o l i d body or t o t h a t of t h e v e c t o r o f angular v e l o c i t y of r o t a t i o n of a n a t u r a l trihedron, considered i n t h e theory of t h i n rods. 'vJe r e c a l l t h a t t h i s vector, according t o t h e Kirchhoff-Clebsch theory, satisfies equations of equilibrium analogous t o t h e dynamic N e r equation determining t h e motion of a r i g i d body about a fixed point. Turning t o t h e equations of nonlinear e l a s t i c i t y theory, we note t h a t t h e d i f f i c u l t i e s a r i s i n g i n t h e general s o l u t i o n o f t h e question o f subdividing t h e problem o f dynamic deformation of an e l a s t i c body i n t o an external problem and an i n t e r n a l problem c o n s i s t s i n a n a l y t i c a l l y expressing a generalized Hooke's law, r e l a t i n g t h e s t r e s s t e n s o r components t o t h e components of t h e f i n i t e deformation t e n s o r and t o t h e a n a l y t i c , e x p r e s s i o n s f o r t h e l a t t e r . It follows
: ; Cf.

E,]i

(Bibl.llb,

pp.125-126).

111

11.1.11111

, . I

, I . .

I . I 111 I

I.

1111.11.1-1111

_ I . . .

........

..

from eqs.(2.6), (2,7), and (2.11) t h a t t h e t e n s o r components D , , sented i n t h e following Yorm:

can be repre-

L24

I f , in accordance with VDV&"zhilov, we consider t h e case of q u a n t i t i e s c i k which are s m a l l i n comparison w i t h u n i t y , then w e o b t a i n approximately+

It Will be c l e a r from e q ~ ~ ( 9 . 1and ) (9.2) t h a t D,, and, consequently, t h e components of t h e s t r e s s t e n s o r 01, contain q u a n t i t i e s r e l a t i n g t o both t h e ext e r n a l and i n t e r n a l problems.
Considering t h e equations of t h e norLLinear theory of e l a s t i c i t y , w e may s t a t e t h a t only t h e conditions of c o m p a t i b i l i t y (3.1) belong exclusively t o t h e e x t e r n a l problem. There i s a resemblance between conditions (3.1) and t h e equ-ations of equil i b r i u m o f t h i n rods. This resid-es i n t h e f a c t t h a t t h e d e r i v a t i v e s o f t h e 3 . 1 ) , a r e analogous t o t h e C h r i s t o f f e l symbols, e n t e r i n g i n t o t h e conditions ( d e r i v a t i v e s o f t h e instantaneous angular v e l o c i t y components of a n a t u r a l trihedron o f t h e axis of t h e rod, which e n t e r i n t o t h e equ.ations o f equilibrium of t h i n rods. The d i f f e r e n c e i s t h a t t h e conditions o f compatibility do not have a k i n e t i c but o n l y a . kinematic meaning. Section 1 0 . Extension of t h e Kinematic R e l a t i o n s of t h e KirchhoffClebsch Thin-Rod Theory t o S h e l l Theory
~~ ~

The theory of t h i n s h e l l s proposed by Kirchhoff and Clebsch i s based on t h e kinematic r e l a t i o n s r e f e r r i n g t o simplified assumntions r e l a t e d with cert a i n concepts on t h e deformation of beams. W e w i l l show t h a t t h e kinematic r e l a t i o n s o f the theory of t h i n rods can b e generalized i n t o t h e three-dimensional problems of t h e theory of e l a s t i c i t y , and first of a l l i n t o t h e problems of t h e theory of shells-x-2. Let us consider, i n t h e deformed body, t h e point and t h e l o c a l coord i n a t e b a s i s associated with it. Let u s superpose on t h i s b a s i s t h e axis of a

M(2)

+> Zquation (9.2) corresponds t o formulas (1, l l l ) . o f t h e a u t h o r f s book (Bib1 I l b )

. .

$:--:t W e

have been guided by t h e e x p o s i t i o n , o f t h e Kirchhoff-Clebsch theory i n Gekk e l e r f s book " S t a t i c s of an n a s t i c Body", OJ!JTI, 1934, Section 3Oa-!+, using it i n t h e g e n e r a l i z a t i o n of t h e a n a l y t i c apparatus of t e n s o r analysis.
70

fixed Cartesian (Eulerian) system of coordinates y,. The neighborhood of t h e point M i s thus determined by t h e coordinates d which a r e functions of t h e coo r d i n a t e s xi. The coordinates d vanish a t point M. Let

Let yk) be a r a d i u s vector drawn from p o i n t PI(&) be t h e displacement o f point N r e l a t i v e t o point M:

;(d,

t o point

N(4,d).
L22

Clearly,

-. v (t, x i ; 0)= 0;

d v ( t , x i ; 0)
.-

-.

ax;-

- 0.
-

It goes withou-t saying t h a t 2q.(a) h a s a meaning i n a c u r v i l i n e a r coordin a t e system only when t h e v e c t o r u ( t , y: ) undergoes displacement p a r a l l e l , i n t h e sense of Levi-Civita, t o point 14.

2,

\le introduce t h e v e c t o r

and i n v e s t i g a t e t h e v a r i a t i o n of t h i s v e c t o r r e l a t i v e t o a moving coordinate base, with t h e o r i g i n being displaced along t h e coordinate l i n e

A network of t h e l o c a l coordinate system yi i s associated with t h e moving base. When t h e base moves through t h e p o i n t s 1.1 and Pi, which a r e fixed i n space, t h e p o i n t s o f t h i s network w i l l be continuously displaced, so t h a t t h e p o i n t s 1 4 and N a r e i n motion r e l a t i v e t o t h e system of coordinates yi with i t s o r i g i n a t t h e fixed point M*($ ).

$.

L e t us assume, f o r d e f i n i t e n e s s , t h a t t h e motion o f t h e coordinate base i s determined by t h e r e l a t i o n s

where t h e index k i s fixed. o r d i n a t e base. Consider t h e v e c t o r

yJ .will be

The parameter sk determines t h e motion of t h e co-

2.

I t s components i n t h e moving system of coordinates

d = ~ j ( t , xi x k ; yi, yk).

(10.3)

7 1

As

for

the vector

$W and MTN, we determine i t s components by t h e equation

; ,

r e p r e s e n t i n g i t s decomposition i n t o t h e components

W e emphasize t h a t eq.(lO.L) The sign P i n d i c a t e s invariance of eq.(lQ.4). must be regarded as t h e d e f i n i t i o n of v e c t o r r, n o t s u b j e c t t o proof, but corresponding i n s t e a d t o elementaqy geometrical concepts. /76

It follows from t h e d e f i n i t i o n of t h i s v e c t o r W e r e v e r t t o t h e v e c t o r rd. t h a t i t s absolute d e r i v a t i v e w i t h respect t o t h e v a r i a b l e & vanishe+: dRi dsk

Further, we f i n d

{;&}ARj = 0.

Consequently,

Noting t h a t it follows from eqs.(l0.2),

(10.3) t h a t

we f i n d f i n a l l y

These r e l a t i o n s a r e an extension o f t h e Kirchhoff-Clebsch kinematic equat i o n s t o t h e three-dimensional problems of e l a s t i c i t y theory, The C h r i s t o f f e l symbols {jk] a s already noted, a r e i n t h i s case associated w i t h t h e known

$5

W e r e c a l l t h a t t h e p o i n t of attachment o f t h e v e c t o r 72

fi

i s t h e point M.

q u a n t i t i e s p, q, r of t h e theory of t h i n rod+. According t o t h e Kirchhoff-Clebsch theory,_;the s t r e s s e d state of t h e elements o f a rod i s determined n o t by t h e v e c t o r u of absolute displacement of an element of t h e rod b u t by t h e v e c t o r of r e l a t i v e displacement v. The Y n t e r n a l t t problem i s solved in t h e components of t h e v e c t o r

; .

Assume t h a t t h e s t r e s s e d state of an element of a s h e l l i s also determined by t h e v e c t o r 7. W e then represent eq.(10.7b) i n t h e following form:

(10.8)

W e also assume t h a t t h e base area of a s h e l l can be chosen such t h a t t h e guant i t i e s V p vi s h a l l b e smll. Then, we f i n d i n first approximation

These r e l z t i o n s permit us t o d e r i v e t h e v e c t o r components vi and t h u s t o solve t h e i n t e r n a l problem. The e x t e r n a l problem i s solved by applying t h e equations of motion. Other a p p l i c a t i o n s of e q ~ ~ ( l 0 . 8 are ) a l s o possible. For example, it f o l -

lows from t h i s equation t h a t ( 1 0 . 1 0 )

I f , as a r e s u l t of t h e smallness of eq.(10.10) t a k e s t h e form
h i ,

we n e g l e c t t h e term Pi j '.vJ,
avi

then

V k u i ( t , x i ; y ~ ) = v ~ ~ i (Of) +, T f ( j i b } M y d dY0

( 1 0 . 1 1)

Equations (10.10) and (10.11) permit u s t o develop a method of a p p l i c a t i o n o f three-dimensional problems of t h e theory of e l a s t i c i t y t o two-dimensional problems, d i f f e r e n t from t h e methods known a t present.

I f we choose t h e base area such t h a t t h e covariant d e r i v a t i v e s such as

+e Compare with pp.86, 87 of t h e above-cited book bg 1.GekJseler " S t a t i c s of an m a s t i c Body", ONTI, 1934.

73

vt}

ui(t, ; 0) o r V k u i (t, 4 ; 0 ) are s u f f i c i e n t l y small in absolute value and t h e y1 a r e also s u f f i c i e n t 1 7 small, then eqs.(10.10) - (10.11) permit a l i n e a r i z a t i o n of t h e expression f o r t h e t e n s o r components D,, and a l i n e a r i z a t i o n of t h e above-derived equations f o r t h e n o n l i n e a r theory of e l a s t i c i t y .

Section 1 1 , P o t e n t i a l h e r g y of Deformation and Kinetic r " t h e E l a s t i c Body Zner,o;y o Without dwelling on t h e well-known conditions o f t h e existence of potent i a l deformation energy as a fvnction of t h e s t r a i n - t e n s o r components, we wish t o s t a t e t h a t , in adopting f o r pheriomenological considerations t h e ilooke-VoigtIhrnaghan lad i n t h e form of eq.(L.e) o r i n t h e more general form o f eq.(b.7), w e i m p l i c i t l y assumed t h a t t h e a3ove-mentioned conditions oi existence were satisfied-:?. The elementary work of deformation h a s t h e following form:

( 1 1 . 1 )

where 1 7 i s t h e volume o f t h e deformed body. The P f a f f form d k 6 D i k i s i n t e - , & The conditions of i n t e g r a b i l i t y a r e satisg a b l e i f eqs. (I,. 7) a r e s a t i s f i e d . f i e d by t h e symmetry p r o p e r t i e s of t h e e l a s t i c t e n s o r s C:kJ8 and C$kj*q*rs, i n d i cated i n 3ect.L.2. Hereafter we shall make use o f t h e l i n e a r Iiookers l a w . t h i s case t h e P f a f f form oik6D,, w e find

Integrating i n

I n t h e more general case,

and here

$ :

Nore d e t a i l s on t h e conditions o f existence of p o t e n t i a l deformation energy as a function of t h e components of t h e t e n s o r D , , w i l l be found i n A.Love*s book "Theory o f ELasticity", and a l s o in ( B i b L l l b ) .

m u a t i o n s (11.2a) ( 1 1 . 2 b ) determine t h e p o t e n t i a l energy of deformat i o n A. The r e s u l t a n t expressions a r e i n v a r i a n t under p o i n t transformations o f t h e coordinates and determine A i n an a r b i t r a r y c u r v i l i n e a r coordinate system. The k i n e t i c energy of an e l a s t i c body i s determined by t h e following equation :

The element of volume d V o f t h e deformed body i s connected with t h e element of volume dV, of t h e body b e f o r e deformation by t h e r e l a t i o n r e s u l t i n g from eq. (6.1):

d V = ( l +2gikDik+.
The limits of i n t e g r a t i o n i n e q s . ( l l 0 2 a ) components of t h e s t r a i n tensor.

. .)dl/,; -

(11 4 1

(11.3) likewise depend on t h e

Section 12. Work and Reciprocity Theorem in Nonlinear E l a s t i c i t y Theory The theorem of r e c i p r o c a l work i n t h e l i n e a r theory of e l a s t i c i t y i s a consequence of t h e i d e n t i t y of two i n v a r i a n t s associated with two states of t h e body : A - =ik ' - aflk 0 'Ik 'ik(12.1) The i d e n t i t y (12.1) r e s u l t s from Hookers l a w ( & . l a ) on replacing t h e tens o r Dr, by t h e t e n s o r of small deformations E,,

W e have i n d i c a t e d elsewhere ( B i b l . 23c ) general considerations p e r m i t t i n g m v a r i o u s g e n e r a l i z a t i o n s of t h e work and r e c i p r o c i t y theorem t o be found. I n Drder t o f i n d a g e n e r a l i z a t i o n of t h e r e c i p r o c a l theorem t o t h e nonlinear elast i c i t y theory, l e t us make use of eq.(4.7), represented i n t h e following form:

where

W e s h a l l c a l l t h e stresses

o & k

reduced s t r e s s e s .

Consider t h e i n v a r i a n t

75

I n t h e 'case of small deformations, t h e i n v a r i a n t A,, invariant A , . From eqs. (12.2a) follows the. i d e n t i t y :

passes over i n t o t h e

or

It W i l l be c l e a r from e q s . ( l l 0 2 a ) t h a t eq.(12,4) expresses t h e property o f r e c i p r o c i t y of t h e work done by t h e reduced stresses of one state of a n elastic body on t h e s t r a i n s of t h e o t h e r s t a t e . This work may be r e f e r r e d , f o r example, t o u n i t volume of t h e undeformed body. W e emphasize t h a t t h e reference of t h e s c a l a r AI, t o u n i t volume of t h e undeformed body i s a r b i t r a r y . I n e x a c t l y t h e same m y , one might use t h e u n i t volume o f t h e body i n t h e first o r second state. To this a r b i t r a r y choice c o r - . respond t h r e e p o s s i b l e i n t e g r a l statements of t h e generalized r e c i p r o c a l theorem, Multiplying eq.(l2.4) by t h e volume element dV, of t h e undeformed body and i n t e g r a t i n g over t h e volume V,, w e find

(12.5

Consider t h e i n t e g r a l

W e have

Further,

and, applying t h e Ostrogradskiy-Gauss

formula, we f i n d

(12.8)

i s t h e u n i t v e c t o r of t h e e x t e r n a l normal t o t h e surface S, of t h e unwhere deformed body.

Using eqso(5.2a) and t h e r e l a t i o n (6.15a), we get

-e

The v e c t o r may be regarded a s t h e f o r c e r e l a t e d t o u n i t volume of t h e undeformed body. I n t h i s case, however, i t must n o t be forgotten t h a t a l l t h e k i n e t i c q u a n t i t i e s i n eq.(12.9) a r e connected with an element o f volume o f t h e deformed body

O n t h e b a s i s of eqs.(12.6)

- (12.9),

eq0(12.5) t a k e s t h e following form:

Equation (12.10) may be regarded as t h e f o r m a l generalization of t h e theorem of work and r e c i p r o c i t y o f t h e l i n e a r theory o f e l a s t i c i t y t o t h e problem of t h e mechanics of a n i s o t r o p i c e l a s t i c bodies with physical and geometrical no;ilinearity. In f a c t eq.(12.10), a t s m a l l deformations and i n t h e absence o f i n e r t i a l forces, y i e l d s t h e c l a s s i c a l theorem of work and r e c i p r o c i t y . Ne note i n conclusion t h a t t h e use of o t h e r methods f o r s e l e c t i n g t h e init i a l i n v a r i a n t A,,, would y i e l d o t h e r i n t e g r a l equations which would l i k e w i s e generalize, i n t h e above sense, t h e r e c i p r o c a l theorem of t h e l i n e a r theory of e l a s t i c i t y . A l l t h e s e g e n e r a l i z a t i o n s do not l i t e r a l l y comespond t o t h e classical theorem s i n c e they contain q u a n t i t i e s which a r e only by convention termed and %odglt forces. The question of t h e p o s s i b i l i t y of proving by u s the r e c i p r o c a l theorem, f r e e of t h e s e a r b i t r a r y elements, s t i l l remains open. S e c t i o n 13. n a s t i c Medium with I n i t i a l S t r e s s e s
I r ,m a n y c a s e s it i s necessary t o i n v e s t i g a t e t h e deformation of elements of an elastiLc body i n an a l r e a d y e s t a b l i s h e d s t r e s s e d state.

/81

The book by A.Love

contains examples of cases i n which t h e i n i t i a l s t r e s s e s

77

cannot be neglected.

One of t h e s e examples i s taken from t h e theory of s h e l l g t

The modern p r a c t i c e of designing reinforced concrete and steel s t r u c t u r e s w i t h prestressed elements likewise f u r n i s h e s numerous examples from t h e f i e l d o f mechanics i n v e s t i g a t e d in t h e present study and demonstrates t h e n e c e s s i t y of formulating a general theory permitting a s u f f i c i e n t l y rigorous mathematical analysis of t h e s e problems and similar ones. The question of t h e r e l a t i o n s between t h e components of t h e t e n s o r o f a d d i t i o n a l s t r e s s e s and t h e t e n s o r of a d d i t i o n a l deformations was posed long ago

Love i n d i c a t e s t h a t , t o e s t a b l i s h t h e r e l a t i o n s between t h e a d d i t i o n a l s t r e s s e s and s t r a i n s , we must t u r n t o a more general theory than t h a t developed i n t h e c l a s s i c a l mechanics of e l a s t i c bodies, o r t o p r a c t i c a l experiments. The outmoded theory of Cauchy and Green i s obviously i n s u f f i c i e n t l y substantiated++. W e present below t h e r e l a t i o n s between t h e a d d i t i o n a l s t r e s s e s and t h e a d d i t i o n a l s t r a i n s r e s u l t i n g from r e l a t i o n s (,!+.la) containing geometrically nonl i n e a r terms. W e s h a l l assume t h a t t h e r e exists an initial undeformed state of t h e body, The body i s then deformed and t h e i n i t i a l s t r e s s e s C T A ~ and displacements k i appear and a r e r e l a t e d by eqs. ( L o l a ) :

( 1 3 . 1)
Equation (13.1) corresponds t o t h e mechanical methods of e s t a b l i s h i n g i n i t i a l stresses. If t h e i n i t i a l s t r e s s e s a r e due t o thermal effects,eqs. ( 1 3 . 1 ) must be supplemented by temperature-dependent terms. Consider c e r t a i n consequences r e s u l t i n g from eqs. (13.1). A s a r e s u l t of t h e a d d i t i o n a l s t r a i n , l e t new s t r e s s e s and displacements a r i s e , connected with t h e i r i n i t i a l values by t h e r e l a t i o n s

(13.2)
It i s here assumed t h a t t h e components v, are s m a l l i n absolute value, L e . , /82 small in r e l a t i o n t o a c e r t a i n c h a r a c t e r i s t i c measurement of t h e body. In t h e s h e l l theory, such a q u a n t i t y i s t h e thickness of t h e s h e l l .
S u b s t i t u t i n g e q ~ ~ ( 1 3 . 2i ) n t o eqs.(l+.la) we obtain, a f t e r a n d e r of transformations and d i s c a r d i n g t h e terms t h a t are nonlinear i n v , and t h e derivat i v e s of vi :
~

$F

Cf. A.Love,

1935, pp.120

- 122.

Mathematical Theory of E l a s t i c i t y (Russian T r a n s l a t i o n ) ONTI,

78

(13.3
Symmetrizing t h e left-hand we f i n d
Tik

s i d e of this equation in t h e i n d i c e s r and s,

1 ( =, +
0

rs

+c f k .jsvju; +c i k n/ rv j d ) ( v s v r

vrvJ

(13.41

Let

where PI i s t h e t e n s o r of small a d d i t i o n a l deformations. t a k e on t h e form of a generalized Hookers l a w :

Then, eqs.(l3.4)

The q u a n t i t i e s xikr may be considered as being components of t h e e l a s t i c tens o r i n a body w i t h i n i t i a l s t r e s s e s .

I n eq.(13.5) w e r e p l a c e t h e t e n s o r componentsVJG by t h e i r expressions i n terms o f t h e t e n s o r components eik and Q , , r e s u l t i n g from eqs.(2.6) - (2.7). W e then f i n d

u , or

It i s c l e a r from eqs.(13.5) and (U9k3) t h a t , under largc2aisplacements of of components o f t h e s t r a i n t e n s o r e and of t h e t e n s o r s1 connected with r o t a t i o n s of elements of t h e body, a p r e s t r e s s e d body on subsequent deformat i o n s must be regarded a s an inhomogeneous body with v a r i e d anisotropy. I n p a r t i c u l a r , an i s o t r o p i c body i s converted i n t o an a n i s o t r o p i c body. These f a c t s a r e a l s o known from geometrical o p t i c s , b u t i n solving t h e problems of t h e mechanics of e l a s t i c bodies they a r e of s u b s t a n t i a l importance only in d i s i placements of $_high $9 modulus, f o r g r e a t absolute v a l u e s of t h e components of t h e t e n s o r s e and 0. A l l above statements a l s o apply t o physically nonl i n e a r e l a s t i c bodies, f o r which t h e r e l a t i o n s (4.7) a r e valid.

79

CHAPTER I11
REDUCTION OF THE THREE-D7MEI\ISIOI'JALPROBLEM OF THE MECHANICS OF ELASTIC B O D I S TO THE TVJO-DDENSIONAL PROBLENS OF THE THEORY OF SHELLS
Section 1 . General Characterization o f t h e Problem

/83

The s o l u t i o n of t h e three-dimensional problems o f e l a s t i c i t y theory involves considerable mathematical d i f f i c u l t i e s . For this reason, long ago, duri n g t h e very development of t h e methods for solving t h e problems of e l a s t i c i t j theory, t w o groups of problems were distinguished, permitting t h e s u b s t i t u t i o n of systems o f elastodynamic equations bg systems of a p p r o a t e equations cont a i n i n g a smaller nmiber o f independent v a r i a b l e s than t h e o r i g i n a l equations. This decrease i n t h e number of independent v a r i a b l e s i s equivalent t o decreasi n g t h e number of dimensions of space, s i n c e t h e independent v a r i a b l e s i n t h e equations of e l a s t i c i t y theory a r e t h e space coordinates and time. The two mentioned groups of problems a r e t h e problems of t h e motion of elements of t h i n e l a s t i c rods and those of t h e dynamically s t r e s s e d and s t r a i n e d s t a t e s of s h e l l s .

In t h e former case, t h e r e l a t i o n s of two s p a t i a l measurements o f t h e body t o t h e t h i r d dimension a r e n e g l i g i b l e , s o t h a t t h e three-dimensional problem of t h e e l a s t i c i t y theory can be reduced t o a one-dimensional problem.
I n t h e s h e l l theory, it i s assumed t h a t t h e r a t i o of one of t h e dimensions of t h e body t h e thickness of t h e s h e l l t o t h e o t h e r dimensions i s small. Then, as w e s h a l l show, t h e three-dimensional problem of t h e t h e o q of e l a s t i c i t y can be approximately redu-ced t o ei two-dimensional problem.

The r a t i o of t h e thickness of a s h e l l t o one of t h e c h a r a c t e r i s t i c paramet e r s determining t h e dimensions o f t h e s h e l l i s l i m i t e d by various conditions i n conventional s t u d i e s of t h e subject matter. These conditions depend primar i l y on t h e accuracy of t h e approximate r e p r e s e n t a t i o n o t h e three-dimensional dynamic boundary problems of t h e two-dimensional e l a s t i c i t y theory. It i s obvious t h a t t h e boundary conditions of t h e problem a r e of g r e a t importance here. It i s t h e r e f o r e impossible t o set up any general absolute c r i t e r i o n which t h e & / thickness of a s h e l l must s a t i s f y , t o ensure a predetermined accuracy i n t h e E s o l u t i o n of equations approximately d e s c r i b i n g i t s s t a t e . W e give below a b r i e f survey o f t h e present methods o f c l a s s i f y i n g s h e l l s according t o t h e i r thickness. This c l a s s i f i c a t i o n a l s o involves concepts on t h e limits of a p p l i c a b i l i t y o f various methods o f a n a l y t i c a l d e s c r i p t i o n of t h e dynamically s t r e s s e d and s t r a i n e d state of a s h e l l . 'de w i l l , make some preliminary remarks::- on t h e general problem o f reducing
-%

M e r e c a l l t h a t t h e r e a r e two-dimensional problems i n t h e e l a s t i c i t y t h e o r i e s t h a t are not connected with s h e l l theory.

80

t h e three-dimensional problem of t h e e l a s t i c i t y theory t o a two-dimensional problem, under t h e assumption t h a t t h e s h e l l i s an e l a s t i c body. Let u s s e l e c t on t h e base surface of an undefsmed s h e l l an a r b i t r a r y coordinate system xi (i= = 1 , 2 ) . The coordinate v e c t o r e, i n t h e undeformed s h e l l will be_;taken, according t o (I, Sect.3) as equal t o t h e unit v e c t o r of t h e normal n t o t h e base area. The v e c t o r s = 1 , 2, 3 ) form t h e l o c a l coordinate base. W e agree t h a t t h e mutual o r i e n t a t i o n of t h e s e v e c t o r s corresponds t o a right-hand coord i n a t e system.

zi(l

The general program of reduction of t h e three-dimensional problem of t h e theory of e l a s t i c i t y t o a two-dimensional problem c o n s i s t s i n constructing anal y t i c expressions f o r t h e q u a n t i t i e s c h a r a c t e r i z i n g t h e s t r e s s e d and s t r a i n e d s t a t e of t h e s h e l l i n terms of new q u a n t i t i e s determined in t h e coordinates xi (i = 1 , 2) of i t s base area, and i n s e t t i n g up t h e equations t h a t t h e s e quant i t i e s must s a t i s f y in t h e region of v a r i a t i o n of t h e v a r i a b l e s xi and on t h e i r boundaries

The equations of s h e l l theory might b e s a i d t o d e s c r i b e t h e s t r e s s e d and s t r a i n e d s t a t e of t h e base area. Obviously, t h e s e equations m u s t n o t contain d e r i v a t i v e s with r e s p e c t t o t h e coordinate 2.
As w e s h a l l show l a t e r , t h e "reduction problem" has no unique solutior+. But t h e solution, on t h e o t h e r hand, cannot be completely a r b i t r a r y . It i s res t r i c t e d by t h e requirements of optimum appr'oximate r e p r e s e n t a t i o n of t h e equaIt i s w e l l known t i o n s of e l a s t i c i t y theory by t h e equations of s h e l l theory. from t h e theory of approximation functions t h a t t h e concept "optbum approximation" i s not e n t i r e l y d e f i n i t e . For example, t h e r e exist optimum approximat i o n s a t a given p o i n t of a manifold t o which an approximation function i s assigned, optimum approximations i n t h e mean i n a c e r t a i n region of v a r i a t i o n of i t s arguments, etc.. To d i f f e r e n t methods of approximation functions t h e r e COP respond d i f f e r e n t methods of approximate reduction of t h e three-dimensional /s5 problems of e l a s t i c i t y theory t o two-dimensional problems.

The s o l u t i o n of t h e reduction problem depends l a r g e l y on t h e choice of t h e approximation method f o r t h e components of t h e stress t e n s o r o r D , , of t h e s t r a i n tensor, considered a s f u n c t i o n s of t h e coordinate 2. Section 2. Remarks on t h e Methods of Reduction given by Poisson, Cauchy, Kirchhoff, and Love General methods f o r t h e reduction of a t h r e e d i m e n s i o n a l s t a t i c problem of t h e e l a s t i c i t y theory t o a two-dimensional problem were developed by Poisson and Cauchy, i n considering t h e equilibrium of a p l a t e . These g r e a t mathematic i a n s applied t h e expansion of t h e s t r e s s t e n s o r components in ascending p o w e r 9 of t h e coordinate z , measured along a normal t o t h e undeformed middle plane of t h e plate. Using t h e equations of equilibrium of an element of a continuous medium, and assuming t h a t t h e boundary planes of t h e p l a t e w e r e free of loads, Cauchy and Poisson obtained a fundamental s y s t e m of equations and boundary con1 The Cauchyd i t i o n s of t h e boundary problem f o r t h e equilibrium of a plate.
%

Here and h e r e a f t e r , f o r b r e v i t y , we w i l l use t h e term rtreduction problemrt.

81

. -.

Poisson method w a s subjected t o a c r i t i c a l a n a i y s i s by Saint-Venant, Kirchhoff, md s e v e r a l o t h e r i n v e s t i g a t o r s . Saint-Venant noted t h a t it was n o t f u l l y just i f i e d t o expand stress t e n s o r components not known i n advance i n t o series i n powers of even t h e r e l a t i v e l y small coordinate Z . These s e r i e s , i n h i s opinion, might converge i n a s u f f i c i e n t l y small neighborhood of an i n t e r i o r p o i n t of t h e p l a t e , b u t t h e i r convergence over t h e e n t i r e range of v a r i a t i o n of t h e coordin a t e 2; might n o t t a k e place. H e r e f e r r e d i n this connection t o t h e i n a c c u r a t e results obtained by t h e Cauchy and Poisson methods i n t h e theory of t h e t o r s i o n of prisms. For this reasonb he p r e f e r r e d d i f f e r e n t methods of s e t t i n g up t h e fundamental system of equations of t h e theory of p l a t e s , including t h e Kirchhoff method, based on well-known simplifying hypotheses.+.

W e s h a l l n o t dwell here on a d i s c u s s i o n of t h e boundary conditions i n t h e theory of p l a t e s , s i n c e w e will r e v e r t t o this s u b j e c t later. Saint-Venantts o b j e c t i o n s t o t h e Cauchy and Poisson methods, t o a considerable degree, ref l e c t e d t h e s t a t e of t h e theory o f e l a s t i c i t y i n t h e t h i r d q u a r t e r of t h e l a s t Zentury. It i s known t h a t , a t t h a t time, only t h e foundations of t h e general s o l u t i o n methods f o r boundary problems of e l a s t i c i t y t h e o r i e s were being prepared, permitting conclusions from t h e a n a l y t i c a l p r o p e r t i e s of t h e s e solutions.
I n t h e l a s t q u a r t e r of t h e 1 9 t h Century, t h e work done by Somigliano, Volt e r r a , and L a u r i c e l l a led t o t h e conclusion t h a t , i n t h e absence of body forces, the s o l u t i o n of t h e problems of t h e s t a t i c s of an e l a s t i c body are a n a l y t i c /86 functions of t h e coordinates of internal p o i n t s of t h e body, i.e., t h e s e solut i o n s can be expanded i n series i n p o s i t i v e powers of t h e coordinates.:-.+. But t h e convergence of t h e s e expansions on t h e s u r f a c e of t h e body r e q u i r e s a separate i n v e s t i g a t i o n f o r s p e c i f i c problems. Moreover, t h e conclusions on t h e a n a l y t i c p r o p e r t i e s of a s o l u t i o n of s t a t i c problems can be extended t o t h e dynamics of e l a s t i c bodies only after a d d i t i o n a l analysis.

W e s h a l l now present f a c t s confirming t h e s i g n i f i c a n c e of t h e method under discussion. W e r e c a l l t h a t t h e method of t h e preliminary introduction of expansions i n s e r i e s , used by Poisson and Cauchy in t h e theory of p l a t e s , i s encountered even today i n v a r i o u s f i e l d s of t h e e l a s t i c i t y theory, f o r example in t h e s t a t i c plane problem. Here, on t h e S a s i s of t h e a n a l y t i c p r o p e r t i e s of sol u t i o n s of t h e plane problem, expansions i n power s e r i e s a r e introduced, and then t h e i r convergence over t h e e n t i r e region of determination o f t h e required functions, including t h e i r boundary, i s confimed i n s p e c i a l case$:-.:-.t. The method o f expansion of t h e required f u n c t i o n s in series in p o s i t i v e powers of t h e coordinate 2, as could have been predicted, w a s found t o be an e f f e c t i v e m e a n s of constructing t h e theory of t h i c k p l a t e s (BibLqb, 2 % ) .
Saint-Venantts c r i t i c i s m of t h e work of Poisson and Cauchy i s given by Clebsch i n t h e book Theory of E l a s t i c i t y of S o l i d Bodies, P a r i s , 1883, pp.722 725.

e Cf., p.135.
JL

f o r example, E.Trefftz,

Mathema-cical Theory of E l a s t i c i t y , ONTI,

1934,

, t + c ~

C f N.I.Muskhelishvili, Some Fundamental Problems of t h e Mathematical Theory of E l a s t i c i t y , USSR Academy of Sciences, 1949.

82

A-ll t h e above permits u s t o a s s e r t t h a t Saint-Venantts c r i t i c i s m of t h e Poisson and Cauchy methods i s not s u f f i c i e n t reason for a b s t a i n i n g from a furt h e r development of t h e s e methods with r e s p e c t t o t h e theory of s h e l l s . Proper caution must, however, b e exerted i n t h e s p e c i a l cases noted below.
The Kirchhoff theory w a s subsequently extended by A.Love t o include t h e theory of s h e l l s . The Kirchhoff-Love theory i s based on t h e well-known postul a t e t h a t t h e normal t o t h e undeformed middle surface of a s h e l l remains normal t o it even a f t e r deformation. This hypothesis i s supplemented by one of two hypotheses about t h e v a r i a t i o n i n i t s length.
According t o t h e f i r s t of these hypotheses, a segment of a normal t o t h e middle surface enclosed. within a s h e l l does n o t vary i t s l e n g t h &der deformat i o n of t h e s h e l l . I n t h i s case, t h e Kirchhoff-Love hypothesis i s appropria t e l y c a l l e d t h e %.ypothesis of s t r a i g h t constant normals**. According t o t h e second version it i s assumed t h a t t h e normal s t r e s s e s ~ 7 ~ a r e small by comparison with axirk 1 ( i , k = 1,2) and can be neglected+. W e /87 note t h a t t h e first hypothesis i s not equivalent t o t h e second, a f a c t which i s not made s u f f i c i e n t l y c l e a r i n c e r t a i n well-known monographs. The llhypothesis of an i n v a r i a n t normal" n a t u r a l l y l e a d s t o a replacement of t h e v e c t o r s of stresses a c t i n g a t t h e boundary of a s h e l l element with generat r i c e s normal t o t h e middle surface by t h e s t a t i c a l l y equivalent system of e emphaf o r c e s applied t o t h e contour of an element of t h e middle surface. W s i z e t h a t this s u b s t i t u t i o n of t h e a c t u a l s y s t e m of stresses by a system of f o r c e s and moments s t a t i c a l l y equivalent t o it i s i n t e r n a l l y i n harmony with t h e "hypothesis of s t r a i g h t and i n v a r i a n t normals*s~-+*. I n t h e remaining cases this agreement does not appear. The Kirchhoff-Love hypotheses introduce an unremovable e r r o r i n t o t h e equations of s h e l l theory, and t h i s e r r o r must be taken i n t o account i n evaluating t h e p o s s i b i l i t y of vari o u s s i m p l i f i c a t i o n s of t h e equations o f ' s h e l l theory. Section 3. Preliminary C l a s s i f i c a t i o n of S h e l l s Connected with t h e Kirchhoff-Love Hypotheses. Linear and Nonlinear Problems The work by V.Novozhilov and R.Finke1tshtey-n (Bibl.28) gives an estimate of t h e e r r o r introduced by t h e Kirchhoff-Love hypothesis i n t o t h e equations o f It w a s shown t h a t this e r r o r i s of t h e o r d e r of max(2hk, ), where s h e l l theory. k, i s one of t h e p r i n c i p a l curvatures of t h e s h e l l . This estimate permits u s t o d i s t i n g u i s h t h e c l a s s of s h e l l s f o r which t h e equations of c l a s s i c a l s h e l l theory based on t h e Kirchhoff-Love hypothesis are s t i l l s u f f i c i e n t l y accurate, permitting, f o r example, a determination of t h e i r s t r e s s f i e l d s with a r e l a t i v e error not exceeding 5%. These s h e l l s are c a l l e d t h i n . A l l o t h e r s h e l l s w i l l
-

Cf.(I,

Here are t h e "physical components**of t h e stress tensor. Sect.5) and ( I , Sect.7).

+t

w* Here and h e r e a f t e r we assume t h e reader t o b e f a m i l i a r with t h e theory of s h e l l s , f o r example t o t h e e x t e n t given i n (Bib1.lla). For this reason, we use c e r t a i n terms without f i r s t giving t h e i r d e f i n i t i o n .

83

here be c a l l e d t h i c k s h e l l s o r s h e l l s of medium thickness-. Let us d e f i n e t h e s e i d e a s using t h e d e f i n i t i o n s given i n s e v e r a l modern studies. W e i n t r o duce t h e n o t a t i o n (Bib1.1):

m a x ( 2 h R , ) = ~ ; max - =q,

('n") -

(3.1)

where cz i s one of t h e parameters determining t h e dimensions of t h e b a s i c surface of t h e s h e l l , The c l a s s i f i c a t i o n of s h e l l s i n t o t h i n and t h i c k i s r e l a t e d primarily t o t h e q u a n t i t y e. Most often, because o f t h e f a c t t h a t t h e r e l a t i v e e r r o r of t h e s o l u t i o n of approximate equations considered i n s h e l l theory i s r e s t r i c t e d t o 5%, a s h e l l i s c a l l e d t h i n i f t h e condition (Bib1.1, lla) /88 1 20 (3.2)

e<-

i s satisfied.
If t h e condition (3.2) i s not s a t i s f i e d , t h e s h e l l i s c a l l e d thick. Of course, t h e condition (3.2) i s somewhat a r b i t r a r y , s i n c e a rigorous estimate of t h e e r r o r of s o l u t i o n s of t h e equations of s h e l l theory, constructed on t h e b a s i s of simplifying assumptions with various boundary conditions, i s very d i f ficult.

The upper l i m i t of values of 1 has been i n s u f f i c i e n t l y studied. S.A.Ambartsumyan (Bibl.1) assumes i n h i s book t h a t TI I; 0.1, w h i l e A.S.Voltmir i n h i s 0.2. The q u a n t i t y can obviously not be d e t e r book (Bibl.4) s t a t e s t h a t 7 mined without a d e l i m i t a t i o n of t h e c l a s s of t h e boundary problems o f s h e l l theory. K.Z.Galimov and Kh.M.Mushtary i n t h e i r monograph (Bibl.10) have pointed o u t a d i f f e r e n t approach t o t h e c l a s s i f i c a t i o n of s h e l l s , obviously based on physical considerations. A s h e l l , according t o t h e s e authors (Bibl.lO), i s c a l l e d t h i n i f it satisfies t h e condition

(3.3)
Whereas, i f it satisfies t h e condition
ep
-~

< 2hf.-' < V E p

(3.4)

of evaluating t h e e r r o r introduced by t h e Kirchhoff-Love hypothesis, advanced by Novozhilov and Finkeltshteyn, has evoked c r i t i c a l remarks by V.M.Darevskiy (Bib1.22).

* The method

Here L i s a l i n e a r dimension, c h a r a c t e r i s t i c a s h e l l i s of medium thickness. f o r a s h e l l or a p l a t e , f o r example one of t h e p r i n c i p a l r a d i i of curvature of t h e b a s i c surface or i t s smallest diameter. The q u a n t i t y cP i s t h e r e l a t i v e elongation corresponding t o t h e proportional l i m i t of t h e m a t e r i a l of t h e s h e l l .
A number of i n v e s t i g a t o r s suggest t h a t no preliminary r e s t r i c t i o n s b e imposed on t h e thickness of t h e s h e l l , 'de s h a l l r e t u r n l a t e r t o t h i s question, when we base our own c l a s s i f i c a t i o n of s h e l l s , as a function of t h e i r thickness, on t h e theory or^ propagation of dynamic wave processes i n such s h e l l s .

It i s well known t h a t p l a t e s and s h e l l s a r e e l a s t i c bodies i n which t h e displacements, deformations, and angles of r o t a t i o n of t h e elements may be so g r e a t t h a t an a p p l i c a t i o n of t h e l i n e a r equations of t h e c l a s s i c a l theory of e l a s t i c i t y would l e a d t o s u b s t a n t i a l e r r o r s . In such cases, nonlinear equat i o n s must be used. I n determining t h e boundaries o f a p p l i c a b i l i t y of t h e l i n e a r theory, one u s u a l l y starts o u t from t h e r a t i o of t h e displacements of t h e p o i n t s of t h e b a s i c surface of t h e s h e l l t o i t s thickness. This method of c l a s s i f i c a t i o n , however, i s a r b i t r a r y , s i n c e here t o o t h e d e c i s i v e influence i s t h a t of t h e preassigned e r r o r limits i n t h e s o l u t i o n of t h e boundary conditions o f t h e theory of s h e l l s . It may be considered, according t o A.S.Vol"ir, that the linear & Q theory i s applicable if : 2h < 1/5, and i s e n t i r e l y i n a p p l i c a b l e i f : : 2h 2 5 (Bibl.4) where 3 i s t h e displacement v e c t o r of t h e b a s i c surface of the shell.

!<I

K.Z.Galimov and Kh.M.Nushtary give a d i f f e r e n t approach t o t h e c r i t e r i a of a p p l i c a b i l i t y o f t h e l i n e a r theory. They d i s t i n g u i s h weak, medium, and strong f l e x u r e s of t h e s h e l l .
A weak f l e x u r e t a k e s place i f t h e following condition i s s a t i s f i e d i n t h e

shell :

maxi1

< < 1,

(3.5)

where Iw'l i s t h e modulus o f t h e v e c t o r of r o t a t i o n of an a r b i t r a r y l i n e a r element under f l e x u r e of t h e s h e l l . It i s found here t h a t , f o r some c l a s s e s of boundary problems, t h e condition (3.5) i s s a t i s f i e d i f

max I W J=2h;

-P

(3.6a)

f o r o t h e r types of boundary problems, t h e condition (3.5) l e a d s t o t h e inequation

For a moderate flexure, t h e displacement of p o i n t s of t h e b a s i c surface, by modulus, equals or exceeds 2h but i s considerably smaller than t h e o t h e r l i n e a r
85

dimensions o f t h e s h e l l .

Here,

(3.7)
The r e l a t i o n (3.7) g i v e s u s t h e r i g h t t o n e g l e c t q u a n t i t i e s of t h e o r d e r of by comparison w i t h unity.

/.'I"

I n a strong flexure, t h e displacements, d i r e c t e d along t h e normal t o .the b a s i c surface-%, a r e g r e a t S e l a t i v e t o 2h and are of t h e order L. I n t h i s case, likewise, t h e q u a n t i t i e s I w l 2 will b e great. The problems r e l a t e d t o a weak f l e x u r e are described by linear systems of equations, while those of moderate and strong flexure belong t o t h e nonlinear theory of s h e l l s . If we confine ourselves t o t h e study of e l a s t i c deformations, then we m u s t impose on t h e problems of moderate and strong flexure a d d i t i o n a l r e s t r i c t i o n s t h a t ensure t h e absence of zones of p l a s t i c i t y . Finally, t h e s e r e s t r i c t i o n s must be connected with methods f o r t h e reduction of a threedimensional problem of t h e theory of e l a s t i c i t y t o a two-dimensional problem It i s d i f f i c u l t t o i n d i c a t e t h e s e reand with s p e c i f i c boundary conditions. s t r i c t i o n s i n t h e general form f o r extensive c l a s s e s of problems. Section

4. Application of Tensor Series. Reduction of t h e ThreeDimensional Problem t o t h e Determination o f an I n f i n i t e Sequence of Functions of a Point o f t h e Base Area of the Shell
_ _ _ ~
~~

m.

L e t us extend t h e methods applied by Cauchy and Poisson i n t h e s t a t i c s o f p l a t e s t o t h e elastodynamic problems of s h e l l theory. To b r i n g o u t t h e fundamental i d e a s of t h e method, l e t u s f i r s t consider t h e l i n e a r equations o f s h e l l theory, holding t o t h e exposition adopted by u s in o t h e r work (Bib1,23a, b). W e w i l l u s e expansions in t e n s o r series, which are generalized Poisson and Cauchy e emphasize t h a t , i n s t e a d of t e n s o r series, expansions in ordinary series. W power s e r i e s can a l s o b e used (Bib1,23a, b and 26). Ekpansions i n t e n s o r ser i e s , i n our opinion, have t h e following advantages:
a ) Such expansions l e a d t o equations v a l i d i n an arbitrary c u r v i l i n e a r syst e m of coordinates, which i s p a r t i c u l a r l y convenient in solving nonlinear problems;
b ) Each term in t h e approximation formulas i s a q u a n t i t y w i t h a d e f i n i t e geometrical meaning. The l a t t e r permits a c l e a r e r and more p i c t o r i a l presentat i o n of t h e meaning of various s i m p l i f i c a t i o n s of t h e equations than t h e use of conventional expansions i n powers of t h e coordinate 2. Hereafter, we w i l l consider s e v e r a l v e r s i o n s of t h e generalized expansions Assume, neglecting t h e i n power s e r i e s . L e t u s make use of eq.(I, 12.3). s t r a i n s in t h e s h e l l , t h a t
(&)I

-. 5 (Ar)2L -. 0;
86

-D

(Ar)3

z.

(4.1)

These displacements are o r d i n a r i l y c a l l e d flexural.

The sign 2 here and h e r e a f t e r , denotes an e q u a l i t y t h a t i s t r u e only i n one d e f i n i t e coordinate+system (a non-invariant equality). The expressions of t h e v e c t o r components A r i n t h e deformed s h e l l w i l l b e given below when we d i s c u s s t h e nonlinear theory;:.

Following t h e general program of s o l u t i o n of t h e reduction problem (Sect i o n 1), l e t us consider t h e expansions of t h e strain t e n s o r components i n t o t e n s o r series, remembering eq. (4.1) :

where D ( z ) are t h e components o f t h e s t r a i n t e n s o r a t t h e p o i n t with t h e coor= 8, while D , , a r e t h e s t r a i n t e n s o r components on t h e b a s i c surface dinate of t h e s h e l l .

W e note t h a t , according t o eq.(I,

1 2 . 3 ) ,

an expression of t h e form

7IZ t i m e s

i s t o be regarded as a component of a covariant t e n s o r of second rank, and a n a expression of t h e form

a s a component of a covariant vector.

Let u s assume t h a t t h e t e n s o r of curvature vanishes i n t h e Lagrangian system of coordinates x i associated with t h e undeformed or deformed s h e l l . Then, according t o eq.(I, 10.9), we have t h e r i g h t t o change, i n t h e multiple covaria n t d e r i v a t i v e s , t h e sequence of operations o f d i f f e r e n t i a t i o n , To simplify . 2 ; t h e calculations, l e t u s assume t h a t t h e system of coordinate l i n e s xi (i = 1 on t h e b a s i c surface coincides with i t s l i n e s of curvature. Using eqsD(I,3.6a.) - (3,6b), (9.5) and (9.S), we f i n d t h e nonzero C h r i s t o f f e l symbols of index 3. W e have

or, f o r z = 0,

Hereafter, t o shorten t h e forrmlas, we s h a l l denote t h e components o f t h e m e t r i c t e n s o r on t h e b a s i c s u r f a c e by g,, and a t an a r b i t r a r y p o i n t by gt;)

T

(Ar)P(Ar)9...

(3ne more remark: The f u n c t i o n s z' are components o f t h e t e n s o r From eqs.(LL.l) and (h.3) it follows t h a t

.+

Vi (Ar)3 = 0

( i s 1, 2).

Eut

i n s p i t e o f t h e first two r e l a t i o n s of eqs.(k.l). For t h i s reason, i n covaria n t d i f f e r e n t i a t i o n , t h e q u a n t i t i e s z c a n m t be regarded as constants. Bearing i n mind a l l t h e t has been said, we f i n d , in expanded fom, t h e ex? a s i o n s o f t h e t e n s o r components of s m a l l defomz.tion in t e n s o r s e r i e s i n powers of Z . >Je have

where

S u b s t i t u t i n g eqs. (L.5a) (L.5f) i n t o eqs.(II, L o l a ) or i n t o eqs.(II,k.5a) (11, 4.33) and r e j e c t i n g t h e norJinear terms, w e o b t a i n expansions of t h e . s t r e s s t e n s o r components i n powers of 2

I n s p i t e of t h e f a c t t h a t all t h e c o e f f i c i e n t s of t h e zm in t h e r e s u l t a n t expansion a r e f u n c t i o n s of t h e coordinates xi, t h e s e expansions do not y e t . n i n f i n i t e set of d e r i v a t i v e s solve t h e reduction problem, since they contain a These d e r i v a t i v e s .a,u, or of covariant d e r i v a t i v e s V, .V,uJ of t h e form &.. may be regarded a s new unknown functions subject t o detemnination. The reduc.V3uJ from t i o n problem will be solved a f t e r eliminating t h e d e r i v a t i v e s V, t h e system of equations of s h e l l dynamics.

.. .

..

Section 5. Reduction of t h e Three-Dimensional Problem t o t h e Determination of S i x Functions of a Point of t h e Base Area of t h e S h e l l W e s h a l l show t h a t t h e reduction problem has a d e f i n i t e a n a l y t i c meaning, Indeed, i.e., t h a t it can be formulated a s a problem of mathematical physics. t h e Lam6 equations (11, 5.5b) permit us, as was f i r s t shown elsewhere (Bib1.23a and b ) , t o express on t h e b a s i c surface t h e d e r i v a t i v e s V3V3 ..uJ, beginning w i t h t h e d e r i v a t i v e of second order, i n terms of t h e d e r i v a t i v e s V 3 u J , of funci.e., derivatives t i o n s o f u J , and of t h e t?tangential?t d e r i v a t i v e s V,V, ...VruJ, with respect t o t h e coordinates 2 and 2 of t h e b a s i c surface. From eq.(II, 5.5b) we f i n d :

Further, d i f f e r e n t i a t i n g eqs0(5.la)

(5.lb) we obtain

(i, s = 1, 2).
89

(5.2b)

S u b s t i t u t i n g i n t o t h e r e s d t a n t expression t h e values of t h e second d e r i v ativesV,V,q from eqs.(5.la) ( 5 . 1 b ) , we f i n d t h e f i n a l expressions f o r t h e derivatives:

(t,s=l,

2).

D i f f e r e n t i a t i n g eqs. (5.3a) - (5.3b) and repeating t h e process o f eliminaV 3 u j , beginning with t h e d e r i v a t i v e s o f second order, t i n g t h e d e r i v a t i v e s V, we can f i n d expressions f o r t h e d e r i v a t i v e s V,...V,uj of a r b i t r a r y m u l t i p l i c i t y i n terms of d e r i v a t i v e s containing t h e o p e r a t o r v , i n an order not higher than t h e first. By m e a n s of t h e r e s u l t a n t expressions f o r t h e d e r i v a t i v e s ofv,... v 3 u j , all q u a n t i t i e s c h a r a c t e r i z i n g t h e s t r e s s e d and s t r a i n e d s t a t e o f t h e s h e l l w i l l be determined by expansions of t h e form of eq~.(1~.5a) (L1..5f), as functions of t h e coordinates o f t h e b a s i c surface x i ( i = 1, 2 ) and e x p l i c i t functions of t h e coordinate x? = z, i f w e b o w t h e six functions xi : xJ and Ne s h a l l consider below c e r t a i n methods of determining these funcV3uJ & tions, which will be c a l l e d fundamental. W e draw t h e reader's a t t e n t i o n t o t h e increase i n t h e o r d e r o f t h e time d e r i v a t i v e s e n t e r i n g i n t o t h e equations, i f terms containing z4 are introduced i n t o t h e expansions.

...

Section
V30..V3~J

6. A m l i c a t i o n of t h e Smnbolic Nethod

The reader has probably noticed t h a t t h e determination of t h e d e r i v a t i v e s i n t h e preceding Section, beginning w i t h d e r i v a t i v e s of t h e second order, in terms of q u a n t i t i e s determined i n t h e i n t e r n a l system of coordinates of t h e b a s i c surface and t h e d e r i v a t i v e s of V3uJ i s e s s e n t i a l l y an a l g e b r a i c operation, This operation i s simplified i n connection w i t h t h e Ricci theorem and t h e vanishing of t h e curvature tensor. W e r e c a l l t h a t t h e Ricci theoren (I, Sect.9) permits u s t o operate with components of t h e metric t e n s o r as with coni ncovariant d i f f e r e n t i a t i o n , while t h e vanishing of t h e curvastant quantities . t u r e t e n s o r permits u s t o vary t h e sequence of d i f f e r e n t i a t i o n in multiple derivatives. These p r o p e r t i e s of t h e operations employed by u s permit t h e use,
90

i n determining t h e d e r i v a t i v e s V3...V3uj? o f me symbolic metnoas developea by Let u s introduce t h e notation: A.I.Lurfye?F.

en'
u y = v,.
n

. .va uj;
times

up)= u , .

(6.la)

=V R.

.-.V,, Fj;

tunes

F j '

= Fp

(6.lb)

The Lame/ equations i n t h e form of eqs.(5.la) ing r e l a t i o n s :

(5,lb) l e a d t o t h e follow-

(i, ~ = l , 2;

n=2, 3,...J.

(60%)

Equations ( 6 .2a) - (6.2b), i n a more e a s i l y v i s u a l i z e d form than rela- bA t i o n s (5.la) - (5 .lb), show t h e recurrence of t h e r e l a t i o n s between t h e SUC? t h a t r e s u l t from t h e Lam& equacessive covariant d e r i v a t i v e s with respect t o Y tions.
Ne s h a l l now introduce abbreviated symbols for t h e d i f f e r e n t i a l operators.

Let

From t h e works of A.I.Lurfye w e here c i t e t h e monograph (Bibl.9b) i n which t h i s method w a s applied t o t h e theory of equilibrium o f an e l a s t i c layer. The Lurfye method w a s a l s o applied by 1.T.Selezov i n h i s d i s s e r t a t i o n "Study of t h e Propagation of E l a s t i c Waves i n P l a t e s and Shells" ( I n s t i t u t e of Mechanics Ukr SSR Academy of Sciences, 1961) i n s e t t i n g up t h e generalized equations f o r t h e t r a n s v e r s e v i b r a t i o n s of p l a t e s , by t h e method developed i n our own work (Bibl. 23a, b )

91

Ma.=-

+ 2P

gS"vsvs+

at" -

a2

M;

(4.3b 1
The o p e r a t o r N i s known t o u s from t h e theory of propagation of waves. Equations (6.2a) - (6.2b) t a k e t h e following form:

(i, s = 1, 2;

n = 2 , 3,...).

The system of eqs.(6.4a) (6.4b) may be regarded as a system of a l g e b r a i c equations permitting u s t o express successively a l l t h e f u n c t i o n s of ~ $ 4 and e shall not here perform t h i s U,b) i n terms of uJ(d and u / ~ ) ( j = 1, 2, 3 ) . W The i n i t i a l s t e p of t h i s operation had operation of successive elimination. been pointed o u t i n the preceding Section. W e note, i n conclusion, t h e t e n s o r p r o p e r t i e s of t h e q u a n t i t i e s introduced by us, and of eqs.(b.&a) - (6.4b).
If we consider p o i n t transformations of coordinates on t h e b a s i c surface of a s h e l l , then with r e s p e c t t o t h e s e transformations t h e q u a n t i t i e s up) (i = = 1 , 2 ) a r e v e c t o r components while t h e q u a n t i t i e s uJn) are s c a l a r s . The proof of t h i s a s s e r t i o n i s obvious. The o p e r a t o r s introduced by u s can a l s o be regarded as symbolic t e n s o r s o f various ranks and s t r u c t u r e s on a s e t o f coordinates x i (i = 1, 2).

Section 7. Expressions f o r t h e " N o m l " P a r t o f t h e S t r e s s Tensor. The m u a t i o n s Determining t h e Fundamental Fun-ctions

~~

W e s h a l l c a l l t h e s e t or" components c13 t h e normal p a r t of t h e s t r e s s ten-

sor. The o t h e r components of t h e s t r e s s t e n s o r form i t s t a n g e n t i a l p a r t . It i s easy t o convince o u r s e l v e s t h a t t h e components d 3 ( i = 1, 2 ) a r e /96 v e c t o r components on t h e set o f i n t e r n a l coordinates of t h e p o i n t s of t h e b a s i c surface, and t h a t t h e component c~~~ i s a s c a l a r on this s e t .
Making use of t h e expansions (4.5d) - (4.5f) extended t o include terms i n of t h e n o t a t i o n s of eqs.(h.la) and (6.1b), and o f Hooke*s l a w (11, 4.5b), we f i n d :
z3,

(i, s = 1, 2).

(7.lb)

? . l a ) and (7.lb) determine t h e s t r e s s t e n s o r coinponents d i s Equations ( According t o t h e R i c c i theorem, placed t o t h e b a s i c s c r f a c e (cf.1. Sect.12). (I, Sect.9), under this p a r a l l e l displacement, t h e m e t r i c t e n s o r g k i s transformed i n t o t h e m e t r i c t e n s o r on t h e b e s i c surface. For t h i s reason, t h e quant i t i e s g e n t e r i n g i n t o eq.(7.15) a r e c o n t r a v a r i a n t conponents of t h e m e t r i c t e n s o r on t h e b a s i c surface.
To avoid misunderstandings, l e t u s n o t e t h e p r o p e r t i e s of t h e covariant d e r i v a t i v e s of ViUJ[n) (i = 1, 2 ; j = 1, 2, 3). T h e s e d e r i v a t i v e s , as before, are determined i n three-dimensional space. I n three-dimensional space, t h e f u n c t i o n s oi uJ(II a r e coinponents of a t e n s o r of rank n + 1. T h i s , according t o eq.(I, 9.12), d e f i n e s t h e meaning of t h e c o v a r i a n t d e r i v a t i v e s V,uJ
To d e t e r x i n e t h e fundamental functions, we make u.se of t h e conditions on t h e boundary s-arfaces of t h e s h e l l (Bib1.23b). O n t h e boundary s u r f a c e s t h e components of t h e e x t e r n a l f o r c e s are u s u a l l y assigned. To s e t f o r t h t h e essence of t h e method, l e t us conline ourselves t o t h e case o f a s h e l l of constant t h i c k n e s s and l e t u s assume t h a t t h e b a s i c s u r f a c e coincides with t h e middle s u r f a c e of t h e s h e l l . Consider t h e boundary condit i o n s (11, 8.2b). Under t h e simplifying hypotheses adopted, t h e s e conditions h i l l be o f t h e following form:

A l l t h e q u a n t i t i e s e n t e r i n g i n t o eqs.(7.2) are assumed t o be displaced p a r a l l e l t o themselves on t h e b a s i c surface, according t o previous statements (I, Sect.11 and 12).

The s e r i e s r e p r e s e n t i n g t h e components of t h e displacement and stresses are assumed t o converge w i t h i n t h e s h e l l and on i t s surface:?. >taking use o f eqs.(7.la) - (7.lb), we o b t a i n t h e following six equations:
ViU3

+u y +

/2

[d,ugl)

+ k , (vp3+u p )+up]+

( continued )

+:- This hypothesis is, as w i l l b e c l e a r from t h e contents of S e c t i o n 2, t h e m o s t

vulnerable p o i n t of t h e r e d u c t i o n method under consideration.

93

(s= 1, 2).

This system may be replaced by i t s equivalent:

m-0'

. .

To t h e s e equations w e must a s s o c i a t e t h e r e l a t i o n s (6.4a)

(6.4b).

m
-

Eliminating from eqs.(?.ha) - (7.l;b) on t h e b a s i s of eqs.(6,ka) (6.kb) t h e q u a n t i t i e s up) ( f = 1, 2, 3 ; n = 2, 3...), we o b t a i n a system of six equat i o n s i n six urdmown functions of uJ and u/l) ( j = l, 2, 3). This system will b e of an order depending on t h e number of terms in t h e expansions. I n turn, t h e number of terms i n t h e expansions will depend on t h e prescribed e r r o r o f Consequently, t h e order of t h e system of equations set up t h e wanted r e s u l t .

94

by us m a y be very higW.

L e t u s return t o e q s O ( 7 . l a ) we f i n d

(7.lb).

Making use of eqso(7.La)

(7.hd),

a33

=
-

&+I3

- &-l3 2

X(+,3

X(--)3

-+
z h

1 +2 (9

- hzj [AgSSv,ui?

+2pupj + . .

( i = 1, 2).
Equations (7.5a) - (7.31) e s t a b l i s h t h e l a w of d i s t r i b u t i o n o f t h e normal o f t h e s t r e s s t e n s o r over t h e thickness o f t h e s h e l l . These equations hold for t h e l i n e a r problems of t h e s t a t i c s and dynamics of shells::-:$.
Dart

Section 8. Further Development of t h e C l a s s i f i c a t i o n of S h e l l s with Respect t o Dynamic Problems L e t u s r e t u r n t o t h e c l a s s i f i c a t i o n of s h e l l s i n t o %bin" and %on-thin'l. A s will be c l e a r from Sect.3, i n t h e s h e l l theory t h e q u a n t i t y 2h i s u s u a l l y considered a small q u a n t i t y i f t h e n a t u r a l u n i t of l e r g t h i s taken as one of the c h a r a c t e r i s t i c dimensions of t h e s h e l l . I n t h e problems of dynamics, such an approach t o t h e c l a s s i f i c a t i o n o f s h e l l s 5-s inadequately motivated. A n a n a l y s i s of t h e question of t h e a p p l i c a b i l i t y limits o f t h e equations of t h e c l a s s i c a l theory of f l e x u r e of p l a t e s t o t h e s o l u t i o n of t h e dynamic problems was performed by G.I.Petrashen (Biblm3O)+Ess. Although t h i s work rel a t e s t o a s p e c i a l kind o f s h e l l , i t s r e s u l t s permit general conclusions t h a t
4s T h i s i s a l s o c l e a r from a study of t h e s t a t i c a l l y s t r e s s e d and s t r a i n e d states of an e l a s t i c medium by t h e method under consideration (cf. Bibl.96). $3:q u a t i o n s analogous t o eqs.(7.5a) (7.B) are presented by us elsewhere (Bib1.23b ). Analogous r e l a t i o n s were given subsequently by (Bibl.16, Bib1.26), and others.

The f u r t h e r development of t h e i n v e s t i g a t i o n by Petrashen i s contained i n t h e paper by LOA.Molotkov It%gineering Functions f o r t h e Vibrations of P l a t e s with Layered Structure", Leningrad Sect. 1nst.Mat. USSR Academy of Sciences, Coll. V: IQuestions of Dynamic Theory o f t h e Propagation of Seismic Wavestt,
\I \I

i=-r

, ,

1961. 95
I

are v a l i d f o r more g e n e r a l problems 03 t h e dymmics of s h e l l s . f o r e b r i e f l y enunerate h i s conclusions (Bib1.3C).

iie s h a l l t h e r e -

This study was based on a solution, exact w i t h i n t h e limits o f t h e linear theory of e l a s t i c i t y , of t h e problem of t h e v i b r a t i o n s or" a n unbouqded e l a s t i c layer under t h e a c t i o n o f a plane and a x i s p e - t r i c s u r f a c e load, and a l s o t h a t o f a normal l o a d unir"0rml.y d i s t r i b u t e d over a c r o s s s e c t i o n o f t h e surface layer. O n t h e b a s i s of an vlalysis o f t h e s o l u t i o n s obtained, Petrashen came t o t h e conclusion t h a t t h e t h i c k n e s s of a p l a t e f o r which t h e a p p l i c a t i o n of t h e theory or" t h i n p l a t e s w a s s t i l l possible, depends s u b s t a n t i a l l y on t h e prop e r t i e s of t h e f o r c e i n f l u e n c i n g t h e p l a t e . I n t h e f i r s t place, t h e width of t h e a p p l i c a t i o n zone of t h e l o a d and t h e zone of i t s appreciable v a r i a t i o n must c o r s i d e r a b l y exceed t h e thiclmess of t h e pla-Le, and I n t h e second place t h e load must vary s l o w l y . The l a t t e r requirement my be represented by t h e i n e q u a l i t y

where T i s t h e d u r a t i o n o f a p p r e c i a b l e v a r i a t i o n o f t h e s u r f a c e load, Id i s a l a r g e number, and

where b-l i s t h e v e l o c i t y o f propagation of t r a n s v e r s e e l a s t i c waves. Consequently, T i s t h e d u r a t i o n of t h e passage o f t h e e l a s t i c t r a n s v e r s e wave through t h e s e c t i o n o f t h e layer, ?.le my a l s o n o t e t h e r e l a t i o n pointed o u t by Petrashen between t h e r e g i o n s of t h e low-frequency spectrum v and t h e t h i c k n e s s o f t h e s h e l l , This r e l a t i o n i s of the fom: vbh <( 1. Thus, by i n c r e a s i n g t h e t h i c k n e s s of t h e s h e l l we decrease t h e region of fre,quencies v i n wi-ich t h e approxiirate theory of p l a t e s does n o t l e a d t o considerable e r r o r s . It I"ol1oi.r~ f r o m Petrashen's study (Eib1.30) i n p a r t i c u l a r , t h a t t h e n a t u r a l u n i t o f l e n g t h t h a t can b e adopted i s t h e qulantity v - l b - l , as will b e c l e a r I r o n eq.(E.3). The corresponding l e n g t h o f t h e s i n e w a e , a s i s g e n e r a l l y known, i s expressed by t h e ecpation

From t h i s follows t h e p o s s i b i l i t y cf choosing -LJL, i n solving c e r t a i n problems O Z t h e elastodynamics of s h e l l s , as a n a t u r a l &it of length. These f a c t s , /lo0 e s t a b l i s h e d by means of analyses of rigorous b u t p a r t i a l s o l u t i o n s o f boundary

96

dynamic problems of t h e theory o f e l a s t i c i t y , can undoubtedly be extended with only minor changes t o t h e s h e l l theory. W e do not, however, know o f any gene r a l i n v e s t i g a t i o n s t h a t would permit introducing a d d i t i o n a l terms i n t o t h e equ-ations of t h e c l a s s i c a l theory of p l a t e s and s h e l l s so as t o make t h e solut i o n s of t h e generalized equations represent, with s u f f i c i e n t accuracy, t h e sol u t i o n s of t h e corresponding boundary conditions of e l a s t i c i t f l - . Apparently, t h e s o l u t i o n of t h e approximation equations can with s u f f i c i e n t accuracy d e s c r i b e only some p a r t of t h e elastodynamic process studied, f o r example, some d e f i n i t e segment of t h e frequency spectrum, t h e phase or group v e l o c i t y o f waves with dispersion, etc. For t h i s reason, t h e d i v i s i o n of s h e l l s i n t o t h e c l a s s e s "thin" and %on-thin" must be subordinated from t h e beginning t o t h e problem of studying c e r t a i n c h a r a c t e r i s t i c s of t h e dynamic process. From t h i s p o i n t of view, t h e conditions (8.1) (8.3) determine t h e c l a s s of t h i n s h e l l s , depending on t h e d e s i r e d accuracy of t h e study of t h e results of perturbing f o r c e s applied t o them. I n t h i s connection we note t h a t it i s a l s o possible in t h e problems of dynamics t o introduce various n a t u r a l units o f length, subordinating them t o t h e fundamental purpose of t h e subsequent i n vestigation.

For instance, l e t u s propose t o study t h e propagation o f e l a s t i c waves of ,, i n an unbounded s h e l l , i.e., i n a s h e l l homeomorphous l e n g t h s not l e s s than & with an unbounded layer. L e t u s put, according t o e q ~ ~ ( 3 . 1and ) ( 8. 4) ,

Let M be t h e number o f f i r s t terms retained i n t h e above-discussed expans i o n s and E t h e prescribed r e l a t i v e d e v i a t i o n of some c h a r a c t e r i s t i c q u a n t i t y ( f o r example, of the'phase o r group v e l o c i t y of waves with dispersion), d e t e r mined - on t h e b a s i s of t h e approximate theory o f s h e l l s - from t h e value of this q u a n t i t y derived from t h e equations of t h e three-dimensional problem. Then, it i s p o s s i b l e t o f i n d

where t h e condition (8.5) w i l l d e f i n e t h e c l a s s of t h i n s h e l l s .

For all types of waves of l e n g t h s a t i s f y i n g t h e i n e q u a l i t y

1 >/ I , ,

(8.7)

t h e s h e l l will l i k e w i s e b e thin. For o t h e r waves, t h e s h e l l w i l l not be t h i n , i.e., t h e number of terms r e t a i n e d will not ensure t h e necessary accuracy of
3:-

The equation obtained by Petrashen (Bibl.30) on t h e b a s i s of t h e s o l u t i o n s of t h e above-mentioned p a r t i a l problems of t h e e l a s t i c i t y theory give no general answer t o t h e question posed.

97

I'

solution.

W e r e c a l l t h a t t h e study of another r e p o r t (Bibl.30) w a s based on an analyConsequently, t h e /lo1 d e f i n i t i o n of t h i n s h e l l s i n d i c a t e d here may require s u b s t a n t i a l additions, or even be u n s u i t z b l e f o r solving dynamic boundary problems i n t h e case of bounded shells.

sis of rigorous s o l u t i o n s f o r an unbounded e l a s t i c layer.

The above statements and those i n Sect.3 l e a d t o t h e conclusion t h a t t h e r e e x i s t s no general c r i t e r i o n t h a t would permit a c l a s s i f i c a t i o n of s h e l l s i n t o t h e s e classes. There i s a l s o no general n a t u r a l u n i t of l e n g t h r e s u l t i n g from t h e p r o p e r t i e s of t h e dynamic processes in s h e l l s . I n concrete problems, however, t h e i n t r o d u c t i o n of a s u i t a b l y chosen n a t u r a l u n i t of l e n g t h may prove useful. W e s h a l l assume below t h a t such a physical or geometrical u n i t has been selected and t h a t t h e q u m - t i t y 2h i s s u f f i c i e n t l y small, i.e., that the conditions (3.1) are s a t i s f i e d , with t h e p o s s i b l e replacement of t h e second cond i t i o n by t h e r e l a t i o n (8.3) or (8.4). The use of t h e conditions (3.3) and (3.4) i s likewise possible. Section 9. Piethod of Successive Approximations (7.4d) c o n s t i t u t e s t h e foundaAlthough t h e system of equations (7.4a) t i o n of one of t h e p o s s i b l e a n a l y t i c statements of t h e problem o f reduction,the complexity of t h i s system and t h e absence of a c r i t e r i o n allowing preliminary conclusions as t o convergence of t h e s e r i e s on t h e left-hand s i d e of t h e s e equat i o n s f o r c e s u s t o t u r n t o methods t h a t permit s o l u t i o n of t h e problem of reduction without i n t e g r a t i n g eqs.(7.4a) (7.4d). Such a method has been given by u s elsewhere (Bib1.23a, b). It i s t h e method o f successive approximations, based on t h e hypothesis t h a t 2h i s r e l a t i v e l y small.

To develop t h e process of successive approximations, l e t u s make use of eqs.(7.4a) and ( 7 . 4 ~ ) . Equations (7,kb) and (7.4d) w i l l not as y e t be applied. Subsequently, eqs.(7.bb) and (7.4d) will permit u s t o develop one of t h e a l t e r n a t i v e versions of t h e s o l u t i o n of t h e reduction problem. If w e do not use t h e s e equations, then a u x i l i a r y equations predetermining t h e statement of t h e problem must be s e t up.
To f i n d t h e first ( i n i t i a l ) approximation, l e t u s r e j e c t from eqs.(7.4a) and (7.Lc) a l l terms containing h. Then,

(i, s = 1, 2).

(9.2)

To f i n d t h e next approximation, l e t u s m a k e use of eqs.(6.4a) - (6.&b), /1C2 p u t t i n g n = 2 and n = 3. Determining t h e f i r s t approximation f o r [ui(a)]l, and

98

[I-&) 11, we r e t u r n t o eqs.(?.4a) and (7.4.~). From t h e s e equations, r e t a i n i n g i n them a l l terms with t h e f a c t o r h2, we will f i n d t h o second approximation of the basic quantities:

(9.3)

(9.4.) The process can be continued f u r t h e r . Applying this method l e t u s f i n d i n e n-th approximation f o r t h e ccmponents ):e : These q u a n t i t i e s w i l l herea f t e r be c a l l e d [elg ] , , (i = 1 , 2, 3 ) . W e have:

W e find: (9.6a) (9.6b) EQuations (9.6a) - (9.6b) permit u s t o d e r i v e formulas r e f l e c t i n g , i n exp l i c i t form, t h e d e v i a t i o n of t h e proposed s h e l l theory from t h e c l a s s i c a l theo ~ f i - . Equations (9.1) - (9.6b) will h e r e a f t e r b e c a l l e d t h e redilction formulas: Let us make a preliminary a n a l y s i s of t h e r e l a t i o n s obtained.
1 . If t h e r e i s no load on t h e boundary s u r f a c e s of t h e s h e l l , then from eqs.(9.l) - (9.2) follow t h e r e l a t i o n s :

These equations express t h e condition of invariance of an element of t h e b a s i c surface normal t o t h e middle surface. Thus t h e first approximation i s c l o s e t o t h e Kirchhoff-Love hypothesis, which i s s t i l l l e s s r e s t r i c t i v e f o r t h e
~~

s+ Here we have somewhat modified t h e n o t a t i o n adopted by us elsewhere(Bibl.23b).

99

s t r a i n s t h z n t'his hypothesis. 2. The nethod of successive approximation given here r e q u i r e s t h e d i f f e r e n t i a b i l i t y or" t h e f u n c t i o n s X ( & l J and p F J .

3. W e n o t e t h e r u l e s for covariant d i f f e r e n t i a t i o n of t h e functions Xi(_+). Although t h e s e f u n c t i o n s a r e e s s e n t i a l l y components of a contravariant vector, they n e v e r t h e l e s s express, according t o t h e reduction formulas, t h e compon- /103 e n t s of a covariant t e n s o r of second rank. This determines t h e r u l e s f o r t h e i r covariant d i f f e r e n t i a t i o n .
4. W e have not been a b l e t o e s t a b l i s h a general proof o f t h e convergence Elsewhere of t h e process o f successive approximation suggested by US. (Eib1.23b) w e have i n d i c a t e d methods f o r t h e preliminary approach t o such a proof i n t h e case of s t a t i c problems. A s f o r t h e problems of dynamics, t h e The question of t h e convergence o d i f f i c u l t i e s here are s t i l l considerable. t h e process of successive approximations may b e approached i n t h e problems of s t a t i c s based on general a n a l y t i c p r o p e r t i e s of t h e s o l u t i o n s o f t h e boundary It can be a s s e r t e d t h a t , problems of t h e e l a s t i c i t y theory mentioned i n Sect.2. f o r t h e cases of t h e a c t i o n of f o r c e s determined by f u n c t i o n s of a p o i n t without a n a l y t i c s i n g u l a r i t i e s , t h e s e r i e s constructed by u s w i l l i n f a c t converge. But these s e r i e s will apparently diverge in t h e neighborhood of t h e p o i n t s of a p p l i c a t i o n of concentrated forces. O f course, concentrated f o r c e s a r e one of e can o b t a i n a t h e forms of l i m i t i n g a b s t r a c t i o n s . It i s c l e a r t h a t even here w s o l u t i o n t h a t i s s a t i s f s c t o q from t h e physical viewpoint by s u b s t i t u t i n g t h e concentrated f o r c e by i t s equivalent load, d i s t r i b u t e d over a small b u t f i n i t e region of t h e body.
The a n a l y t i c p r o p e r t i e s of t h e s o l u t i o n s of dynamic problems of t h e elast i c i t y theory as i n v e s t i g a t e d t o date, do not permit a d e f i n i t e answer t o t h e question whether t h e successive approximations developed by our method a c t u a l l y e must consider t h e proposed method as merely an converge+. For t h i s reason, w algorithm for obtaining approximate equations of t h e d y n m i c s o f e l a s t i c s h e l l s . These equations a r e m b j e c t t o f u r t h e r experimental and t h e o r e t i c a l v e r i f i c a tion. There a r e i n d i r e c t confirmations of our methods. For erample, in t h e work it i s shown t h a t t h e method under by M.P.Petrenko and t h a t by I.T.Selezov:+ discussion permits obtaining equations for t h e l o n g i t u d i n a l and t r a n s v e r s e vib r a t i o n s of rods and t h e t r a n s v e r s e v i b r a t i o n s of p l a t e s which y i e l d , a s spec i a l cases, t h e equations found by o t h e r methods and by o t h e r authors. I n t h i s manner, it i s p o s s i b l e t o o b t a i n a g e n e r a l i z a t i o n of t h e d i f f e r e n t i a l equation
3: The s t a t e of t h e general theory of s o l u t i o n of t h e problems of elastodynamics i s i n d i c a t e d i n V.D.KupradzeTs book ?Boundary Problems of t h e Theory of Vibrations and I n t e g r a l Equations", Gostekhizdat, 1950.

x-x- C i ' . t h e d i s s e r t a t i o n by M.P.Petrenko ffLongitudinal and Transverse Vibrations Arising i n Short Rods of Constant and Variable Thickness under t h e Action of an Impact" I n s t i t u t e of Mechanics, Ukr SSR Academy of Sciences, 1961, and the above-cited d i s s e r t a t i o n by 1.T.Selezov.

100

of t h e l o n g i t u d i n a l v i b r a t i o n s of rods found by S,P.Timoshenko, as w e l l as vari o u s g e n e r a l i z a t i o n s of known equations for t r a n s v e r s e v i b r a t i o n s of p l a t e s , f o r example t h e equations given by Ya.S.Uflyand, et a l . These r e s u l t s appar-/lGlL e n t l y confirm t h e expedience o f t h e method proposed here. Returning t o t h e question of t h e convergence of t h e above-suggested method o f . successive approximations, it i s u s e f u l t o c i t e t h e concepts by A.M.Krylov on t h e convergence i n purely a n a l y t i c a l and applied research+. Obviously, rlconvergencefr i s important here in view of t h e f a c t that,, a f t e r a f i n i t e l y s m a l l number of approximations, it w i l l y i e l d s u f f i c i e n t l y exact eqcations,i.e., equations whose s o l u t i o n s w i l l s a t i s f y t h e equations of t h e mathematical theory of e l a s t i c i t y and t h e boundary conditions, with an e r r o r s u f f i c i e n t l y small from t h e viewpoint or" p r a c t i c a l requirements. The study (Bibl.30) on t h e c l a s s i c a l theory o f p l a t e s shows t h a t t h e above-mentioned I f p r a c t i c a l convergence" will t a k e pla,ce whenever t h e r e s t r i c t i o n s i n d i c a t e d i n Sect.8 a r e imposed on t h e a c t i n g forces. The question of t h e a p p l i c a b i l i t j limits of t h e equations obtained by t h i s method r e q u i r e s f u r t h e r investigation. Section 1 0 . Expansion of t h e S t r a i n Tensor i n t o a Tangential P a r t and a Rormal P a r t

L e t u s r e t u r n t o t h e expansions (L.5a) - ( b . 5 ~ ) of t h e components of t h e s t r a i n t e n s o r ~2;) (i, k = 1 , 2). These components were not used by u s i n solve s h a l l show, f o r z = 5 they d e s c r i b e t h e dei n g t h e reduction problems. A s w formation o f t h e b a s i c surface, i.e., they determine, with an accuracy t o quant i t i e s of t h e second o r d e r of' smallness, t h e v a r i a t i o n s o f t h e metric t e n s o r components of t h e b a s i c surface. They a l s o contain terms dependkg on t h e vari a t i o n oiP t h e curvature of t h e b a s i c surrace.
z) de w i l l denote t h e s e t of terms o f e if k deterriining t h e v a r i a t i o n o f t h e metric t e n s o r Components o f t h e b a s i c surface, a s t h e t a n g e n t i a l p a r t of t h e s t r a i n tensor. de s h a l l c a l l t h e s e t of i u a n t i t i e s belonging t o 2; ) and depending on t h e curvature v a r i a t i o n of t h e b a s i c surface, t h e normal p a r t o f t h e s t r a i n tensor. Let us represent eqs.(&.ra) - (L.5c) i n t h e following form:

(1O.lb) (1G.lc ) The q u a n t i t i e s cik form t h e t a n g e n t i a l p a r t o f t h e strain tensor. These /lo5 q u a n t i t i e s are determined by t h e formulas:
ell = dlItl -

1
~

2gn % C f . W o l l e c t i o n of Works of Academician A.N.Krylovf', Sciences, 19h8, pp.205-2C6.

' L , % g 1 ,

i + 2g2,

___ ' L z Q h

-klgIl4,
Vol.X,

(1~2a)
USSR Academy of

10 1

( 10.29

(10.2c)

The q u a n t i t i e s .%ik a r e t h e components o f t h e symmetric covariant t e n s o r o f second rank on t h e b a s i c surface. This i s known as t h e t e n s o r of v a r i a t i o n s of curvature. The c o m e c t i o n between t h e q u a n t i t i e s .%ik and t h e v a r i a t i o n s of curv a t u r e will be c l e a r from eq.(l+.2), bearing i n mind t h a t t h e d e r i v a t i v e s 0 , are absolute d e r i v a t i v e s i n t h e d i r e c t i o n o f a normal t o t h e b a s i c surface and maki n g use o f eqs.(I, 3;6a) - (I, 3.6b). The q u a n t i t i e s t l i k , corresponding t o t h e n-th approxirnation, are determined on t h e b a s i s of eqs.(4.5a) (lF.5c) and t h e (9.6b): reduction formulas (9.6a)

a r e absent from t h e equations of t h e c l a s s i c a l The terms containing theory. They obviously c h a r a c t e r i z e t h e influence of t h e l o c a l l o a d s on t h e curvature of t h e s h e l l . The r e l a t i o n s (10.2a)

( 1 0 . 3 ~ ) c o n s t i t u t e t h e f i r s t (kinematic) group

/ l d

102

of equations of t h e s h e l l theory s e t up on t h e c l a s s i c a l plane. The terms of t h e expansions of et:) containing m , where n 2 2, have no s p e c i a l names and t h e i r geometrical meaning i s more complicated than t h a t of c i k and H i k e The scope of t h e present study does n o t permit u s t o go more deeply i n t o t h e s e kinematic investigations.

1 . Two Kethods of S e t t i n g up t h e Q u a t i o n s of t h e Theory of Section 1 Shells, b o t h Connected with t h e Method of Successive Approximations. F i r s t Version of Establishment of t h e Gastodynamic Systems of Equations
~

The method of successive approximations r e q u i r e s t h e use of t h r e e equa('7./+d). I n essence, t i o n s derived from t h e system of six equations (7,La) t ' s method i s one of t h e methods of eliminating t h e t h r e e unknown functions I . $ (i = 1 , 2, 3 ) from t h e six unknowns i n t h e system of equations (7,ha) (7.4d). To o b t a i n a complete system o f equations of t h e s h e l l theory, t h r e e more equations must be s e t up with unknown functions u , ( i = 1, 2, 3 ) . This may b e done by two methods.

The first method i s based on t h e use of t h e t h r e e equations of t h e system (7.Ld) t h a t had n o t been used by u s i n d e r i v i n g t h e reductiori formulas. By eliminating t h e q u a n t i t i e s q ' " from t h e s e equations on t h e b a s i s of t h e reduction formulas, we o b t a i n a system of t h r e e equations with unknown functions u l (i = 1 , 2, 3 ) .

(7.4a)

The second method c o n s i s t s i n s e t t i n g up t h e conditions of equilibrium of m element o f t h e s h e l l as a whole, followed by t h e a p p l i c a t i o n of r e l a t i o n s res u l t i n g from Hookets l a w , and i s t h e most widely used i n modern s h e l l theory, uld e s s e n t i a l l y corresponds t o t h e construction of t h e c l a s s i c a l theory. Consider t h e elastodynamic system of equations of t h e theory of s h e l l s , derived from eqs.(7.&a) - (7.4d) and fro m t h e reduction formulas, and l e t u s e must e s t a b l i s h t h e r e l a t i v e accuracy of t h e r e t a i n t h e first version. F i r s t w required system of equations. W e s h a l l conditionally d e f i n e t h i s accuracy by t h e highest power o f h in t h e t'erms retained i n t h e equations. The equations o f t h e c l a s s i c a l theory of s h e l l s were u s u a l l y confined t o terms containing h3, b u t introduced only some of t h e s e terms+$. During t h e l a s t 1 0 or 12 years, a nurriber of s t u d i e s on t h e dynamics of p l a t e s and c y l i n d r i c a l s h e l l s have been published, containing terms i n h3 b u t a l s o omitting a number of terms of t h i s order, without giving s u f f i c i e n t reasons f o r t h e legitimacy of neglecting them. Below, we a l s o confine ourselves t o s e t t i n g up t h e equa/lo7 t i o n s of s h e l l dynamics, containing a l l terms up t o and including t h e f a c t o r kr? However, t h e method employed by u s makes it p o s s i b l e t o set up equations cont a i n i n g a l l terms t o an a r b i t r a r y power of M*r.
~

x- W e have given elsewhere (Bib1.23b) a d e t a i l e d a n a l y s i s of t h e completeness of t h e c l a s s i c a l system of equations of t h e s t a t i c s of shells.

+:*I n t h e above-cited d i s s e r t a t i o n by I.T.Seleeov, t h e equations of t h e vibrat i o n s of p l a t e s , containing terms up and including h5, w e r e set up by this
method i n t h e expanded form.

103

_To obtain t h e required degree o f accuracy, t h e twc f i r s t terms and t h e seThen w e r i e s mist b e r e t a i n e d in t h e left-hand s i d e s of eqsm(7.Lb) - (7.Ld). o b t a i n t h e followirig approximtlor, fo;mulas:

( 1 1.lb )
L e t us make use of t h e r e c u r r e n t r e l a t i o n s (&.ha) - (6.1Lb) and t h e reduct i o n formulas, and l e t u s ROW mal:e a l l t h e c a l c u l a t i o n s w i t h t h e required den the b a s i s o f eqs.(S.lla) (643) w e f i n d successively gree of detail. O

1Olr

( i , r , s= 1,2).

( 11.L i b)

Now, applying t h e reduction f o m u l a s , we can exclude t h e q u a n t i t i e s uJ[ from t h e r e s u l t a n t equation. Here, i n view of t h e prescribed a r b i t r a r y accuracy of t h e equations, w e w i l l introduce t h e fi-rst approximations i n t h e expresand u / ~ ) ,w h i l e t h e remaining q u a n t f t i e s e n t e r i n g i n t o e q s . ( l l . l a ) sions ( 1 l . l b ) can be determined by t h e second approximations.

d3)

I n order t o execute t h i s program, w e must r e t u r n t o eqs.(9.l) - (9.2) and introduce t h e r e our newly adopted notation, and then f i n d t h e f i r s t approximat i o n of t h e q u a n t i t i e s uJ(), up), uJ(*). Then ,we w i l l be able t o w r i t e eqs. (9.3) - (9.4) i n t h e expanded form and complete s e t t i n g up t h e system of equations From eqs.(6.3a) - (6.3b), we f i n d (1l.la) (11.1b).

(11.5)
W e put

(11.6)

lie have here considered t h e remarks i n Sect.9 on t h e meaning of t h e functions Fquations (9.1) - (9.2) then t a k e t h e following form:

X ( * )

Further elementary b u t unwieldy c a l c u l a t i o n s l e a d t o t h e Pollowing gen- / l C 9 e r d expressions o f t h e wanted q u a n t i t i e s :

(Ui2jm =( P ( n ) ] m ~ l

+ [Rl)],,p + s [Qj2],;

(11.8a )

' 1, ;%(2("31m[si( 2 n - 9 1 m y ~Is(2n-1) 8 J, are The expressions [P(2n)]m [q(2n) d i f f e r e n t i a l o p e r a t o r s depeniing on t h e o r d e r of approximation. W e shall i n d i c a t e below t h e form of t h e s e o p e r a t o r s of t h e approximations introduced by us. The q u a n t i t i e s [ are "force terms" containing t h e d i f f e r e n t i a l operations on t h e body f o r c e s and s u r f a c e forces.

QP)],

To start with, we i n d i c a t e t h e values of t h e o p e r a t o r s i n eys.(ll.8a) ( l l . 9 b ) f o r t h e first approximation. W e have

(1l.lOa)

( 1 1 . l o b) ( 1 1 1la)
( 1 1llb )

. .

(11.12a)

(i, f , s = 1, 2).
(11.123)
The o p e r a t o r s have t h e following meaning :

(11.13a)
(lLl3b)

(1 1 1Lb )

(11.1Lc )

( l l . % ) ,

Quztions ( 1 1 . 1 G a ) - (ll.llJd), t o g e t h e r with t h e r e l a t i o n s (ll.8a) determine t h e first approximation. Let us now consider t h e second approximation.
To s e t ~p e q s . ( l l . l a )

lrforcelr o p e r a t o r s contained i n t n e expressions or [uJ(l) Iz and have

(1l.lb) with t h e necessary accuracy it is s u f f i c i e n t t o consider t h e second approximations o f t h e q u a n t i t i e s u ; ) and ( j = 1, 2, 3 ) . Again s t a r t i n g from eqs.(ll.Sa) - (l109b), we s h a l l give t h e values of t h e d i f f e r e n t i a l and

u12)

Iz

W e

P [S1], = L, A*P

( 11.15~~ )

[P*],=M ; [Rj:? = L i :

A L,N~: -+Afp

(11.16b )
p u t h e r , we f i n d t h e force operators;

/ 1 1 1

(11.18a)

(11.18b )
L e t us return t o eqs,(ll.la) ( 1 l . l b ) . Making use of t h e n o t a t i o n (11.5) we represent t h e s e equations i n t h e following form:

(11.19a)

(i, s = 1, 2).
where

(11.19b )
( 1 1 . 2 0 )

/112

(i, s = 1. 2)

(11.21b)

X e s h a l l now make s e v e r a l preliminary remarks on t h e system of equations ( 1 1 . 2 1 ~ ~) (11.21b).

1 . The system of equations (11.21a) (11.21b) i s of t h e t w e l f t h order. W e r e c a l l t h a t t h e o r d e r of t h e s y s t e m of equations i n t h e c l a s s i c a l theory i s eight. The i n c r e a s e i n t h e order of t h e system i s due t o t h e i n t r o d u c t i o n of a l l terms with f a c t o r s h a up t o h " inclusive. The o r d e r o f t h e system of equat i o n s (11.21a) (11.21b) i s lower than t h e o r d e r o f t h e i n i t i a l system ( 1 1 , l a ) - (11.1b). The o r d e r of t h e i n i t i a l system,as i s o b v i o u s , i s 21. Here t h e higher d e r i v a t i v e s i n e q s . ( l l . l a ) and (1l.lb) will be mixed d e r i v a t i v e s with r e s p e c t t o t and t h e coordinates xi(i = 1 , 2). The system of equations (1l.la) - ( 1 l . l b ) w i l l be o f t h e f i f t e e n t h o r d e r i n t h e d e r i v a t i v e s with respect t o t h e coordinates x i . The lowering of t h e o r d e r of t h e system as a result of t h e app l i c a t i o n of t h e method of successive approximations i s due t o elimination of t h e terms containing f a c t o r s of hn where n 2 1,; . i n introducing t h e formulas of (11.1b). reduction i n eqs. ( 1 l . l a )

2. The system o f equations (11.21a) (11.21b) approximately d e s c r i b e s t h e dynamic process i n e l a s t i c s h e l l s of a r b i t r a r y form. A s we have a l r e a d y noted, t h e s e equations contain a l l terms with f a c t o r s P where n S 3. Neglecting t h e remaining terms n a t u r a l l y limits t h e s i g n i f i c a n c e of t h e derived equations. This f a c t will become obvious when considering t h e boundary and i n i t i a l conditions.

3 . The system of equations (11.21a) (11.21b) "symbolicallyfr i s resolved i n t o a system of two equations containing t h e t a n g e n t i a l components o f t h e d i s placement v e c t o r and an equation with t h e nomtal component. This r e s o l u t i o n , however, i s i l l u s o r y s i n c e t h e covariant d e r i v a t i v e s , forming t h e b a s i s of t h e o p e r a t o r s introduced by u s t o shorten t h e notation, are sets containing all com-

ponents of t h e v e c t o r u . accomplished.

Only i n t h e plane problem i s t h i s r e s o l u t i o n a c t u a l l y

4.. The system of equations (11.21a) (11.21b) contains t h e wave operat o r s 14. However, t h e question of t h e existence of t h e a c t u a l c h a r a c t e r i s t i c s of eqs. (11.21a) - (11.21b) r e q u i r e s s e p a r a t e analysis.

5. The system of equations (11.21a) - (11.21b) obtained b y ' o u r a n a l y t i c methods i s very complex and permits only approx5mate i n t e g r a t i o n , n e g l e c t i n g a number of terms. It seems u s e f u l , however, t o introduce such a s y s t e m i n t o t h e a r s e n a l of d e s c r i p t i v e means of t h e theory of s h e l l s as a p e c u l i a r ttstagert, /ll3 permitting u s t o judge t h e accuracy of t h e equations obtained by .other, more p i c t o r i a l method 9 :

Equations (11.21a) - (11.21b) do not d e f i n e t h e statement of t h e dynamic boundary conditions of t h e s h e l l theory. The boundary and i n i t i a l conditions I n order t o do t h i s , we must first f i n d approximation exmust be considered. pressions f o r t h e stress t e n s o r components. Section 12. Approximate Ekpressions f o r t h e Components of t h e Displacement Vector and t h e Components o f t h e S t r e s s Tensor In considering t h e expansions o f t h e displacement v e c t o r components, t h e question a r i s e s a s t o t h e number of terms t h a t must be retained i n these expansions. e will r e t a i n Based on t h e r e l a t i v e accuracy of eqs.(ll.2la) - (11.21b), w i n t h e expansions o f t h e displacement vector components a l l terms including components with f a c t o r s z3. Here, however, w e have a c e r t a i n inconsistency, since (11.21b) contain terms depending on c o e f f i c i e n t s of z4 in t h e eqs.(ll.2la) expansions o f t h e displacement v e c t o r components. This'inconsistency, however, i s apparently one o f s e v e r a l c o n t r a d i c t i o n s of t h e theory under consideration. Below, we will d i s c u s s t h e c o n t r a d i c t i o n s i n t h e approximate theory of s h e l l s i n e find more d e t a i l . I n t h e n o t a t i o n adopted by u s w

(i= 1, 2, 3).

( 1 2 . 1 )
we obtain

Making use of t h e r e l a t i o n s (11.8a)

(11.9b),

+?

The d e s i r a b i l i t y of i n v e s t i g a t i o n s t o obtain a r b i t r a r i l y "exact" equations of t h e s h e l l theory, permitting a judgment from t h e p r o p e r t i e s of t h e r e j e c t e d terms, w a s discussed a t t h e Conference on S h e l l Theory held i n October 1960 at

Kazanf

1 1 0

Consider t h e expansions o f t h e stress tensor components. t h e n o t a t i o n o f eq.(11.5), we f i n d

Introducing

/ 1 1 L

Equations (11.8a) ( l L 9 b ) , together with t h e values o f t h e i r o p e r a t o r s expressed by eqs. (1l.lOa) (11.18b), permit t h e approximate r e p e s e n t a t i o n of W e t n e stress t e n s o r components in a form analogous t o eqs0(12.2a) (12.213). s h a l l not write o u t t h e s e expressions i n view of t h e i r g r e a t length. With respect t o t h e expressions found by us f o r t h e components of t h e displacement vect o r and of t h e stress tensor, we may remark t h a t they contain terms depending on t h e a c c e l e r a t i o n of an element of t h e s h e l l . These terms w i l l h e r e a f t e r b e designated i n e r t i a l r v .

The presence of i n e r t i a l terms d i s t i n g u i s h e s o u r approximation expressions f o r t h e displacement v e c t o r components and t h e stress t e n s o r from t h e expresIt i s obvious t h a t t h e s e expressions / W j sions known from t h e c l a s s i c a l theory. contain a number of non-inertial terms, which are a l s o absent from t h e r e l a t i o n s of t h e c l a s s i c a l theory. Section 13. Boundary Conditions The equations o f motion of an element of t h e s h e l l were obtained by u s from t h e equations of motion of a three-dimensional body. It w a s n a t u r a l a t first s i g h t t o t u r n t o t h e boundary conditions of t h e three-dimensional problem o f t h e theory of e l a s t i c i t y t o obtain t h e boundary conditions of t h e theory of s h e l l s . This i s e x a c t l y what we did. Consider two fundamental boundary problems. I n t h e f i r s t problem t h e displacements on t h e contour surface a r e prescribed and, i n t h e second problem, t h e stresses (11, Sect.8). The contour surface w i l l be 2 n a l y t i c a l l y detemnined by t h e following conditions imposed on t h e u n i t vector n of t h e e x t e r n a l normal:

1 . The F i r s t Boundary Condition

O n t h e contour surface C, l e t t h e displacements

(rt);L))c = cp, ( X I ,

2,

t)

(i = 1, 2, 3;

j = 1, 2).

(13.2)

b e prescribed. Expanding t h e prescribed displacements i n t e n s o r s e r i e s i n powers of z, w e find

(UP)), = 8 1( X I ,

0,

t ) zcpl (x, 0,t )

+r 2

.2cp5*

(x, 0, t)

Equating t h e first f o u r terms of t h e expansion (13.3) t o t h e first f o u r terms 2 f t h e expansions (12.2a) - (12.3b), we f i n d

112

- - ~ . ~ - ~ - - - ~ ~ 1 1 1 1 1 . . 1 ~ 1 1 . 1 1 1 1 1 . 1 1 1 1 1 . 1 1 1 1 1 1 . 1 1 . .

I1111.111111111

1111 I 1 1 1 1 1 1 1 1 1 . 1 . 1 1 1 . 1 1 1

I I1 I 1111 11111 1111111111

where C i s an a r c of t h e contour of t h e b a s i c surface of t h e s h e l l , The condit i o n s (13.4a) - (13.4.b) were obtained by us as a r e s u l t of a formal operation. The t o t a l number of t h e s e conditions i s twelve. W e have two remarks t o make on t h e conditions (13.4a)

(13.4b).

m
-

1 . The compatibility o f t h e conditions (13.4a) - (13.4b) with e q s . ( l l 0 2 l a ) (11.22b) i s not obvious. Apparently some of t h e s e conditions (13.4a) (13.4b) cannot be s a t i s f i e d by s o l u t i o n s of t h e system of equations (11.21a) I n f a c t , t h e o r d e r o f t h e s y s t e m of equations (11.21a) - (11.2lb) i s (11.21b). twelve. I f we r e c a l l t h a t t h e s o l u t i o n of a p a r t i a l d i f f e r e n t i a l equation of second order permits s a t i s f a c t i o n o f one boundary condition+, while t h e solut i o n o f a biharmonic equation satisfies two boundary conditions, then t h e solut i o n s of t h e system of equations (11.21a) - (11.21b) must s a t i s f y six boundary conditions. I n o t h e r words. w e s h a l l have t o confine ourselves t o t wo terms i n t h e expansions (13.3) and aicordingly t o two terms in t h e expansions (12,2a) ( 1 2 . 2 b ).

Obviously, t h e arguments presented here a r e not rigorous. The mentioned questions r e q u i r e s p e c i a l i n v e s t i g a t i o n . W e s h a l l r e t u r n t o them l a t e r . 2. I n problems of t h e s h e l l theory, t h e functions cp(xj, z, t ) a r e u s u a l l y n o t prescribed b u t it i s required t o s a t i s f y , by conditions imposed on t h e wanted functions, weaker r e s t r i c t i o n s on t h e contour of t h e b a s i c surface. Thus, t h e above-mentioned d i f f i c u l t i e s do not arise i n p r a c t i c e . 2. Second EoundaSy Problem Let u s a s m e t h a t , on t h e contour surface, t h e stress v e c t o r

i s prescribed.
35

Ekpanding this v e c t o r i n a t e n s o r Taylor+%eries i n powers of z,

The D i r i c h l e t and PJeumann problems are examples.

-H- W e

r e c a l l t h a t an expansion i n a t e n s o r power series b r i n g s about t h e operat i o n of p a r a l l e l displacement i n t h e Levi-Civita sense.

113

we find

L e t us now make use of eq.(II,

$.a):

where O i k and $+are, respectively, t h e components of t h e s t r e s s t e n s o r and o f t h e u n i t v e c t o r n of t h e e x t e r i o r normal t o t h e contour surface, displaced para l l e l t o themselves on t h e b a s i c surface o f t h e s h e l l along t h e normal t o t h i s surface. / 1 1 7 The p a r a l l e l displacement of t h e stress t e n s o r i s accomplished by expand-_, i n g i t s component,s i n t e n s o r s e r i e s . The p a r a l l e l displacement of t h e v e c t o r n The p o s s i b i l i t y i s performed on t h e basis of previous statements 11, Sect.11). of a separate displacement of t h e s t r e s s t e n s o r 2~ and t h e v e c t o r results from t h e fundamental p r o p e r t i e s of t h e operation o f p a r a l l e l displacement i n t h e Levi-Civita sense (I, Sect.10).

It follows from (I, 11.13) t h a t t h e r e l a t i o n s (13.1) remain v a l i d for t h e a r e deThe remaining components of t h e displaced v e c t o r displaced v e c t o r termined by equations r e s u l t i n g from (I, 11.13) and (I, 11.18):

z.

(i = 1, 2; do n o t sum over i1). Here, ni a r e t h e components of t h e u n i t v e c t o r p o i n t o f ' t h e s h e l l with t h e coordinate 2 = z.

t o t h e contour surface a t t h e Then e q ~ ~ ( 1 3 . 6 l) ead

Let u s expand ntZ, i n a Taylor s e r i e s i n powers of t o t h e following expressions :

Z.

To s e t up t h e boundary conditions w e must b e a r i n mind eqs.(a) and t h e expansions (12.3a) - (12.3c), eq.(13.5), and f".iLa (13.7). W e have

I n t h e expanded form, t h e s e equations a f t e r equating t h e c o e f f i c i e n t s of equal powers of z on t h e l e f t and right-hand s i d e s l e a d t o t h e following system of boundary conditions:

114.

(p=O,l,2,3;

i,j,r,s=l,2).

The s e l e c t i o n of t h e approximation m i s so performed t h a t , i n t h e conditions ( 1 3 . 8 ) , no terms of t h e 'rorderlt h4 will enter. The o r d e r s of h and z are The summation over r i s de-&g e assume t h a t = taken t o be t h e same-::. W noted by t h e u s u a l convention.

us'") us.

The system of r e l a t i o n s (13.9a) - (13.9b) contains twelve conditions. The above remark 1, on t h e number o f conditions of t h e boundary problems of s h e l l theory i n our formulation, a l s o a p p l i e s here. The question as t o t h e number and meaning of t h e boundary conditions i n s h e l l theory i s not new. Over a hundred y e a r s ago t h e r e w a s a discussion between t h e followers of Poisson's theory, according t o whom f i v e f o r c e condit i o n s had t o be s a t i s f i e d on t h e contour of t h e c e n t r a l plane or middle surface o f a p l a t e , and adherents o f t h e theory o f Kirchhoff, who a s s e r t e d t h a t t h e number of t h e s e conditions d i d not exceed four. The Kirchhoff theory, using t h e well-known simplifying static-geometrical hypotheses, i s g e n e r a l l y recognized a t t h e present time. The i m p o s s i b i l i t y of s a t i s f y i n g a l l t h e boundary conditions of t h e f i r s t or second boundary problems of s h e l l theory n a t u r a l l y l e a d s t o t h e i d e a t h a t t h e r e must be some i n t e r n a l c o n t r a d i c t i o n i n t h e theory developed by u s a s a whole. Indeed, t h e accuracy of t h e boundary condition t h a t can be s a t i s f i e d will be lower than t h e accuracy of t h e system of equations (11.21a) - (11.21b),which n a t u r a l l y r a i s e s t h e question whether t h e s e equations a r e not excessively acc u r a t e and u n j u s t i f i a b l y complex.

It is, however, easy t o prove t h a t t h e theory developed here contains no l o g i c a l contradictions. W e s h a l l r e t u r n l a t e r t o i t s evaluation. However, t h e s e and similar questions encourage t h e study of o t h e r a n a l y t i c a l approaches t o t h e mathematical d e s c r i p t i o n o f t h e s t r e s s e d and s t r a i n e d s t a t e of s h e l l s . One of them 5s based on t h e use of t h e v a r i a t i o n a l p r i n c i p l e s of t h e mechanics of e l a s t i c bodies+:-:$. M e s h a l l make use of this method l a t e r , and s h a l l then
$5 I n o t h e r words, terms with f a c t o r s haan, where m + n > 3 , must not e n t e r int o t h e equations.

-x% V.V.Bolotin

has c a l l e d our a t t e n t i o n t o t h e advantage of making extensive u s e of v a r i a t i o n a l methods i n t h e general theory of s h e l l s .

lI-5

r e t u r n t o t h e general, a n a l y s i s of t h e formulation of t h e boundary problems of t h e dynamics of s h e l l s . Section

14. I n i t i a l Conditions. General Remarks on t h e F i r s t

Version of t h e Solution of t h e Problem o f Reduction

To complete our b r i e f o u t l i n e of t h e general formulation of t h e dyramic boundary problems of t h e theory of s h e l l s i n t h e first version, l e t u s consider the i n i t i a l conditions. W e s h a l l s t a r t out from t h e i n i t i a l conditions of t h e Let dynamics of a three-dimensional e l a s t i c body ( 1 1 , $.la-b).
%o ( X i ) =

si (x,

2);

uio =8, (x,

2).

(14.1)
z, we f i n d

m a n d i n g t h e s e v e c t o r s i n t e n s o r s e r i e s i n powers

01

Making use of t h e expansions (12.2a) t i a l conditions:

(12.2b), we f i n d t h e following ini-

116

I n a l l , twenty-four i n i t i a l conditions (14.3a) (14.3d) must be s a n s f i e d . From t h e expressions f o r t h e o p e r a t o r s ( 1 1 , l O a ) - ( l l . U b ) , it i s c l e a r t h a t t h e system of equations (11.21a) - (11.21b) i s of t h e order twelve with (11.21b) i s of t h e r e s p e c t t o t h e t i m e t. Each of t h e equations (11.21a) f o u r t h order i n t, containing t h e wave operator X2. It i s c l e a r from t h i s t h a t t h e s o l u t i o n s of t h e system of equations (11.21a) (11.21b) can s a t i s f y only twelve i n i t i a l conditions. The remaining twelve conditions w i l l not be satis(11.21b) cannot, f i e d . Consequently, t h e s o l u t i o n s of t h e system (11.21a) ! o r with t h e accuracy prescribed by us, i.e., with an accuracy t o terms o f t h e ! der" h3, describe t h e i n i t i a l d i s t r i b u t i o n of t h e displacements and veloci- /120 t i e s i n t h e shell+. Obviously, even i n f u t u r e motion, t h e s o l u t i o n s of t h e system of equations (11.21a) (11.21b) w i l l not describe t n e f i e l d s of d i s placements and v e l o c i t i e s with t h e required accuracy.

A l l t h i s f o r c e s u s t o conclude t h a t s a t i s f a c t i o n of t h e boundary and i n i t i a l conditions with an accuracy t o terms of t h e '!order'* h " i s possible only i f t h e order of this system of equations (11.21a) - (11.21b) i s increased, which can be accomplished by introducing i n t o t h e s e equations terms with f a c t o r s h*, h5, h6, and h7, The system of equations (11.21a) - (11.21b) with terms t o t h e "ordertt h3 i n c l u s i v e may be u s e f u l i n t h e study of dynamic processes t h a t do n o t r e q u i r e rigorous s a t i s f a c t i o n of t h e i n i t i a l and boundary conditions. These problems include t h e problem of t h e propagation of p e r t u r b a t i o n s i n unbounded rods, p l a t e s and s h e l l s , t h e problem of l o c a l and very b r i e f i n f l u e n c e s caused by impact, etc.

This method permits obtaining eq;. (11.21a) - (11.21b) t h a t contain terms which can be i n t e r p r e t e d t o be a r e s u l t of t h e influence of shear s t r e s s e s and of t h e i n e r t i a of r o t a t i o n of an element of t h e shell's*. The appearance of t h e s e terms i n e q s . ( l l 0 2 l a ) - (11.21b) involves none of t h e kinetic-geometrical hypotheses t h a t have been introduced i n a number of modern works, b u t ils instead t h e r e s u l t of t h e a n a l y t i c construction of eqs.(ll.2la) - (11.21b).

02:)

To summarize, it may be s a i d t h a t t h e above method of expansion i n s e r i e s corresponding t o t h e b e s t approximation of t h e required functions !!at a point" permits u s t o construct+ssc a l o g i c a l l y non-contradictory technique f o r reducing t h e three-dimensional problems of t h e theory of e l a s t i c i t y t o two-dimensional
if In t h e absence of s u r f a c e forces, e q s 0 ( l l . 2 l a ) terms with t h e f a c t o r s ho and h2

(11.21b) w i l l contain and

M See, f o r instance, t h e above- si.ted d i s s e r t a t i o n s by 14.P.Petrenko 1.T.Selezov.

\ , > , _ ~ _ ~ ' optimum ~ ' ~ r e p r e s e n t a t i o n of "in-the-meantl The
# .

function-. -.qd i t s applicatfon

t o s h e l l theory w i l l be discussed l a t e r .

117

problems.

W e note two shortcomings of t h e method.

1 . The s a t i s f a c t i o n of boundary and i n i t i a l conditions with a prescribed accuracy by convention r e q u i r e s a r e l a t i v e l y high accuracy of eqs. (11.21a) (11.21b). For example, t o satisfy t h e boundary and i n i t i a l conditions with an accuracy t o terms containing f a c t o r s of t h e o r d e r of z3 reqi!.ires u s t o r e t a i n terms with f a c t o r s up t o h ' i n c l u s i v e , i n e q s . ( l l 0 2 l a ) - (11.21b). This shortcoming i s i n p a r t due t o t h e i t e r a t i o n process employed by us, which lowers t h e However, as will be c l e a r (7.4d). o r d e r of t h e system of equations (7.La) from t h e concluding remaz-ks t o Section 1 1 , t h e o r d e r o f t h e system o f equatior! (7.ka) - (7.4d) i s a l s o i n s u f f i c i e n t t o s a t i s f y t h e boundary and i n i t i a l con-/121 d i t i o n s with an a r b i t r a r y accuracy equal t o t h e a r b i t r a r y accuracy o f these equations.

Consequently, r e l a t i v e l y s l i g h t e r r o r s i n t h e preliminary de5ermination of t h e s t r e s s t e n s o r components oi3 lead t o g r e a t e r e r r o r s i n t h e subsequent determination of t h e f i e l d s of displacement, t h e v e l o c i t y of displacement, and t h e s t r e s s t e n s o r a s a whole. The index o f v a r i a b i l i t y i s o f considerable s i g n i f i c a n c e i n t h e problem o f s e t t i n g up approximation formulas t h a t d e s c r i b e k i n e t i c phenomena i n s h e l l s with s u f f i c i e n t accuracy (Bibl.5, 27, 29). According t o m o t h e r author(E'ibl.27) we may assume t h a t n e g l e c t i n g t h e terms t h a t contain t h e f a c t o r hn w i l l i n t r o duce an e r r o r of t h e o r d e r o f a n , where r i s t h e index of v a r i a b i l i t y , ar!d a

+lT

i s a dimension c h a r a c t e r i s t i c f o r t h e b a s i c s u r f a c e of t h e s h e l l . But t h e qv.est i o n of evaluating t h e e r r o r may become more complicated when w e consider t h e s o l u t i o n s o f refined equations. This i s confirmed by t h e existence o f boundary e f f e c t s of new types, discovered on an a n a l y s i s of t h e s o l u t i o n of t h e refined s t a t i c a l equations given by E.Reissner (Bibl. 20a).
Questions connected w i t h t h e c h a r a c t e r i z a t i o n o f t h e accuracy or" approximate dynamic equations by means of t h e index of v a r i a b i l i t y a r e s t i l l i n t h e s t a g e of study, and w e s h a l l not consider them here.". 2. The system of equations (11.21a) - (11.21b) i s very complicated. It i s e n t i r e l y possible t h a t t h e r e e x i s t s i m p l i f i c a t i o n s o f t h i s system, which have only a n e g l i g i b l e e f f e c t on t h e f i e l d s of displacement and s t r e s s . The method employed gives no answer t o t h i s question.

L e t u s pass now t o o t h e r methods o f solving t h e reduction problem and o f formulating t h e dynamic boundaqy problem o f s h e l l theory.
x- The s t a t u s of t h e problem a t t h e present time i s given by another author (Bibl.2Clb). The complexity of t h e problem i s increased by t h e introduction of

i n e r t i a l terms i n t o t h e boundary conditions, when c e r t a i n methods of reduction a r e used.

1 1 8

Section 15. ApFlicatior, of t h e General B u a t i o n OZ D~mamicst o t h e S o l u t i o 3 of t h e Problem of Reduction Let, us make u s e of t h e general equation of dynamics, s e t up with respect t o t h e motion o f an e l a s t i c body:>:

(15.1)

where Xi a r e t h e f o r c e s a c t i n g on t h e surface S on t h e body, and 6k i s t h e e l e x e n t a v wo& o f deformation defined by (11, 11.1). The o t h e r n o t a t i o n s a r e arnil i a r

iJe r e c z l l t h a t eq.(15.1) includes a l l t h e forms of t h e equation of mo- /122 t i o n 01 an e l a s t i c body. Q u a t i o n (15.1) y i e l d s t h e s o l u t i o n of t h e reduction problein and m k e s it p o s s i b l e t o Tormulate t h e dTynamic boundary problems of t h e theory of s h e l l s . The fundamental method of reduction r e s u l t i n g from eq.(15.1) i s an approxima-Lion o f t h e components of t h e displacement v e c t o r and t h e s t r e s s t e n s o r by f i n i t e sums of f u n c t i o n s of t h e coordinate z, s e l e c t e d i n a d e f i n i t e 1,.ray and having c o e f f i c i e n t s depending on t h e i n t e r i o r coorclinates x j of t h e b a s i c surfa.ce o f t h e s h e l l . This scheme includes most of t h e methods known today f o r solving t h e problem of reduction o f a three-dimensional problem of t h e theory of e l a s t i c i t y t o a two-dimensional problem. An exception i s t h e method considered i n t h e l a s t few Sections, since it does not involve an i n t e g r a t i o n o f approximation funct i o n s over t h e coordinate Z .
The p o s s i b i l i t y of applying t h e general equation o f d;mamics t o t h e solut i o n o f t h e problem of reducing t h e three-dimensional s t a t i c problem of t h e theory of e l a s t i c i t y t o a two-dimensional problem o f t h e theory of s h e l l s has been noted by V.Z .Vlasov i n h i s monograph (Eibl.3a). Kh.M.Nushtari and 1.G.Teregulov d i s c u s s this problem i n t h e nonlinear forrmilation i n g r e a t d e t a i l (Bibl. 27 )

The method t h e method used by u s below, i n > a i l i n Section

of reduction i n d i c a t e d in a monograph (Bibl.3a) d i f f e r s from i n m o t h e r paper (Bib1.27) as w e l l a s f r o m t h e method developed t h a t it i s l e s s general. W e compare t h e s e methods i n more de-

30.

The eleConsider i n succession t h e q u a n t i t i e s e n t e r i n g i n t o eq.(15.1). nent of volume d V and t h e element of a r e a dS(*].of t h e boundary s u r f a c e s of t h e s h e l l a r e expressed by t h e following equations.

-%

C f . , or example, L.S.Leybenzon, .Acadeq of Sciences, 1951.

Collection of Works, Vol.1,

pp.193-194.

USSR

119

W e confine o u r s e l v e s here t o t h e consideration of s h e l l s of constant thickness 2h. The r e l a t i o n s (I, 2.6b) and (I, 3.6a'- 3.6b) are used here (Bib1.13). The element of area dSC of t h e contour s u r f a c e C i s defined by t h e equation

a r e t h e parametric equations o f t h e contour of t h e b a s i c s u r f a c e of t h e s h e l l . Since w e will make u s e of segments o f t e n s o r power s e r i e s , which approximately determine t h e v e c t o r u, t h e v a r i a t i o n s h i , and t h e stress t e n s o r d k on t h e b a s i c surface, we s h a l l d i s p l a c e t h e v e c t o r s pF, and X, t o t h e b a s i c surface, u s i n g t h e o p e r a t o r s of p a r a l l e l displacement (I, 11.20). To avoid complicating t h e formulas, we s h a l l r e t a i n t h e previous n o t a t i o n f o r t h e displaced pFI a n d m X , . W e put f u r t h e r

uY)= U,+ZU:') + 1 z@) + 1

2 The q u a n t i t j 2s q ( ' ) , up), shell.

+ ....
1 ~ )

3 1

(15.5)

uy'.

..

a r e t h e generalized coordinates o f t h e

W e shall- h e r e a f t e r confine ourselves t o t h e same conditional accuracy adopted by u s i n considering t h e f i r s t version of t h e s o l u t i o n of t h e reduction problem. I n view of t h e f a c t t h a t eq.(15.5) determines t h e v e c t o r displaced t o t h e b a s i c surface, w e find

(do not sum over i I ) . Consider now 6A.

By (11, 11.1) we have+:-

Let u s now t a k e up t h e transformation of t h e sum

WZ' = a;;
x- Cf.

SOL).

a l s o t h e above-cited work by L.S.Leybenzon. 12G

Bearing i n mind t h e commutativity of t h e operations o f v a r i a t i o n s and covariant differentiation,

6VU!) =Vi6U6),

(15.9)
Vi6Ut.

we o b t a i n
6wz= 1 a,) iA 2

(visut+Vk6UF)) =a;)

(15.lo)

W e find, further, sWz = 6W

1 z26w(2) 1 z36w(3) +Z6W +++ ... =

3 1

The c o e f f i c i e n t s a i t ! , o f t h e expansions o f t h e s t r e s s t e n s o r i n powers o f z .have t h e following meaning :

Here,

(15.13)
1 2 1

a r e t h e components of t h e stress t e n s o r on t h e b a s i c surface. To prepare all t h e summands e n t e r i n g i n t o t h e v a r i a t i o n a l equation (15.1) f o r t h e forthcoming transformation, l e t u s consider t h e sum-a 2 q W , r e t a i n i n g

a t2

i n it a l l t e m s up t o terms with t h e f a c t o r z3 i n c l u s i v e . e q ~ ~ ( 1 5 . 5) (15.6), we f i n d

Making use o f

(15.14)

D i f f e r e n t i a t i o n w i t h r e s p e c t t o time i s denoted here and h e r e a f t e r by tremas.

1 5 . 1 1 ) , and N o w l e t us s u b s t i t u t e t h e expressions (15.3), (15..L), (15.6), ( (15.lh) i n t o t h e v a r i a t i o n a l equation ( 1 5 . 1 ) . L e t u s i n t e g r a t e over z y under

t h e assumption t h a t t h e b a s i c surface coincides with t h e c e n t r a l plane of t h e m s h e l l , and confining o u r s e l v e s t o summands with t h e f a c t o r s h, h2y and h" inle introduce t h e n o t a t i o n

Here w e made u s e of eq.(15.4). The components of t h e s t r e s s v e c t o r X, can be expressed, i f convenient, i n terms of t h e s t r e s s t e n s o r by eqs.(l3.E). W e vectors erators are the r e c a l l again t h a t when w e apply t h e general equation of d y r " i c s , a l l of f o r c e s a r e f i r s t displaced t o t h e b a s i c surface by means of t h e opof p a r a l l e l displacement (I, 11.22). The q u a n t i t i e s 4caJi and S ( a ) * r e s p e c t i v e generalized f o r c e s on t h e b a s i c and contour s u r f a c e s corre122

sponding t o t h e generalized coordinates uif

.
.

After s e v e r a l transformations and a p p l i c a t i o n o f t h e Ostrogradskiy-Gauss theorem, t h e general equation of dynamics (15.1) t a k e s t h e following approximate fornr;:t:

45

Ne w r i t e o u t t h i s equation, r e t a i n i n g terms up t o t h e order h3 i n c l u s i v e , i n t h e semi-developed form, t o make t h e book e a s i e r t o read. O f course, it i s q u i t e simple t o shorten t h e n o t a t i o n here.

123

(i, k = 1, 2).

( 1 5 . 1 6)

where w i s the area. o f t h e b a s i c (middle) surface, and C i s t h e contour or" t h e w and t h e element of a r c d s of t h e conrniddle surface. The element o f a r e a d C t o u r of t h e middle surface, based on t h e r e l a t i o n ( 1 5 . 4 ) , a r e expressed as f o l lows :

The v a r i a t i o n a l equation (15.16 ) y i e l d s a system o f approximation equat i o n s f o r t h e v i b r a t i o n s o f a s h e l l , together with t h e boundary conditions. Section

1 6 . D i f f e r e n t i a l Fiquations of t h e O s c i l l a t i o n s

of a S h e ll

/127

Assume t h a t t h e only c o n s t r a i n t s inyosed on t h e s h e l l a r e on t h e contour surface. Then, t h e v a r i a t i o n s &up) i n t h e region u: a r e a r b i t r a r y independent q u a n t i t i e s , and from t h e i r v a r i a t i o n a l equation (15.16) follows t h e vanishing o f t h e c o e f f i c i e n t s of t h e s e v a r i a t i o n s . FQuating t h e s e c o e f f i c i e n t s t o zero, we o b t a i n t h e following s y s t e m of d i f f e r e n t i a l equztions:

1 2 1 ,

.. .

(16 2 a )
2h3 . . ( I ) 2h3 -pa3 - -p (k, k,) k3 3 3 2h3 1 -V i (k, ka) ai3 - h30 ( ) : 3 3

-2Vi at:))

- Q(*) =0;

(16.~~)

( 16 4a)

Thus, from t h e v a r i a t i o n a l equation (15.1), with t h e accuracy of approximation adopted by us, we obtained a system of twelve equations which, taken tog e t h e r with t h e r e l a t i o n s (15.12a) - (l5.l2c), determine twelve unknown functions q @ ) (m = 0, 1, 2, 3; i = 1 , 2, 3). Let u s make a b r i e f a n a l y s i s of /128 t h i s system.
1 . Equations (16.4a) - (16.4b) permit a d i r e c t determination of t h e "norm a l part" of t h e s t r e s s tensor. W e know from t h e first version of t h e s o l u t i o n o f t h e reduction problem t h a t t h e determination of d 3 ( i = 1, 2, 3 ) i s s u f f i cient f o r i t s solution i f w e have recourse t o t h e Lame/ equations.

The mechanical meaning of eqs.(16.4a) - (16.kb) i s t h a t they express one = of t h e g e n e r a l i z a t i o n s of t h e Kirchhoff-Love hypothesis. I n f a c t , i f Q(3) - Q ( 3 ) 3 = 0, then it follows from eqs.(16.4a) - (l6.4b) t h a t
&3

= a3R = 0

(k = 1 , 2).

(16.5)

(16.5) a r e l e s s accurate than t h e expressions (7.5b). Nevertheless, t h e f a c t of a d i r e c t donnection between t h e Kirchhoff-Love 11hy-pothesis17 and t h e approximation equations of motion r e s u l t i n g It may b e s t a t e d t h a t from t h e v a r i a t i o n a l equation (15.1) deserves a t t e n t i o n . t h e Kirchhoff-Love ??hypotheses" are a simple a n a l y t i c consequence of t h e condit i o n a l and prescribed accuracy f o r t h e equations of t h e two-dimensional problerr of t h e theory of e l a s t i c i t y and f o r t h e s p e c i a l hypotheses about t h e f o r c e s a c t i n g on t h e s h e l l .
O f course, t h e r e l a t i o n s

(7.5a)

2. Bearing i n mind eqs.(l5.l2a) (15.12c), we can f i n d t h e order o f t h e system of equations s e t , u p by us, including eqs.(16./+a) - (16.kb) i n t h i s syst e m , a u a t i o n s (16.la) - (16.3b) are equations of t h e second order i n deriva, 2 ) and t i v e s of t h e unknown f u n c t i o n s with respect t o t h e coordinates x i ( i = 1 t o t h e time t. Equations ( 1 6 , b ) - (16.Ib) are equations of t h e first o r d e r i n d e r i v a t i v e s w i t h r e s p e c t t o t h e coordinates. Time d e r i v a t i v e s do not e n t e r i n t o t h e s e equations. consequently, w e have obtained a system of t h e t w e l f t h ord e r i n d e r i v a t i v e s with r e s p e c t t o t h e coordinates, and of t h e eighteenth o r d e r i n d e r i v a t i v e s with r e s p e c t t o time. N e r e c a l l t h a t t h e system o f equations (7.La) (7.4d) i s a system of t h e twenty-first order and t h e system of equations (11.21a) - (11.21b) a system o f t h e t w e l f t h order. The mixed d e r i v a t i v e s w i t h r e s p e c t t o t h e coordinates and t o time belong t o t h e highest order with r e s p e c t t o t h e d e r i v a t i v e s e n t e r i n g Mixed d e r i v a t i v e s o f t h i s type do not e n t e r i n t o i n t o eqs.(7.1La) - (7.Ld). eqs. (16.1~~) - (16.4b).

3. In s e t t i n g up e q s 0 ( l 6 . l a ) - (16.1Lb), t h e o p e r a t i o n o f d i f f e r e n t i a t i o n i s not performed on t h e components o f t h e force vectors. a u a t i o n s (7.4a) (7.fl.d) a r e s e t up under t h e assumption t h a t a d i f f e r e n t i a t i o n o f t h e components of t h e v e c t o r s o f body f o r c e s i s permissible. T h i s g i v e s a c e r t a i n advantage t o eqs.(lb.la) - (16.4b) over t h e equations s e t up according t o t h e f i r s t vers i o n (Sect . u ) .
14. The d i f f e r e n c e i n t h e composition o f e q s 0 ( l 6 . l a ) (16.1kb) and t h a t o f eqs.(7.ka) - (7.I+d) can be explained b y t h e theory o f approximation functions. Q u a t i o n s (".ha) - (7.Ld) a r e s e t up according t o one o f t h e methods of opti-/129 mum "at-a-point" approximation ?unctions, while eqs. (16.la) - ( 1 6 4 1 ) correspond t o one of t h e methods o f optimum representation of "in-the-mean" functions.
Section 17. Natural Boundary Conditions Derived from t h e Gariational Equation (15.16) Let u s pass now t o t h e consideration o f t h e i n t e g r a l over t h e contour of A study of t h i s i n t e g r a l pert h e b a s i c surface t h a t e n t e r s i n t o eq.(15.16). mits us t o e s t a b l i s h v a r i o u s v e r s i o n s of t h e boundary conditions. Iiere, t h e vanishes if a l l t h e components under t h e s i g n of i n t e g r a t i o n l i k e integral

(C)

wise vanish-::-.

For t h e s e summands t o vanish, one o f t w o conditions must be s a t i s f i e d : E i t h e r t h e corresponding v a r i a t i o n &I$) ' must vanish o r t h e c o e f f i c i e n t o f t h a t v a r i a t i o n must vanish. Terrns i n t h e v a r i a t i o n s 6 % ' do not e n t e r i n t o t h e expression under t h e i n t e g r a l s i g n over t h e contour of t h e b a s i c surface. Thus, t h e v a r i a t i o n a l equation (15.16 ) y i e l d s nine boundary conditions, corresponding t o t h e v a r i a t i o n s 6 q , 6rq('), 6 % ( " ) .
Since t h e order of t h e system o f ecpations (16.la) - (16.1Lb) r e l a t i v e t o
:+ The n e c e s s i t y o f this condition i s proved i n courses on t h e p r i n c i p l e s o f anal y t i c mechanics.

126

t h e d e r i v a t i v e s with r e s p e c t t o t h e coorciinates i s twelve, it may be assumed t'nat t h e system of n a t u r a l boundary conditions r e s u l t i n g from t h e v a r i a t i o n a l equation (15.16) i s compatible with t h e s e eqcations. Obviously any sim-plificat i o n of t h e system o f equations (16.1~~) (16.4b) must be accompanied by a change in t h e boundary conditions. The question of t h e compatibility of t h e boundary conditions with t h e system of fundamental equations must be subjected t o a s p e c i a l a n a l y s i s i n s p e c i f i c problems.

L e t u s now consider, as an e-uample, s e v e r a l versions o f t h e bomdary conditions.

1 . With Rigidly Attached Contour Surface

I n t h i s case, w e obviously have

Ye do not impose conditior;s on since t h e v a r i a t i o n s do not e n t e r into the integral and t h e conditions imposed on ui(3) will not be natural.
(C 1

up),

2. wit;? Free Contour Surface

/13 0

I n this case, t h e v a r i a t i o n s & u p )(i = 1 2, 3 ) z r e a r b i t r a r y . >/e o b t a i n t h e n a t u r a l conditions by equating t h e c o e f f i c i e n t s of these v a r i a t i o n s t o zero. On t h e b a s i s of eq.(15.16), we f i n d

I n t h e o t h e r cases, t h e boundary conditions are mixed. Their formulation depends on t h e scope of t h e s p e c i f i c problems of t h e mechanics o f s h e l l s . Section 18. I n i t i a l Conditions The question o f t h e i n i t i a l conditions cannot be solved by analogy t o t h e

127

question of t h e boundary conditions from a d i r e c t study of t h e v a r i a t i o n a l equae s t a r t . f r o m considerations similar t o those set f o r t h i n 1 5 . 1 6 ) . If w tions ( Sect.14, we will again o b t a i n twenty-four i n i t i a l conditions which will not be s a t i s f i e d by s o l u t i o n of t h e system (16.la) - (16..4b), s i n c e t h e o r d e r of this system wtth respect t o t i n e i s eighteen. Obviously some of t h e generalized coo r d i n a t e s mst obe t h e i n i L i a l conditions, f o r example only t h e generalized coordinates u l , q q(2) (i = 1 , ' 2, 3). I n t h a t case, t h e number of i n i t i a l conditions w i l l be eqv.al t o t h e order of t h e system.

(5,

The l i m i t a t i o n imposed on t h e number of i n i t i a l conditions i s confirmed also by considerations based on t h e p r i n c i p l e s of a n a l y t i c a l mechanics of d i s c r e t e systems. A comparison of t h e formulation of t h e problems under study w i t h t h e formulation of t h e c l a s s i c a l problems of a n a l y t i c a l mechanics permits u s t o r e f i n e t h e meaning o f t h e i n i t i a l condition sought. Xe r e c a l l t h a t t h e s e t o f q u a n t i t i e s determining t h e i n i t i a l conditions i n t h e problems of t h e motion of systems of material p o i n t s e n t e r i n t o t h e t o t a l time d e r i v a t i v e s which appear on t h e left-hand s i d e of t h e general equation 1131 of dynamics;':-. 'de s h a l l perform t h e transformation of only one summand i n t h e 1 5 . 1 4 ) . This will permit left-hand s i d e of t h e general equation of dynamics ( e have us, by analogy, t o w r i t e o u t t h e required expression completely. W

Hence we conclude t h a t t h e function determining t h e i n i t i a l conditions i s o f t h e following form:

$5

These q u a n t i t i e s subsequently form terms o u t s i d e t h e i n t e g r a l sign, which appear i n t h e proof o f t h e Ostrogradskiy-Hamilton p r i n c i p l e and vanish when t h e paths of comparison a r e properly chosen.

128

pit3 (k,
Uk

1 + k,) +3 h3piiz)] 6u,+

pgkkU61

+ [- 2h3 3pgkk ( k , + k,) +2h3 3

aut+

( 1 8 . 1 )
The composition of t h e f u n c t i o n Q confirms our preliminary statement t h a t t h e i n i t al conditions n problems of t h e dynamics of s h e l l s , under t h e condit i o n a l accuracy of t h e equations here adopted, a r e expressed as follows:

( 1 8 . 2 )
Consequently, here too, t h e conditional accuracy or t h e eqLations per- /132 mits u s t o f i n d t h e displacements only w i t h an accuracy t o terms containing t h e f a c t o r z i n c l u s i v e , although f o r s e t t i n g up the B u d m e n t a l system of equat i o n s we use terms of t h e form-1 z%(~ a d t h e eqcations contain terms with f a c t o r s h3

31

I n s p i t e of t h e presence of terms with t h e f a c t o r h3, i t wf.11 be noted t h a t t h e c o n d i t i o n a l accuracy of e q s 0 ( l 6 . l a ) - (lS.4b) i n c e r t a i n cases i s determined by t h e o r d e r of t h e terms containing h . I n Tact, i f t h e s u r f a c e vanish, then, as i s obvious from t h e equations of generalized forces X f o r c e s (15.15 a ) , t h e generalized f o r c e s w i l l be of t!ie order h. A f t e r term-byterm d i v i s i o n of t h e equations by h, w e obtain equations w i t h terms containing h i n a power n o t higher t h a n t h e second,
To s u m a r i z e our r e s u l t s , w e may note t h a t t h e a p p l i c a t i o n of t h e general equation of dynamics permits u s t o o b t a i n b e t t e r c o m p a t i b i l i t y o f t h e system of equations s a t i s f i e d by t h e wanted f u n c t i o n s on t h e b a s i c surface with t h e boundary arld i n i t i a l conditiofis, than can be obtained by using t h e methods i n d i cated i n S e c t e l l .

Section 19. O n Concentrated Forces de mentioned above t h a t t h e method considered i n Sect.11 r e q u i r e s a m i l l t i p l e d i f f e r e n t i a b i l i t y of t h e v e c t o r components of t h e a c t i v e f o r c e s applied t o t h e s h e l l . The p e r m i s s i b i l i t y of t h i s method i n t h e case of t h e a c t i o n of conc e n t r a t e d surface or body f o r c e s on t h e s h e l l becomes doubtful. Of course, t h e s e doubts a r e connected with an o b j e c t having no physical existence, namely,

1 11111 I I

, ,,

,,, ,, ,-,

_.... ._

t h e concentrated force. A11 t h e sane, t h e concept o f concentrated force, for a l l i t s a b s t r a c t n e s s , occupies a d e f i r i t e p o s i t i o n among t h e concepts o f t h e mathematical theory or" e l a s t i c i t y and appears i n t h e form ol" a n a l y t i c singulari t i e s o f t h e corrponents of t h e displacement vector, t h e deformation tensor, and t h e stress tensor. It i s t h e r e f o r e n a t u r a l t o s t r i v e or an a n a l y t i c a l l x corr e c t i n t r o d u c t i o n o f concentrated f o r c e s i n t o t h e approximate r e p r e s e n t a t i o n s of t h e applied t h e o r i e s o f e l a s t i c i t y , and e s p e c i a l l y i n t o t h e s h e l l theory. The theory developed i n Sects.15 - 18 i n p o s e s fewer r e s t r i c t i o n s on t h e p r o p e r t i e s of t h e f o r c e s applied t o t h e s h e l l t h a n t h e f i r s t version o f t h e sol u t i o n of t h e reduction problem. The construction o f e q s 0 ( l 6 . l a ) - (16.Lb) does n o t r e q u i r e t h a t t h e components of body and s u r f a c e f o r c e s b e d i f f e r e n t i a b l e . These equ-ations a l s o app l y t o cases of t h e a c t i o n o f concentrated f o r c e s i f t h e components of t h e conc e n t r a t e d f o r c e s are expressed i n terms of t h e Dirac d e l t a function. Herea f t e r , i n solving eqs.(16,la) - (16.!+b) we must make u s e of o p e r a t i o n s t h a t do n o t include t h e d i f f e r e n t i a t i o n of t h e components of concentrated forces. It i s w e l l known t h a t t:his raquirerLent i s s a t i s f i e d i n a number o f s p e c i a l prob/133 The a p p l i c a t i o n of t h e theory o f general.ized f u n c t i o n s considerably exlems. pands t h e c l a s s o f t h e s e problems. Xe s h a l l a l s o i n d i c a t e a method t h a t does n o t r e q u i r e an e x p l i c i t applicat i o n o f t h e theory o f genera.l;ized functions. L e t us assume, f o r d e f i n i t e n e s s , t h a t t h e concentrated f o r c e P with components P, i s applied t o t h e boundary s u r f a c e z = +h. Then, from t h e well-known d e f i n i t i o n o f concentrated f o r c e s , w e have

where S , i s t h e region on t h e boundary surface t o which t h e p o i n t M of applicat i o n of t h e v e c t o r P belongs.

L e t u s denote by L(m) 1 (a, u ) t h e c o e f f i c i e n t s o f G U ~ ( ~ i) n t h e double i n t e Obviously L(") i s the gral e n t e r i n g i n t o t h e v a r i a t i o n a l equation (15.16). d i f f e r e n t i a l o p e r a t o r d e f i n i n g a cer22i.n s e t of o p e r a t i o n s t o b e pezformed on t h e components o f t h e s t r e s s t e n s o r I S and t h e d i s p l . a c c + n ~ + :r?ctor , u . Let

2-3

) i s a complete system of l i n e a r l y independent f u n c t i o n s o f two where (;tpq)(xJ are a r b i t r a r y . v a r i a b l e s XJ on t h e b a s i c surface, while t h e c o e f f i c i e n t s a&, Then, i n s t e a d of t h e system of equations ( 1 6 . h ) - (16.4b) which we obta' ed by under t h e sign o f i n t e g r a t i o n equating t o zero t h e c o e f f i c i e n t s of
(E)

i n eq.(15.16), i n g form:

we o b t a i n an i n f i n i t e s p t e m of i n t e g r a l r e l a t i o n s o f t h e follow-

p q ) (x')

(1 - R,h) (1 - R,h) h*
IM

=0 .

(19.3)

where t h e values of a l l q u a n t i t i e s i n b r a c k e t s a r e taken a t t h e p o i n t M o f app l i c a t i o n of t h e concentrated force P. r"urther determination 0;" t h e required q u a n t i t i e s f r o h eqs.(19.3) u s u a l l y l e a d s t o t h e s o l u t i o n of i n f i n i t e systems o f a l g e b r a i c equations. The modific a t i o n s of eqs.(19.3) i n t h e case of t h e a c t i o n o f a concentrated body f o r c e a r e obvious.

/l34 Let u s consider t h e case of t h e a c t i o n of a concentrated force on t h e contour surface of a s h e l l . I f t h e concentrated f o r c e P i s applied t o t h e cont o u r surface a t t h e p o i n t bl(sz-, z;'), then t h i s w i l l introduce i n t o t h e quantit i e s S(.) 1 t h e following a d d i t i o n a l terms:

x 6 (s, -sD.
where S ( s ,

(19.4)

sz-) i s t h e d e l t a function.

Let u s put, on t h e contour surface,

where J . ( p ) ( s c ) i s a complete system of l i n e a r l y independent functions of t h e , of t h e contour, while t h e c o e f f i c i e n t s b ( p ) arc s on those p a r t s o f t h e cont o u r t h a t are f r e e from kinematic c o n s t r a i n t s are a r b i t r a r y .
L e t u s denote t h e c o e f f i c i e n t s or" Gui@)i n t h e i n t e g r a l
2? ! $

..

The meaning of t h i s symbol i s v a r i a t i o n a l equation (15.16) by P I ( " ) f) analogous t o t h a t of t h e symbol L(m) i (0,U ) Then, t h e system of boundary cond i t i o n s w i l l l e a d t o a system o f i n t e g r a l r e l a t i o n s of t h e form

. .

entering i n t o the
(C)

131

The i n t e g r a l i s extended t o segwhere o(sC s = ) i s t h e Heavisidt function. ments of a r c of t h e contour t h a t are free from kinematic c o n s t r a i n t s . I f t h e concentrated f o r c e i s applied a t a p o i n t where, because of kinematic considerat i o n s , we must put 6u,(") as equal t o zero, then t h e corresponding boundary cond i t i o n s (19.6) l o s e t h e i r meaning. Equations (19.6) t i o n (19.3). Section 20. Second Version o f t h e S o l u t i o n T f t h e Problem of Reduction' I n Sect.11 we pointed o u t t h a t t h e r e e x i s t two methods of solving t h e ree duction problem i f w e s t a r t from t h e s y s t e m o f equations (7.4.a) - (7.&d). W did n o t d i s c u s s t h e second method and confined ourselves t o t h e statement t h a t this method w a s close t o t h e c l a s s i c a l theory o f s h e l l s . Sections 15 - 1 9 do n o t r e l a t e t o t h e second version, since we d i d not r e f e r t o eqs.(7.4.a) (7.4.d). Here, likewise, w e s h a l l not make use of t h e s e equations b u t s t a r t from t h e general equation o f dynamics ( 1 5 . 1 ) . The r e s u l t a n t equations w i l l be close /135 e will t h e r e f o r e i n form t o t h e equations of t h e c l a s s i c a l theory of s h e l l s , W e a r e r e a l l y going beyond t h e l i m speak here o f "the second version", although w i t s o f t h e scheme given i n Sect.11, supplement t h e system o f conditions of t h e form of equa-

5;Je Let u s r e t u r n t o t h e v a r i a t i o n a l equation (15.1) and t o eq.(15.10). Let u s expand t h e q u a n t i t i e s V, 6%(z) i n tens h a l l a l s o make use of eqs.(15.5). sor s e r i e s iri powers o f z , d i s p l a c i n g t h e m t h u s t o t h e b a s i c surface of t h e s h e l l , and l e t u s a l s o d i s p l a c e t h e s t r e s s t e n s o r components t o t h e b a s i c surf a c e along t h e coordinate l i n e s d = const, 2 = const, without expanding them i n s e r i e s , b u t using i n s t e a d t h e operators o f p a r a l l e l displacement given preThese components of t h e displaced s t r e s s t e n s o r w i l l be v i o u s l y (I, Sect.11). From eqs.(I, lLl), (I, 11.12) - (I, 11.U) and (I, 1 1 . 1 8 ) , we denoted by +k. have
& = A;Ai,aPq = is;

+a ; ) (6; + 0;)

ap4,

(a>

or
,til = (1 - k,Z)(1 - kjZ)uti;
Ti3

= (1 - R,z)013;

(20. l a )

733

=0 3 3

(20.lb)

(i, j = 1, 2; do not s u m over i and .jI).

As a r e s u l t we o b t a i n from eq.(l5.10;, ing relation:

8w'=' = TikViBU,

i n s t e a d of eq.(15.11),

t h e follow-

T i 3 (Vlbu3

+6uj") +

733&&)+

132

(20.2)

L e t us introduce t h e n o t a t i o n

+h

T'"')" = -!- zmuif (1 -- k , z ) (1 -biz) (1 - k,z) (1 - k,z) dz, m!

-h
.

+h
n

--h

T ( m ) 33 = nt !

-h

'-h

zmaJ3 (1

- R,z) (1

-k g j dz

The q u a n t i t i e s T ( . J i J a r e components of t h e second-rank t e n s o r on t h e midd l e surface, L e . , i n t h e set of coordinates xJ( j = 1 , 2 ) ; t h e q u a n t i t i e s T") l 3 are v e c t o r components in this s e t , and t h e q u a n t i t y T(m)33 i s a scalar. From the analytic-functional viewpoint, these q u a n t i t i e s are generalized f u n c t i o n a l moments about e, of an o r d e r d i f f e r e n t from that of t h e s t r e s s t e n s o r components.

W e s h a l l now r e t u r n again t o t h e general equation of dynamics ( 1 5 . 1 ) , and a f t e r transformationa we reduce it t o t h e following form:

133

The v a r i a t i o n a l up a two-dimensional t h e n a t u r a l boundary n o t a s y e t solve t h e

equation obtained here, l i k e eq. (15.16 ), permits s e t t i n g / l 3 7 system of d i f f e r e n t i a l equations of motion and t o i n d i c a t e conditions. These equations in themselves, however, do reduction problem.

Section 21. F i r s t Group of Bastodynamic EQuations of t h e Theory of S h e l l s Equating t o zero t h e v a r i a t i o n a l c o e f f i c i e n t s of t h e generalized coordin a t e s e n t e r i n g i n t o t h e expression under t h e s i g n of i n t e g r a t i o n over t h e midd l e surface w of t h e s h e l l i n eqs(20.5), we now f i n d t h e following equations of motion:

ph3(k1

+k,) gk'at) +

134

The system of equations (21.la) (2l.5) c o n t a i n s f i f t e e n equations w i t h t h i r t y - t h r e e unknown functions. The unknowns in t h e equation a r e t h e twentyf o u r moments about z of t h e stress t e n s o r components and t h e nine c o e f f i c i e n t s of t h e expansion in power s e r i e s of t h e displacement v e c t o r components. Thus, /138 i n s e t t i n g up e q s 0 ( 2 l . l a ) (21.5) we have been g u i l t y of a l o g i c a l inconsistency, caused by t h e s e l e c t i o n of t h e system of generalized coordinates and l e a d i n g t o equations containing both moments and c o e f f i c i e n t s of an ekpansion i n power s e r i e s . This inconsistency can be eliminated by s e l e c t i n g t h e generali z e d coordinates i n a d i f f e r e n t way. However, t h e system of equations (21.la) (21.5) obtained by t h e mixed method permits u s t o e s t a b l i s h a d i r e c t connect i o n with t h e equations o f t h e c l a s s i c a l theory of s h e l l s and t o analyze t h a t theory from t h e p o s i t i o n of a n a l y t i c a l mechanics.

Equations (21.4a) (21.5) do n o t contain i n e r t i a l terms. These equations permit f i n d i n g t h e moments of t h e t h i r d and second o r d e r of t h e components of t h e normal p a r t of t h e s t r e s s tensor. The meaning of equations (2l04a)-(2l.5) i s analogous t o that of t h e Kirchhoff-Love hypotheses, s i n c e they make it pos(2L.5) s i b l e t o solve t h e reduction problem. The system of equations (21.4a) i s indeterminate and must be supplemented by equations r e s u l t i n g from Hookets

law.
Section 22. Second Group of ELastodynamic .Equations of t h e Theom o f S h e l l s *
~

L e t u s express t h e moments of t h e stress t e n s o r components in terms of t h e c o e f f i c i e n t s of expansions i n s e r i e s i n powers o f z of t h e displacement vector, making use of Hookets law. Following t h e *"jxedt' method, we s h a l l make use of eqs.(20.3), expanding t h e stress t e n s o r components -rlk i n t o t e n s o r power series, i n ascending powers of z, t h u s accomplishing t h e p a r a l l e l displacement of t h e stress t e n s o r to t h e basic surface.

135

Making use of eqs.(15.12)

- (15.12c),

we f i n d

n -0

(i, k = 1, 2; do not sum over i and k 1 ) . The choice of N depends on t h e conditional accuracy prescribed f o r t h e equations t o be set up. If t h e equations mst n o t contain terms w i t h f a c t o r s h t o a power higher than t h e t h i r d , then N S 2. The number of generalized coord i n a t e s t o be determined in this case is twelve. S u b s t i t u t i n g t h e expresf139 sions (22,la) - ( 2 2 . 1 ~ ) i n t o eqs.(20.3), we f i n d t h e r e l a t i o n s between t h e moments of t h e components of t h e stress tensor and t h e generalized coordinates. W e have

136

..

! E g" ( V i U , + .I")+ .. . ,
3

A t t h e conditional accuracy adopted here, a l l t h e moments T(3) equated t o zero:

77'3' x.3

'

must be/l/O (22.5)

- 0; __

743133

- 0.

Equations (22.5) coincide with eqs.(2l.5). The systems (22.2a) (22.5) form t h e second group of elastodynamic equat i o n s of t h e s h e l l theory i n t h e version given here f o r s e t t i n g up t h i s system. It includes eighteen equations supplementing t h e f i r s t group. The first and second groups together contain t h i r t y - t h r e e equations. Obviously, under t h e method of c a l c u l a t i o n adopted here, t h e system of equations (21,la) - (21.5) and ( 2 2 . 2 ~ ~ ) (22.5) i s equivalent t o t h e system of equations (16.h) - (16,4b), supplemented by t h e r e l a t i o n s (15.12a) (15.12~). Therefore, w e need not d i s cuss here t h e general p r o p e r t i e s of t h e systems (21.la) - (21.5) and (22.2a) (22.5), since we intend t o do this i n our comparative a n a l y s i s of t h e equations of t h e c l a s s i c a l theory.

Section 23. Boundary and I n i t i a l Conditions FQuating t o zero t h e components under t h e sign of i n t e g r a t i o n over t h e w e obtain a system of n a t u r a l bouncontour of t h e b a s i c surface i n eq.(20.5), dary conditions. This system does not d i f f e r b a s i c a l l y from t h e conditions W e s h a l l , therefore, give here only t h e conditions on considered i n Seat.17. t h e free part of t h e contour surface that d i f f e r i n a n a l y t i c form from t h e conThe conditions on t h e attached part of t h e contour d i t i o n s (17.2a) (17.4b). surface remain unchanged.

Thus, on t h e free part of t h e contour surface, t h e following conditions

137

are satisfied:

The main stetements in W e s h a l l n o t analyze t h e conditions (23.1). Sect.17 n a t u r a l l y apply t o t h i s somewhat mbdified formulation o f t h e boundary problem. O f course, t h e i n i t i a l conditions considered i n Sect.18 a l s o apply t o t h e s o l u t i o n s of t h e systems of equations ( 2 1 . h ) (21.5), (22.2a) (22.4~).

Section 24. Generalized Conclusions and Further Development o f t h e Anal~ic Mechanics of S h e l l s 23, we may remark t h a t t h e discussed Analyzing t h e contents o f Sect.15 methods are based on s e l e c t i n g t h e generalized coordinates i n such a manner a s t o r e s t r i c t t h e number of degrees of freedom of t h e s h e l l i n t h e d i r e c t i o n of a normal t o t h e b a s i c surface. The reduction of t h e three-dimensional problem of t h e t h e o q o f e l a s t i c i t y t o t h e twodimensional problem goes back t o this same r e s t r i c t i o n . Similarly, t h e method of expansion i n t e n s o r series, a method /141 based on t h e use of v a r i o u s kinetic-geometrical hy-potheses ( f o r example, t h e hypotheses of s t r a i g h t i n v a r i a n t normals), in some form o r o t h e r limits t h e numb e r of degrees of freedom of t h e s h e l l in t h e d i r e c t i o n of a normal t o i t s b a s i c surface. Evaluating t h e s e methods o f s o l u t i o n of t h e reduction problem, we must recognize t h a t t h e most 1o.gically c o n s i s t e n t are t h e methods of reduction based These methods on an a p p l i c a t i o n of t h e general equation of dynamics (15.1). permit u s t o formulate a s y s t e m o f n a t u r a l boundary c o n d i t i o n s and t o f i n d i n i tial conditions t h a t do n o t e x p l i c i t l y c o n t r a d i c t t h e p r o p e r t i e s of t h e solut i o n s of t h e p r i n c i p a l system of equations. The method of expansion i n s e r i e s g i v e s even l e s s d i s t i n c t grounds f o r est a b l i s h i n g t h e system of boundary and i n i t i a l conditions. There a r e a l a r g e nuniber of methods of reducing t h e three-dimensional problems of t h e theory of e l a s t i c i t y t o t h e two-dimensional problems of t h e theory o f s h e l l s . All t h e s e methods a r e based on v a r i o u s s e l e c t i o n s of t h e system o f generalized coordinates. W e s h a l l here s t a t e two choices which, in our opinion, a r e of fundamental i n t e r e s t .

1 . Choice of G e n e r a l i z d Coo.@inates Corresponding to t h e

Optimum Q u a d r a t i c Approximations W e already c a l l e d t h e readervs a t t e n t i o n t o t h e c o r r e l a t i o n between t h e reduction problem and t h e methods of approximate r e p r e s e n t a t i o n of functions. The method of expansion in s e r i e s is one of t h e methods of optimum repres e n t a t i o n of in-point functions. In t h e case of t h e presence of a n a l y t i c sing u l a r i t i e s n e a r t h e approximation functions, however, t h e process of approximat i o n by segments of a Taylor s e r i e s may prove t o be divergent. This i s p a r t i c -

138

u l a r l y important f o r us, s i n c e t h e existence of concentrated f o r c e s a c t i n g on an e l a s t i c body cause t h e appearance of a n a l y t i c s i n g u l a r i t i e s n e a r t h e componFor t h i s reason, w e e n t s of t h e displacement v e c t o r and t h e stress tensor. will consider a choice of generalized coordinates l e a d i n g t o a determination of t h e c o e f f i c i e n t s of t h e expansion of t h e required f u n c t i o n s in Fourier s e r i e s over t h e segment (-h, + h ) o f a normal t o t h e b a s i c surface of t h e s h e l l , and i n d i c a t e t h e p o s s i b i l i t i e s f o r f u r t h e r development of this method. W e r e c a l l t h a t segments of a Fourier series accomplish optimum q u a d r a t i c approximation t o t h e function t o b e approximated*.

L e t t h e displacement v e c t o r . undergo p a r a l l e l displacement t o t h e b a s i c surxace along a normal t o t h a t surface. L e t u s denote t h e d i s p l a c e d v e c t o r /l42 From eqs.(I, 11.12) (I, ll.U+), and (I, 1 1 . 1 8 ) , we have by V.

vi = ZLi (1 - k,z); 213 = ua (i =.l, 2; do n o t sum over i 1 ) .

Let u s expand t h e components vi i n Fourier s e r i e s :

where

The r e p r e s e n t a t i o n of v1 by t h e s e r i e s (24.2) i s equivalent t o abandoning a determination of vi on t h e boundary surface of t h e s h e l l . I n f a c t , if t h e f h t h i s series s e r i e s (24.2) i s convergent, then, a t t h e boundarg s u r f a c e s z

will converge t o t h e value 1 (vi(+, + vi(-) ), where vi(+)a r e t h e values of t h e vect o r components

v'

on t h e boundary s u r f a c e s of t h e s h e l l .

Within t h e i n t e r v a l (-h, + h ) t h e s e r i e s w i l l converge t o t h e values of v i . W e t h e r e f o r e adopt t h e d e f i n i t i o n of v* by t h e series (24.2) as a s i m p l i f i c a will be t i o n whose meaning w i l l be given below. The c o e f f i c i e n t s vi(.) and d(mj F u r t h e r , we have considered a s generalized coordinates.

'1 zvi = -- 8vfo, 2

mxz ~vf,,cos - -h

G W ~ sin , ~ )-

m-1

(24.4)

Cf., f o r example, V.L.Gonchar:v, Functions, ONTI, 1934.

Theory of I n t e r p o l a t i o n and Approximation

Let us e s t a b l i s h t h e connection between t h e c o e f f i c i e n t vi(Bj, +(Bl and t h e e Fourier c o e f f i c i e n t s of t h e stress t e n s o r components. Using HookeTs law, w f i n d t h e following r e l a t i o n s :
5ik

hg'"irv'

+ ).g'"3vJ +p

(gi+7pk

Ti3 -= p
233'=

(g"vrv3 v3v'), 1Vrv' ( h 2p) P 3 V 3

+ +

+gk"p'),

(24.5a) (24.5b (24.5c

1 1

(i, k , r = 1 , 2; do not sum over i and k ! ) . The v a l i d i t y of these r e l a t i o n s f o r Euclidean space within a s h e l l i s obviou#.

To find t h e Fourier c o e f f i c i e n t s o f t h e stress t e n s o r components we

multiply eqsm(24.5a) - ( 2 4 . 5 ~ ) term by term by cosh h

mst/lk3

m z and sin-mTIZ and i n t e -

g r a t e over z from -h t o +h. L e t u s first consider t h e result of t h i s operation performed on t h e covariant d e r i v a t i v e V3vi, bearing of course i n mind t h a t a l l t h e covariant d e r i v a t i v e s i n eqs.(24.5a) ( 2 4 . 5 ~ ) a r e determined in t h e metric e have of t h e space adjoining t h e b a s i c surface. W

(i, j = 1 , 2; do not sum over i t ) .

+h

-h

wfmI3 =k

's
-h

+h

m= I v3ui sin mxz -dz = - -u(,) - Riw[,)

h h

(i = 1 , 2, 3; do not sum over i!),

* I n Euclidean

space, an a u x i l i a r g Cartesian system can always be introduced i r which t h e eqs.(24,5a) ( 2 4 . 5 ~ ) are d i r e c t l y confirmed. But t h e tensor equat i o n s a r e i n v a r i a n t [cf. ( I , Sect.6) and (Bibl.7)d.

In t h e s e equations f o r i 3, we m u s t set & = 0 . Here we have borne i n mind t h e above remark on t h e p r o p e r t i e s of t h e series (24.2) a t e = *h. When vi i s represented by t h e series (24.2), t h e d i f f e r e n c e vi+ vi- must be taken a s zero.

The covariant d i f f e r e n t i a t i o n Vi f o r i = 1 , 2 and i n t e g r a t i o n over z a r e commutative, since t h e d e r i v a t i v e s Vi are determined i n t h e metric of a space adjoining t h e b a s i c surface.

L e t t h e expansion in a Fourier series of t h e stress tensor components have t h e following form#

Then, f r o m eqso(24.5a) m d , w e find

( 2 4 . 5 ~ ) and bearing eqs.(24,6a)

(24.6b) in

/l&

Equations (24.8a) (24.lCb) form one o f t h e groups o f elastodynamic equat i o n s of t h e s h e l l theory, To s e t up t h e second group (equations of motion), we must again r e t u r n t o t h e general equation of dynamics. Consider f i r s t t h e v a r i a t i o n 6W(*) of t h e s p e c i f i c p o t e n t i a l energy of deformation. On t h e basis of eq.(15,10), we f i n d

* Here,

too, t h e above statement on t h e series (24.2) is valid.

1W

Since t h e f a c t o r (1 k l z ) (I k,z) which, according t o eq.(15.2), enters i n t o t h e expression f o r t h e element of volume dV, complicates t h e d i r e c t t r a n s formation of t h e general equation of dynamics ( 1 5 . 1 ) , l e t us put

Vi=(l -kk,z)(l -kk,z)v'. (i = 1 , 2, 3).

Hence, it follows t h a t

(24.12)

YjVi

1 ___ vivi; ( 1 - k , z ) ( I - k,z)

v3vi=

1 __ V3 V' (1 - k , z ) (1 - k , z )

+
(24.13 1

a + v i 82 - ( l - ~ ~ ~ ) - ~ ( l - k ~ z ) - l( i = l , 2 , 3 ;
and
- . \;;.cv' =

+1,2)

1 1 - v i w ; 60,v' = 5v3V'+ ( 1 - k , z ) (1 -k2z) (1 - k , z ) (1 -k,z) +;vi-a ( 1 -k,z)-I(l - k,z)-' ( i = 1, 2, 3; J = 1 , 2).

az

Let us r e p r e s e n t Vi and t h e v a r i a t i o n s SV, by Fourier s e r i e s :

O f course, t h e Fourier c o e f f i c i e n t s of t h e f u n c t i o n s Vi and vi as w e l l as 6 v and 6v' a r e i n t e r r e l a t e d . W e s h a l l n o t consider t h e s e r e l a t i o n s , b u t note that, t h e a r b i t r a r i n e s s and independence of t h e v a r i a t i o n s 6vi(,] and 6wilrl results i n t h e a r b i t r a r i n e s s and independence of t h e v a r i a t i o n s SIP(,) and SWi,,.
Let us expand t h e q u a n t i t i e s V3Vi i n a Fourier series. W e have

Consequently

Here, from eqs.(24,6a)

(24.6b), after an obvious change of notation, w e find

.
W e recall that & = 0, After obvious transformations we find, from the relations (24.U.) (24.18b):

-h

Here, w e have introduced the notation

143

-h

(i = 1. 2, 3).

The right-hand s i d e s of eqs.(2&.20a) (24.2Cb) can be represented by serFor s u f f i c i e n t l y t h i n s h e l l s , howand e*& i e s with terms expressed by ever, t h e r e i s no p o i n t i n complicating t h e statement of t h e problem by considering t h e s e r e l a t i o n s . Indeed, f o r t h i n s h e l l s t h e function
k , k, - 2R,k,z 1 - (k, k,) z k,k,z2

.+

f (x/

2) =

is monotonous f o r fixed values of xJ ( j = 1, 2 ) i f 2; v a r i e s over t h e i n t e r v a l e find (-h, +h), Then, applying t h e theorem on t h e i n t e g r a l mean, w

where C1 and c2 are c e r t a i n values of z on t h e i n t e r v a l (-h, +h), which in gena y approximately /147 eral depend on i and m. For s u f f i c i e n t l y t h i n s h e l l s , we m Put f ( x j , Ci) = IC, k,. (24.23)

In s p e c i a l cases, t h e function f(xJ t h e case of p l a t e s this function vanishes.

, z) i s

simplified.

For example, f o r

L e t u s continue t h e transformation of t h e q u a n t i t i e s e n t e r i n g i n t o t h e The v i r t u a l work of t h e f o r c e s of inertia is v a r i a t i o n a l equation (15.1). transformed as follows:

L e t u s introduce t h e following n o t a t i o n f o r t h e generalized f o r c e s on t h e b a s i c surface:

The generalized f o r c e s on t h e contour surface are expressed as follows:

-h

(24.26b )

where

Equations (24.25a)

(15.15a)

- (15.1%).

- (24.26~) are a n d o g o u s in meaning t o

expressions

W e s h a l l n o t consider t h e s u b s t i t u t i o n of t h e r e s u l t a n t expression into /148 t h e general equation of dynamics (15.1) nor t h e simple transformations connected with such s u b s t i t u t i o n , s i n c e they are analogous t o those given above i n Sect i o n s 15 23.

W e s h a l l now present t h e system of d i f f e r e n t i a l equations of motion and of t h e n a t u r a l boundary conditions r e s u l t i n g from t h e v a r i a t i o n a l equation (15.1) The equations of motion have with our s e l e c t i o n of generalized coordinates. t h e following form:

( m = 1 , 2 , . . . ; j = l I 2, 3; i = l ,

2;

do not sum over j l ) .

Here

e3 = 1; k3

= 0 .

The n a t u r a l boundarg conditions on t h e portion of the contour surface f r e e from kinematic c o n s t r a i n t s are:

- L(o) = 0; Tifm)jtzi - L(,, = 0 ; 0'. .Itl - M ( m ) j = 0 (m)J ( m = 1, 2,. .; j = 1, 2, 3; i-

~4,

(24,301 (24.31
1, 2).

(24.32)

The conditions on t h e clamped edge will not be w r i t t e n out. These a r e obvious, Since a l l equations derived here a r e of t h e second order w i t h respect t o time t, t h e system of i n i t i a l conditions does not d i f f e r from those known from Courses i n t h e p r i n c i p l e s of mechanics.
If we make use of eqs0(24.22a) (24.23), then t h e systems of equation (24.8a) (24.lCk) and (24.27) (24.29), taken together with t h e boundary con(24.32), permit us t o fonrmlate autonomous boundary condid i t i o n s (24.30) t i o n s t o determine t h e Fourier c o e f f i c i e n t s of t h e wanted quantities.

These problems are all o f t h e same type. For m # 0, each of t h e boundary problems l e a d s t o solution of t h e system of equations of t h e twelfth order with unknown Fourier c o e f f i c i e n t s o f t h e displacement vector components. The s o h For m = 0 t h e system of equations t i o n s must s a t i s f y six boundary conditions. w i l l be of t h e s i x t h order, and the number of boundary conditions w i l l be three. There a r e two additional remarks t o be made on t h e application of Fourier

series expansions t o solution of t h e problem of reduction.

1 . W e have selected t h e segment (-h, h ) a s t h e interval of expansion. It follows from t h e theory of Fourier s e r i e s t h a t t h i s i n t e r v a l can be extended i f we indicate t h e a n a l y t i c prolongations of t h e components of the displacement /149 vector beyond t h e segment (-h, h). Since t h e values of the displacement vector components vi beyond t h e segment (-h, h ) are a r b i t r a r y , they can be chosen such t h a t t h e expansions obtained over t h e extended i n t e r v a l s h a l l not contradict t h e conditions on t h e boundary surface of the s h e l l .
2, The above selection of t h e generalized coordinates can be so modified t h a t t h e operation of d i f f e r e n t i a t i o n with respect t o z of the proposed expansions of t h e vector components vi s h a l l not be e x p l i c i t l y nor i m p l i c i t l y en-

countered. For this, it i s s u f f i c i e n t t o s t a r t from t h e expansions of t h e components V 3 v 1 , and then t o use t h e system of l i n e a r d i f f e r e n t i a l equations of t h e first o r d e r f o r f i n d i n g t h e vi. In t h i s case, t h e generalized coordinates w i l l b e t h e q u a n t i t i e s vic,, and 3. A similar s e l e c t i o n of generalized coo r d i n a t e s will be discussed below in Subsect.2. 2. -@e of t h e N e w Versions of t h e Choice of Generalized Coordinates

I n most of t h e p r e s e n t Chapter we have been considering methods of reduct i o n based on t h e approximate r e p r e s e n t a t i o n of t h e displacement v e c t o r compone n t s as f u n c t i o n s of t h e coordinate e. To determine t h e stress t e n s o r compone n t s as f u n c t i o n s of z we had t o d i f f e r e n t i a t e t h e expressions of t h e d i s p l a c e ments Kith r e s p e c t t o e. Such a method of reduction cannot be c a l l e d opti". This shortcoming w a s t o some e x t e n t compensated by t h e use of Fourier expansions.
W e s h a l l now i n d i c a t e another p o s s i b l e approach t o e l i m i n a t i o n of t h i s drawback. L e t u s approximately r e p r e s e n t t h e covariant d e r i v a t i v e s V 3 u , by t h e equations :

where fl i s a function approximately representing t h e d e r i v a t i v e V 3 q , while 4 1 ', qb) a r e parameters considered t o be generalized coordinates. They are , 2 ) and of t h e time t. W f u n c t i o n s of t h e coordinates x J ( j 1 e recall that t h e q u a n t i t y 16 must b e put equal t o zero.

...,

I n t e g r a t i n g t h e d i f f e r e n t i a l equation (24.33),

we find

( i = 1, 2, 3).

N o w it i s no l o n g e r necessary t o d i f f e r e n t i a t e with r e s p e c t t o e i n calcul a t i n g t h e stress t e n s o r components. Let u s r e p r e s e n t t h e function f , by t h e polynomial

m
1

( i = 1, 2, 3).
(24.35

147

Then,

w8

obtain

[i= 1, 2, 3).

(24.36b )

- (I, 11.13)

Displacing t h e s e q u a n t i t i e s t o t h e b a s i c surface, we f i n d , f r o m and (I, l l . 2 0 ) ,

(I, ll.12)

(i = 1, 2, 3).

(24.37b

W e then set up t h e d i f f e r e n t i a l equations of motion and t h e boundary conW e l e a v e this task t o t h e d i t i o n s by means of t h e v a r i a t i o n a l equation (15.1). reader as an exercise. W e also c a l l a t t e n t i o n t o t h e coincidence of t h e r i g h t (24.37b) with t h e segments of t h e t e n s o r s e r i e s hand s i d e s of eqs.(24,37a) considered by u s a t t h e beginning of this Chapter. This again confirms t h e int e r r e l a t i o n between t h e o p e r a t i o n s of expansion of t e n s o r f u n c t i o n s in generali z e d Taylor series and t h e parallel displacement of t m s o r q u a n t i t i e s over f i n i t e distances.

Section 25. Application of Analytic Methods t o t h e Theory of O s c i l l a t i o n s of Layered S h e l l s Consider a s h e l l c o n s i s t i n g of p a r a l l e l i s o t r o p i c l a y e r s of constant nonThe b a s i c surface i s superposed on t h e boundary surface of identical thichess. t h e s h e l l having a u n i t v e c t o r of t h e n o m 1 (I, 3.1) d i r e c t e d inwards i n t h e material of t h e s h e l l . This choice permits u s t o obtain s e v e r a l p a r t i a l matheO f course, such a s e l e c t i o n of t h e b a s i c surface i s matical simplifications. n o t s u f f i c i e n t l y general, nor i s it opthum. O t h e r methods of choosing t h e e b a s i c surface are p o s s i b l e and have various advantages (Bibl.15, 21, 24). W shall not go i n t o t h i s question here.

In s e t t i n g up t h e equations of motion, t h e conditions of connectivity /I51 of t h e l a y e r s (11, 8.14) must be borne in mind. Of s p e c i a l importance here i s t h e choice of t h e generalized coordinates such as t o ensure maxi" s i m p l i c i t y t o t h e s o l u t i o n of t h e problem. W e n o t e t h a t t h e conditions (11, 8.14) impose, upon t h e generalized coordinates, r e s t r i c t i o n s t h a t do not depend on t h e l a w of motion, and must be considered as equations of constraint. O f t h e various above methods of introducing t h e generalized coordinates, l e t u s d i s c u s s t h e method

i n d i c a t e d i n Sect.24.2 and s u b j e c t this method t o a c e r t a i n extension. of eq. (24.33 ) , l e t us set

Instead

(25.1)

where 0, i s t h e Heaviside u n i t function, wr(k) are t h e excess generalized coordin a t e s introduced t o s a t i s f y t h e conditions of c o n n e c t i v i t y of t h e l a y e r s , z, a r e t h e z-coordinates of t h e s u r f a c e s o f contact, and m is t h e number of t h e s e surfaces

Making use of eqs.(24.34)

(24.36b),

W e find

up

+ z u y +1 z*up +2 6

2 3 ui 1311

It i s c l e a r from eq0(25.2a) t h a t t h e displacement v e c t o r components and the derivatives o, u, ( i = 1 , 2 ) a r e continuous f u n c t i o n s of z . The q u a n t i t i e s o,u, have f i n i t e d i s c o n t i n u i t i e s on t r a n s i t i o n a c r o s s t h e i n t e r f a c e of t h e l a y e r s . The magnitude of t h e d i s c o n t i n u i t y i s (1 k, 4 ) X #)(xi, t). We can f i n d t h e q u a n t i t i e s wit) from t h e condition of c o n t i n u i t y of t h e s t r e s s t e n s o r components u13, which we w i l l demonstrate:

z, ; l e t t h e La e l a s t i c c o n s t a n t s of t h e k t h l a y e r be lk and
(k +

L e t t h e 8-Goordinates on t h e boundarg s u r f a c e s of t h e k t h l a y e r be and ~ r , Then f o r t h e l a y e r we have

According t o Hookets law, t h e stress t e n s o r components ulg i n t h e k t h and

-.....I.

.-

.-

.. .. .. ....

..

_.

. .

(k

l a y e r s on t h e i r i n t e r f a c e are expressed by t h e following equations:=

The conditions of c o n t i n u i t y of t h e s t r e s s t e n s o r components lead t o t h e following values f o r t h e excess coordinates w/k)(.j = 1, 2, 3):

Equations (25.4a) (25.4b) permit u s t o express t h e excess coordinates w j ' " ) i n terms of t h e independent coordinates up). As an example of t h e a p p l i c a t i o n of t h e a n a l y t i c methods here presented, l e t u s consider t h e equations o f vibrat i o n of a two-layer s h e l l . Section 26. Equations of O s c i l l a t i o n o f t h e Two-Layer S h e l l Consider a s h e l l of constant thickness 2h, c o n s i s t i n g of two l a y e r s of res p e c t i v e thickness hl and Using eqse(25.2a) (25.2b) and performing a p a r a l l e l displacement t o t h e b a s i c surface, we f i n d , by analogy t o eqs0(24.37a) (24.3%):

(26 lb )

Eiquations (26.la) (26.lb) permit u s t o obtain t h e a n a l y t i c expressions for vi and V3vl in t h e first and t h e second layer. W e have

150

L e t us now eliminate from t h e s e equations t h e excess coordinates W e have i n g use of eqs.(25.&a) (25.4b).

wp),

mak-

Here we have introduced t h e n o t a t i o n

(26.5) Further, we f i n d

(26.6a)

I n s e t t i n g up t h e s e expressions we r e t a i n all terms containing t h e generalized coordinates L@ ( j I , 2 , 3; m = 0 , 1 , 2 , 3 ) r e g a r d l e s s of t h e convent i o n a l o r d e r of smallness in t h e s e terms i n connection with t h e presence of a f a c t o r of t h e form hf an.

L e t us now put eqs.(26,6a)

(26.6b) i n t o t h e following form:

where

Further,

152

(26 1ck )

(26.11)

( 2 6 . 1 2 )

Equations ( 2 6 . 8 ) , t o g e t h e r with eqs.(26.9a) (26.10) do not formally d i f f e r from eqs.(15.5), and t h e following r e l a t i o n obtained by expansion of t h e displacement v e c t o r components i s a t e n s o r s e r i e s . The d i f f e r e n c e i s t h a t t h e q u a n t i t i e s VJ0 and V , ' a r e piecewise-continuous f u n c t i o n s of z, which a r e cons t a n t on t h e segments of a normal t o t h e b a s i c surface enclosed within t h e layers.

L e t u s now f i n d t h e s t r e s s t e n s o r components, noting t h a t i n covariant d i f f e r e n t i a t i o n with r e s p e c t t o XJ ( j = 1 , 2) t h e q u a n t i t i e s z and u 0 ( z hl ) can be considered as constants. W e have

3
Ti8

z z
m-0

I m! Z m y )

(i, k = 1, 2, 3).

(26.13)

The right-hand s i d e of e q . ( 2 6 . 1 3 ) does n o t d i f f e r i n form from a segment of f . a r e piecewisea Taylor t e n s o r s e r i e s , b u t t h e c o e f f i c i e n t s l i m f k , except 7 continuous f u n c t i o n s of z, constant on t h e segments of a normal t o t h e b a s i c s u r f a c e enclosed between t h e boundary s u r f a c e s of t h e layer. The c o e f f i c i e n t s T (m)ik are expressed by t h e equations

',

153

=AgrrvrV$"

+ ( A +2B) YL"'")
m

( 26 14c )

(i, h, r -- 1, 2:

-0,

1, 2, 3; do not sum over i and k 1 ) .

Equations (26.14a) ( 2 6 . 1 4 ~ ) a r e analogous t o t h e r e l a t i o n s (15.12a) (l5.l2c), found f o r a single-layer homogeneous s h e l l . These equations, taken t o g e t h e r with eqs.(26.9a) (26.1clb) and (26.12), permit us t o express t h e stress t e n s o r components i n t h e first and second l a y e r s in terms of t h e genera l i z e d coordinates u p ) and t h e i r d e r i v a t i v e s .

Consider t h e v a r i a t i o n s 6v, :
3

za, =
m-0

m!

z"6 V!"

( i = 1, 2, 3).

From eq. (26.12) t h e r e r e s u l t s , by analogy,

/156

The v a r i a t i o n s 6Vjo) and 6V,(1) are piecewise-continuous functions of I , cons t a n t on t h e segments of t h e normal t o t h e b a s i c surface included within t h e P ) as new generalized layers. This does n o t permit t h e use of t h e q u a n t i t i e s V n t h e b a s i s of eqs.(26.9a) (26.9b) and (26.1Oa) (26.lCb), WE coordinates. O have

(26.16a ) (26 16b )

where

I 1 - 1

I II

(26.17a)

The presence of covariant d e r i v a t i v e s on t h e right-hand s i d e s of equations (26.17a) - (26.1%) leads t o fundamental d i f f i c u l t i e s , as will be seen from what follows.

2h

p m i i k --

111 !

JZ , +
0

(1 -- k,z) (1 - k,z) dz

( 26 18a)

(i, k - 1, 2, 3);

(26.18b)

( 26 18c )

/157
Ktm)I3m!
L1*Zm?.i:l

(1 - k,z) (1 - k , z ) dz

l 0 i ( i = l , 2, 3; i , k = l ,

(26.18d)

2).

Further, l e t u s consider t h e q u a n t i t i e s connected with t h e i n e r t i a l forces:

(26.19a )
2h

(26.19b )

155

2h
a2v3

m!
0

dt2

(1 -k,z)(1 - R,z) dz

(26 19d)

(26.19d) are expressed i n The q u a n t i t i e s determined by eqsm(26,18a) terms of t h e generalized coordinates u p ) , t h e generalized a c c e l e r a t i o n s u/d and t h e i r d e r i v a t i v e s w i t h respect t o t h e coordinates xJ of t h e b a s i c surface of t h e s h e l l by means of formulas (26.8), (26.14a) ( 2 6 . 1 5 ) .

L e t u s now introduce t h e generalized forces.

W e put

2h
Q(m)i =

- S p P P (1 - k , z ) (1 - k , z ) d z + X f + , (1 - 2k,h) m!
0

x
(2 6 20a )

x ( 1 - 2k,hj (2h)m;

p!miR = nt! If y d P z m
0

(1 - k , z ) (1 - k , z ) dz

-t

x;+) x
XL) x

x (1 -- 2/<,h) ( 1 - 2k&) (2/z)*;

(26 21b )

2h

p i 3

- - J - I I 2 p z -( I -- k,z) (1 - k,z) dz
JIL !
0

+[&I

x (1 - 2 k l h )(1 - 2kJz) (2iz)m

( J ' = 1, 2; m =0, 1).

(26.21~)

where [p12],

[yr2] and [A,,]

a r e t h e values o f p 1 2 , y12 ,and

)clz

at z

2h.

On t h e contour surface we determine t h e following q u a n t i t i e s :

2h

L " "

1 =m!
0

pl2Xi(p(2) dz,

(26.22b)

L(m)3

2! TTl2Fqdz,
Nl

(2)

(26 22c )

0
2h

( i = 1, 2, 3; j = 1, 2).

where q ( z ) i s expressed by t h e formula (24.26~). Let u s now consider again t h e general equation of dynamics ( 1 5 . 1 ) , making use of eq.(24.11) i n i t s transformation. in The transformation of eq.(15.1), t h i s case, has two stages. The first stage l e a d s t o t h e following result:

157

(26.23) where
3 .

(26 24 )

The presence, on t h e left-hand s i d e of t h e v a r i a t i o n a l equation (26.23), of terms w i t h t h e f a c t o r s g r v , (6vr , (6~3 i n d i c a t e s substanand O t i a l d i f f e r e n c e s between t h e cases considered e a r l i e r and this problem. These f a c t o r s can be eliminated from t h e surface i n t e g r a l by means of i n t e g r a t i o n by p a r t s , i.e., by repeating t h e transformation of e q . ( 1 5 . 1 ) , l e a d i n g t o equation T h i s i s t h e second s t a g e of t h e transformations mentioned above. The (26.23). use of term-by-term i n t e g r a t i o n does not r e q u i r e t h a t t h e conditions of d i f f e r e n t i a b i l i t y of the s u r f a c e load components be s a t i s f i e d , s i n c e t h i s load i s eliminated from t h e sums -h Pc0) + P(') and -hl R ( O ) + R(l) The f a c t o r s gr'Vr' ( 6vr )I = h, and Ok ( 6v3 z = hi will, however, l i k e w i s e e n t e r under t h e sign of i n t e g r a t i o n over t h e contour C of t h e b a s i c surface. Here, t h e s e terms cann o t be excluded. Consequently, t h e r e i s a s u b s t a n t i a l a d d i t i o n t o t h e n a t u r a l boundary conditions.

)Izxh,

) I,=,,

)I

W e w i l l d i s c u s s this question l a t e r i n t h e text, b u t f i r s t l e t u s s e t up t h e system of d i f f e r e n t i a l equations o f motion of t h e two-layer s h e l l which results from eq.(26.23) a f t e r t h e second s t a g e of transformations. S e c t i o n 27. D i f f e r e n t i a l Equations of Motion .o f a Two-Layer S h e l l W e introduce t h e notation:

/160

158
. .. .. ..

Performing the above transformation on the variational equation ( 2 6 . 2 3 ) , we f i n d by the usual method the following system of differential equations of motion:

159

/161
(5, k = 1 , 2; do n o t sum over k t )

(270%)

The s y s t e m of equations (27.2a) (27.5b) c o n s i s t s of twelve d i f f e r e n t i a l The d e r i v a t i v e s equations of t h e t h i r d o r d e r in twelve unknown f u n c t i o n s u p ) The general o r d e r of t h e entering i n t o 0 ,uk and v k Z are of t h e highest order. system i s 36. Obviously, t h e o r d e r of t h e system i s so high because a l l t h e terms containing t h e generalized coordinates introduced by u s were r e t a i n e d i n t h e equations, r e g a r d l e s s of t h e i r conventional value, deterxnined by t h e expone n t m i n t h e f a c t o r s H. O f course, such a system of equations i s u n s u i t a b l e f o r p r a c t i c a l calculat i o n s , b u t it my s t i l l be used f o r purposes of comparison w i t h o t h e r systems obtained by v a r i o u s s i m p l i f i c a t i o n s . A q u a l i t a t i v e a n a l y s i s of eqs. (27.2a) (27.5b) may l i k e w i s e introduce new elements i n t o t h e r e p r e s e n t a t i o n of t h e osc i l l a t o r y processes in a layered s h e l l .

Section 28. Natural Boundary Condition

Two b a s i c forms o f boundary conditions result from t h e v a r i a t i o n a l equat i o n (26.23).

1 . With Contour Surface Kinematically not Free O n t h e unfree contour s u r f a c e t h e q u a n t i t i e s uJ must be assigned. This assignment determines t h e i r d e r i v a t i v e s w i t h r e s p e c t t o t h e a r c of t h e conFrom t h e composition of t h e integrand expression i n t h e i n t e g r a l over t o u r C. it i s c l e a r t h a t on t h e unfree part of a contour t h e contour C i n eq.(26.23), s u r f a c e we must a l s o p r e s c f i b e t h e d e r i v a t i v e , with r e s p e c t t o t h e normal t o t h e contour, of v , and of t h e sum v,. By this assignment, t o g e t h e r w i t h t h e e n t of t h e f u n c t i o n s u[;) t h e covariant d e r i v a t i v e s V,V, J Z 5 , , , and The t o t a l number of boundary condivkv3 Z.h,, a r e determined on t h e contour C. t i o n s here i s n o t single-valued.

assir

I n f a c t , we may p r e s c r i b e a l l t h e d e r i v a t i v e s with respect t o t h e normal t o t h e contour C of t h e q u a n t i t i e s uJ I n t h a t case, we w i l l o b t a i n twentyf o u r boundary conditions. This number of boundary conditions does not c o r m - m spond t o t h e order of t h e system (27.2a) (27.5b).

If we d i r e c t l y p r e s c r i b e t h e normal d e r i v a t i v e s of g r v ; and v3 h,, then we w i l l have f o u r t e e n boundary conditions. The uniqueness of t h e s o l u t i o n of t h e boundary problem i n t h i s case r e q u i r e s f u r t h e r i n v e s t i g a t i o n .

Iz.h,

l2.

2. With t h e Contour Surface Free

If t h e boundary s u r f a c e s of t h e s h e l l a r e free from kinematic c o n s t r a i n t , then t h e v a r i a t i o n a l f i e l d ofbuj) i s e n t i r e l y a r b i t r a r y within t h e region w and may a l s o be reprealong i t s boundary. An a r b i t r a r y v a r i a t i o n a l f i e l d of 6uJ@)
160

This shows t h a t on t h e boundary sented beyond t h e boundary of t h e region W. t h e v a r i a t i o n gr6vrV , I I must b e regarded as an a r b i t r a r y , independent qnan-

tits.

I n a s s i g n i n g t h e v a r i a t i o n s 6uj(') on t h e contour surface, t h e d e r i v a t i v e s 6vkv3 w i l l b e bound by a linear r e l a t i o n r e s u l t i n g from t h e expression o f t h e . For this reason, f o u r t e e n d e r i v a t i v e s of 6v, over t h e a r c of t h e contour C boundary conditions can b e obtained from eq. (26.23) :

(28.34

They are/163 The c o e f f i c i e n t s 4, are f u n c t i o n s of t h e a r c of t h e contour C. determined a f t e r e l i m i n a t i o n of one of t h e d e r i v a t i v e s o k 6 ~ I3 z = h by means of t h e expression of t h e a b s o l u t e d e r i v a t i v e of 6 ~ 13 z - h over t h e a r c of t h e con. The q u a n t i t i e s N ( ' ) 3 are derived from t h e term-by-term i n t e g r a t i o n of tour s i n eq.(26.23), containing t h e a b s o l u t e t h e " n a n d under t h e i n t e g r a t i o n s i g n
(Cl

d e r i v a t i v e of 6v3 I z = h l over t h e a r c S. This summand a l s o r e s u l t s from eliminae w i l l not 6v3 z= h,, as already mentioned. W t i o n of one of t h e d e r i v a t i v e s give t h e d e t a i l s on t h e s e c a l c u l a t i o n s , b u t r a t h e r pass t o b r i e f conclusions g e n e r a l i z i n g t h e v a r i o u s methods of reduction of t h e three-dimensional problems

161

o f t h e e l a s t i c i t y theory t o two-dimensional problems. Section 29. C l a s s i c a l Theory o f S h e l l s

The i n v e s t i g a t i o n made by u s n a t u r a l l y i n c l u d e s a b r i e f a n a l y s i s of t h e c l a s s i c a l theory of s h e l l s and of c e r t a i n works t h a t have expanded t h e f i e l d of e will d i s c u s s t h e c h a r a c t e r i z a t i o n of t h e a n a l y t i c p r o p e r t i e s t h i s theory. W of t h e fundamental q u a n t i t i e s which a r e t h e o b j e c t of i n v e s t i g a t i o n i n t h e c l a s s i c a l theory, and on t h e methods of i n v e s t i g a t i o n .

1 . Forces and Moments

I n t h e Kirchhoff-Love theory of s h e l l s , t h e s t r e s s t e n s o r i s replaced by a system of f o r c e s and moments determining t h e p r i n c i p a l v e c t o r and t h e p r i n c i p a l moment of t h e i n t e r n a l f o r c e s i n t h e s h e l l , reduced t o a p o i n t l y i n g on t h e cont o u r of an element of t h e middle surface and r e f e r r e d t o u n i t l e n g t h of t h e corresponding coordinate l i n e .

It i s easy t o convince ourselves t h a t t h e components of t h e f o r c e s and moments so determined are n o t components of v e c t o r s obeying t h e r u l e s of transformation of t e n s o r q u a n t i t i e s . The f o r c e s and moments can e v i d e n t l y be connected w i t h t h e v e c t o r components of t h e stress t e n s o r (Bibl.8). W e s h a l l not go i n t o d e t a i l s on this approach,
A f t e r c o n s t r u c t i n g t h e v e c t o r s of t h e s t r e s s e s a c t i n g on t h e contour surf a c e s of an element of t h e s h e l l , w e w i l l s u b j e c t them t o p a r a l l e l displacement t o t h e b a s i c surface, on t h e b a s i s of ( I , 1 1 . 1 8 ) , and will then r e l a t e them t o u n i t l e n g t h of t h e corresponding coordinate l i n e . W e obtain
+h
n

Here t h index (k) i n d i a t e s t h e number of t h e coordinate l i n e normal t o that(164 p a r t of t h e contour s u r f a c e of t h e s h e l l element on which t h e s t r e s s v e c t o r a c t s , y i e l d i n g t h e q u a n t i t i e s Ti(k) and p ( k . , The o t h e r n o t a t i o n i s conventional. Further, making use of (I, 8.6), which d e f i n e s t h e covariant components of a v e c t o r product, we f i n d t h e components of t h e moments of t h e annexed couples, produced when t h e stress v e c t o r s a r e reduced t o t h e b a s i c surface, r e f e r r i n g them t o u n i t l e n g t h of t h e Corresponding coordinate l i n e . Neglecting all terms t h a t a r e nonlinear w i t h r e s p e c t t o t h e s t r a i n t e n s o r components, we f i n d
+,h z

M ~ ~ ~ , = - g l l l , J ~ o 2J (1 - z k , ) ( I --zk2)2dx3dz=
-h 0

162

1 --Zzk2) dz;

m u a t i o n s (29.la) (29.2f) d i f f e r i n form from those familiar from t h e c l a s s i c a l theory of shelld6. But t h i s d i f f e r e n c e i s n o t s u b s t a n t i a l , s i n c e most works on s h e l l theory replace t h e components o f t h e t e n s o r q u a n t i t i e s by t h e i r "physical components", which may b e obtained by using eqs.(I, 5.20) (I, W e have, f o r example, 5.2) and noting t h e remark i n (I, Sect.7).

and a l s o

S u b s t i t u t i n g t h e expressions ( a ) and ( b ) i n t o one of eqsD(29.1a), we find

T,lj.xl= -h

s*

a.slxl (1

- zk,) dz.

The r e l a t i o n ( c ) i s known from t h e c l a s s i c a l theory.

B y analogy, p u t t i n g

36

C f . Arthur Love, The Mathematical Theory of E l a s t i c i t y , ONTI, 1935, o r (Bibl.D5,11)D

163

we f i n d . f r o m eq.(29.&):

.
+h

M(l)Sl= -h

zaxlx(1 - zkJ

dz.

This i s t h e well-known expression f o r t h e t o r s i o n a l moment. Thus, t h e method employed here, based on t h e theory of parallel displacement of t e n s o r q u a n t i t i e s , l e a d s t o t h e results of t h e c l a s s i c a l t h e o r y i f we remain within t h e limits o f t h e l i n e a r theory of e l a s t i c i t y . I f we r e t a i n t h e nonlinear terms, however, this method l e a d s i n s t e a d t o r e s u l t s t h a t r e f i n e t h e c l a s s i c a l theory. as shown by u s elsewhere (Bib1.23a, b). W e s h a l l n o t consider t h e n o n l i n e a r theory here

W e w i l l now d i s c u s s t h e connection between t h e f o r c e s , moments, and t h e - ( 2 0 . 4 ~ ) . The q u a n t i t i e s T ( O ) i J (i, j = 1 , 2 ) a r e connected by l i n e a r r e l a t i o n s with t h e f o r c e s and moments. I n f a c t , eqo(2O.4a) y i e l d s

quantities T ( m )l k as determined by eqs. (20.4a)

Ti0)'j =

J dj(1- k i Z ) (1 - k,z) (1
ck

+h

--

k,z) dz -

-h

From t h i s t h e above-mentioned r e l a t i o n i s obtained.

The q c a n t i t i e s T ( l ) i J cannot be expressed i n terms of t h e f o r c e s and moW e note t h a t ments. This a l s o applied t o t h e q u a n t i t i e s T(m) i 3 and T(') 33 only i n t h e approximate theory which contains e r r o r s t h a t permit n e g l e c t i n g a l l tems o f t h e o r d e r of h k , are t h e q u a n t i t i e s T ( O ) 1 J and T ( O ) proportional t o the f o r c e s , and t h e q u a n t i t i e s T ( l ) i J proportional t o t h e moments.

A l l t h e above again l e a d s t o t h e conclusion t h a t t h e approximate replacement of t h e stress t e n s o r by a system of f o r c e s and moments i s j u s t i f i e d only i n the case of." a preliminary i n t r o d u c t i o n o f simplifying hypotheses analogous t o those of Kirchhoff-Love. For t h i s reason, v a r i o u s g e n e r a l i z a t i o n s of t h e c l a s s i c a l theory involving t h e use of reduced f o r c e s and moments a r e o f l i m - /166 it& value. In p a r t i c u l a r , t h e nonlinear theory of p l a t e s and s h e l l s , cons t r u c t e d w i t h t h e u s e of reduced f o r c e s and moments, contains e r r o r s t h a t decrease t h e s i g n i f i c a n c e of t h e i n t r o d u c t i o n o f t h e n o n l i n e a r terms.

2 . Equations of Equilibrium and Motion

The shortcomings connected with t h e d e s c r i p t i o n of t h e s t r e s s e d s t a t e of a s h e l l element by a system of f o r c e s and moments are e s p e c i a l l y pronounced when we consider t h e equations of motion o r , i n p a r t i c u l a r , t h e equations o f equilibrium of t h i s element. W e r e c a l l t h a t an element of a s h e l l has a f i n i t e dimension i n t h e d i r e c t i o n o f t h e l i n e 9 . The equations of motion of an e l e ment of t h e s h e l l i n t h e c l a s s i c a l theory are set up a s t h e equations of a r i g i d body. They r e s u l t from t h e theorem on t h e motion of t h e c e n t e r of inert i a of a s h e l l element and t h e theorem on t h e v a r i a t i o n of i t s k i n e t i c moment. C l e a r l y such an approach t o s e t t i n g up t h e equations of motion i s based on t h e preliminary a p p l i c a t i o n of one of two methods of reduction of t h e threedimensional. problems of t h e theory of e l a s t i c i t y t o two-dimensional problems; t h i s i s t h e method based on t h e a p p l i c a t i o n of t h e Kirchhoff-Love hypotheses,or expansions i n t e n s o r series followed by elimfnatAon of t h e d e r i v a t i v e s . . , V E s s e n t i a l l y , t h e use o f t h e equations of motion of an element "as a V3u, whole", by d e f i n i n g t h e g e n e r a l statement o f .iic reduction problem, permits exc l u s i o n of eqs.(7.4b) and (7.4d) from considzration. W e have mentioned this f a c t i n Sect.7 and i n t h e subsequent discussion.

V- s h a l l n o t here consider a l l t h e c l a s s i c a l equations of motion of a she;.'. element, b u t r a t h e r focus our a t t e n t i o n on t h e s i x t h equation, containing t'.a component of t h e moment of e x t e r n a l forces P ? , r e f e r r e d t o u n i t a r e a of t h e b a s i c surface o f t h e s h e l l .

Assume t h a t t h e deformed s h e l l i s *.--?;'a~*ir.ed t o t h e system of coordinates xi The coordinate l i n e s 1?. and x? on t h e dafomed b a s i c s u r f a c e coincide with i t s l i n e s of curvature, while t h e v e c t o r e 3 nf t h e coordinate b a s i s i s d i r e c t e d along t h e normal t o it, and i s equal, modulo, t o unity. Then, we may make use o f eqs.(29.la) (29.2f), b u t we must remmber t h a t a l l t h e q u a n t i t i e s e n t e r i n g i n t o them r e l a t e t o t h e deformed s h e l l .

Following our o t h e r work (Bib1.23a, b), we shall show t h a t t h e s i x t h equat i o n of equilibrium i s n o t s a t i s f i e d i f the component $ of t h e p r i n c i p a l moment of e x t e r n a l f o r c e s does n o t vanish. In connection with t h e vanishing of M(1).3 in t h e s e formulas ( 2 9 . 2 ~ ) and (29.2f), t h e s i x t h equation of e q w l i b s y s t e m of coordinate s e l e c t e d by u s has t h e following form ( s e e B i b l 23a, b ) :

Making use of eqs.(29.la)

(29.2f),

we f i n d

J . -h,

I n t h e s e equations, hl and h, a r e t h e d i s t a n c e s along t h e normal from t h e de-

formed b a s i c surface t o t h e boundary surfaces.

Further, l e t u s use eqs. (4.4):

S u b s t i t u t i n g i n t o eq.(29.3)

t h e r e l a t i o n (g)

- (.e>,

we f i n d

Consequently, eq. (29.3 ) reduces t o t h e condition

~3

= 0.

(29.4)

Thus, t h e s y s t e m of f o r c e s and moments (29.la) (29.2f) reduced t o t h e b a s i c surface, cannot balance t h e e x t e r n a l forces, i f they a r e reduced t o a

/168

166

couple l y i n g i n a plane p a r a l l e l ' t o t h e plane of a tangent t o t h e b a s i c surface'k.

W e s h a l l now make two remarks on t h e r e s u l t .

1 .W e have used a s p e c i a l s e l e c t i o n of t h e coordinate system connected w i t h t h e deformed s h e l l . However, t h e i n v a r i a n t p r o p e r t i e s of t e n s o r equations of equilibrium permit u s t o a s s e r t t h a t t h e result i s v a l i d in any system (I, Sect. 6 )

O f course, i f t h e choice of t h e coordinate system i s a r b i t r a r y , we w i l l But in t h i s case, o n l y two of t h e t h r e e equations of n o t o b t a i n eqs.(29.4). equilibrium containing moments of i n t e r n a l f o r c e s w i l l be independent.
I

2. I f t h e component $ vanishes, then t h e s i x t h equation of equilibrium i s s a t i s f i e d i d e n t i c a l l y only on t h e b a s i s of t h e expressions of f o r c e s and moThe i d e n t i c a l s a t i s f a c t i o n of t h i s equation i s enments (29.la) (29.2-f). t i r e l y unconnected with t h e r e l a t i o n s between f o r c e s , moments, components of t h e s t r a i n t e n s o r of t h e b a s i c surface, and t h e t e n s o r of v a r i a t i o n of i t s curv a t u r e (Section 10) r e s u l t i n g from Hookets law. For t h i s reason, t h e i d e n t i c a l s a t i s f a c t i o n of t h e s i x t h equation of equilibrium by r e l a t i o n s r e s u l t i n g from equations approximately expressing Hookers law must be considered only a s an i n d i c a t i o n t h a t t h e approximation adopted can i n f a c t be s a t i s f i e d .

Returning t o t h e i n t r o d u c t o r y remarks on t h e equations of equilibrium and t h e motion of t h e s h e l l , we note t h a t t h e condition (29.4) imposed on t h e ext e r n a l f o r c e s r e v e a l s t h e i n s u f f i c i e n c y of t h e d e s c r i p t i o n of t h e s t r e s s e d s t a t e of t h e s h e l l by a system of f o r c e s and moments reduced t o t h e b a s i c surface. This i n s u f f i c i e n c y . has no e f f e c t on t h e s o l u t i o n s of most t e c h n i c a l problems of t h e s h e l l theory, s i n c e i n t h e s e problems t h e condition (29.4) i s u s u a l l y satisfied. Section 30. B r i e f Survey of Recent R e s u l t s of Reducing t h e Three-Dimensional Problem of t h e Theory of a a s t i c i t y t o t h e Two-Dimensional Problem of t h e Theory of S h e l l s I n conclusion, l e t u s give a c h a r a c t e r i z a t i o n of t h e r e s u l t s obtained i n solving t h e reduction problem during t h e l a s t q u a r t e r century. Here, we w i l l n o t analyze t h e outstanding work by F.Krauss w r i t t e n i n l 9 2 F b u t merely remark t h a t he posed t h e problem of c o n s t r u c t i n g a s t a t i c s f o r s h e l l s t h a t d i d n o t r e l y on t h e Kirchhoff-Love hypotheses.

This was f i r s t e s t a b l i s h e d by a d i f f e r e n t method by F.Krauss in h i s paper Fundamental Equations of Shell Theory, Math. Ann., Vol.101, 1929. This proof w a s mentioned by u s elsewhere (Bib1.23a) in 1938. See a l s o t h e monograph by V .Z .Vlasov (B i b l 3a)

. .

Y & See preceding footnote. found elsewhere (Bib1.23b).

A b r i e f a n a l y s i s of Krausst i n v e s t i g a t i o n s w i l l b e

167

1 . Reduction by t h e Use of S e r i e s . Application of t h e

DfAlembert-Lagrange P r i n c i p l e

In 1938 1940 (Bib1.23a, b ) one v e r s i o n of t h e a n a l y t i c s t a t i c s of s h e l l s was studied, based on t h e use of expansions of stress and strain t e n s o r compone n t s in YacLaurin t e n s o r series i n powers of t h e coordinate x? = z . T h i s method, w i t h t h e necessary general equations, i s given a t t h e beginning of this
Chapter f o r dynamic problems. The reduction method based on t h e expansion of t h e wanted q u a n t i t y i n power s e r i e s of z was applied t o problems of t h e s t a t i c s of s h e l l s and p l a t e s by A.I.Lurtye i n 1940 1942 (Bibl..25a, b). I n one paper (Bibl.2%), he studi e d t h e equilibrium of a plane F l a t s and showed t h a t t h e displacement of any p o i n t of t h e p l a t e could b e expressed i n terms of c e r t a i n functions determined by t h e loads on t h e f a c e s of t h e p l a t e ( f o r z = f h) i n t h e form of s e r i e s f o r which t h e form of t h e nth term was established. This same method was employed i n a monograph (Bibl.9b) i n studying t h e equilibrium of a plane Layer. The results w e r e obtained by t h e symbolic method. The work of A.I,Lurfye confirms t h e s i g n i f i c a n c e o f t h e reduction method based on an expansion i n power s e r i e s of 2.

W e f i n d t h e i d e a of t h e combined use of t h e general equation of s t a t i c s and expansion i n power s e r i e s of a of t h e s t r e s s and s t r a i n t e n s o r components, w i t h t h e o b j e c t o f solving t h e reduction problem (Bibl.3aP. Here t h e hypothe s e s of Kirchhoff-Love a r e used, and t h e s h e l l element i s regarded as an absol u t e l y r i g i d body w i t h six degrees of freedom. Clearly, under t h e s e assumpt i o n s , t h e a p p l i c a t i o n of t h e general equation of s t a t i c s introduces no substant i a l l y new elements i n t o t h e s o l u t i o n of t h e problem of reducing t h e threedimensional problem of e l a s t i c i t y theory t o a two-dimensional problem, and a s a r e s u l t we o b t a i n t h e equations o f t h e c l a s s i c a l s t a t i c s of s h e l l s . This method was f u r t h e r developed by Kh.M.Mushtary and 1.G.Tereplov (Bib1.27), who studied t h e reduction problem f o r t h e s t a t i c problem i n nonline a r formulation, u s i n g expansions of t h e displacement v e c t o r components ii1 series i n powers of t h e v a r i a b l e 2 = Z . W e used a s i m i l a r device i n t h e l i n e a r 23, when we i n v e s t i g a t e d t h e problems of elastodyf o r m l a t i o n in Sect.15 namics.
4

The general equation of dynamics (15.1) holds l a t e n t p o s s i b i l i t i e s f o r t h e development of a reduction theory. Certain a p p l i c a t i o n s of t h i s equation t o t h e new formulations of t h e dynamic boundary problems o f s h e l l theory have a l r e a d y been i n d i c a t e d by u s i n SectO2EW*. Of course, even t h e s e r e s u l t s do not ex- / l 7 O haust dl1 t h e f a c t s obtainable from eq.(15.1). The general equation of dynamics was e v i d e n t l y applied by A.Basset t o cons t r u c t i o n of t h e equations of t h e c l a s s i c a l theory of s h e l l s , f o r c y l i n d r i c a l and s p h e r i c a l s h e l l s , as far back as 1890. See A.Love, Mathematical Theory of E l a s t i c i t y , ONTI, 1935, pp. 559-561. The reader will f i n d s e v e r a l d a t a on t h e development of i n v e s t i g a t i o n s on t h e reduction problem i n another paper (Bib1.26).

168

Methods of reduction based on t h e use of s e r i e s expansions have been developed a s e a r l y as 1942 by Epstein, Kennard, and o t h e r s who studied t h e dynamics of c y l i n d r i c a l s h e l l s and were apparently unacquainted with t h e work of Soviet scientist*. Beginning from about 1948, these generalizations, mainly of t h e dynamics of p l a t e s and s h e l l s , became widespread everywhere. The cause of this new i n t e r e s t i n s t u d i e s which c e r t a i n s c i e n t i s t s formerly c l a s s i f i e d a s t h e o r e t i c a l i n v e s t i g a t i o n s without p r a c t i c a l value, was the need t o e s t a b l i s h a dynamics of p l a t e s and s h e l l s s u i t a b l e f o r t h e study of various high-frequency v i b r a t i o n s and t r a n s i e n t s of dynamic loading.

It must be mentioned again t h a t p r i o r i t y i n t h e development of t h e genera l i z a t i o n s of t h e theory of p l a t e s and s h e l l s belongs t o Ukrainian and Soviet s c i e n t i s t H-

2. The "Semi-Inverse"

Method of Reduction

During t h e l a s t decade, a new trend has developed i n the methods of reduction of t h e three-dimensional problems of the theory of e l a s t i c i t y t o the two-dimensional problems of the mechanics of p l a t e s and shells. These methods may be called vtsemi-inversevy, since t h e i r d i s t i n c t i v e f e a t u r e i s t h e prelimina r y determination of c e r t a i n components of t h e s t r e s s o r s t r a i n tensor by cert a i n functions of the coordinate 2 = z.
I n chronological order, i n this respect, w e must l i s t the work of E.Reissner on the theory of equilibrium of t h i n plate-. A n analysis of Reissnerts work and i t s possible generalization i s given i n another paper (Bibl.20a). /171 Reissner expressed components o f t h a t p a r t of t h e stress tensor tangential t o t h e middle plane by l i n e a r functions of t h e coordinate z, and determined the components of t h e normal p a r t from t h e equations o f equilibrium, finding the indeterminate elements of t h e solution from t h e conditions on the boundary surface-. H e thus obtained a solution s a t i s f y i n g the conditions of e q u i l i -

a ) P.S.Epstein, O n t h e Theory of E l a s t i c Vibrations i n P l a t e s and Shells. J. Math. and Phys., Vol. 21, 1942 b ) E.H.Kennard, The N e w Approach t o S h e l l Theory: Circular Cylinders.IAM, V01.20, No.1, 1953; Cylindrical Shells: Energy, Equilibrium, Addenda andErratum. IAM, V01.22, No.1, 1955; Approximate Ehergy and Equilibrium Q u a t i o n s f o r CylA M , Vo1.23, No.4, 1956; A Fresh T e s t of the Epstein Equations i n d r i c a l Shells. I A M , Vo1.25, No.4, 1958. See a l s o U S S R Abstract Journal of Mef o r Cylinders. I chanics, N o . 2 , Abstract No.802, 1953 The generalized equation f o r transverse vibrations o f rods w i t h allowance f o r t h e e f f e c t of shear and i n e r t i a of r o t a t i o n w a s found by S,P,Timoshenko i n These r e s u l t s were extended t o the theory of vibrations of p l a t e s l921-1322. by Ya.S.Uflyand i n his paper '?Propagation of Waves i n Transverse Vibrations of Rods and Plates", PMM, Vol.XI1, No.3, 1948 +sw E.Reissner, a ) On t h e Theory of Bending of ELastic Plates. J . Math. And Phys., Vol.XXII1, 1944. b ) On Bending of m a s t i c Plates. Quar. Appl. Math.,

V01.5,

N o . 1 ,

The semi-inverse method of constructing t h e stress f i e l d f o r a s h e l l of a r b i t r a r g configuration was given by A.Love. Cf. A.Love, Mathematical Theory of E l a s t i c i t y . ONTI, 1935, p.560

1947

169

b r i m of t h e theory of e l a s t i c i t y , and boundary c o n d i t i o n s of s p e c i a l form on t h e boundarg s u r f a c e s applying t h e C a s t i g l i a n o p r i n c i p l e , s a t i s f i e d i n t e g r a l l y Saint-Venantqs c o m p a t i b i l i t y conditions, and derived t h e n a t u r a l boundary cond i t i o n s on t h e contour surface. Thus, here t h e semi-inverse method l e d t o a r a t h e r complete and convincing a n a l y s i s of t h e question. The semi-inverse method of solving t h e reduction problem i s a l s o found i n t h e work of S.A.Ambartsurnyan (Bibl.l6a-c), and A.A.Khachatryan (Bib1.33). I n t h e s e s t u d i e s , t h e components (i = 1 , 2 ) of t h e s t r e s s t e n s o r were

first expressed by t h e product o f a c e r t a i n prescribed f u n c t i o n f ( z ) and t h e , 2 ) which had t o be determined. The component d3was function tpi (xJ ) ( j = 1 taken as zero. The function f ( z ) was most o f t e n expressed by t h e equation
f (2) = - (22- h2).

1 2

More general forms of t h e f u n c t i o n f ( z ) w e r e a l s o considered. The expressions f o r f ( z ) similar t o (a), as w e l l as t h e condition t h a t s h a l l vanish, do n o t permit s a t i s f a c t i o n of t h e boundary conditions on t h e boundary surfaces, except f o r t h e case when t h y r e i s no load on them.

g3

The displacement v e c t o r components u* (i = 1 , 2 ) were determined from t h e expressions f o r t h e components d 3 on t h e b a s i s of Hookeqs l a w i n terms o f t h e , and t h e d e r i v a t i v e s of u3 w i t h r e s p e c t t o t h e coordifunctions f ( z ) , cp* (xJ) n a t e s XJ. T h i s solved t h e problem of reduction, and t h e f u r t h e r formulation of t h e problem proceeded i n t h e u s u a l context of s h e l l theory. I f we t u r n t o our approximation equations (7.5a) (7.5b), it will be noted that, i n c l u d i n g terms w i t h t h e f a c t o r (z h2 ), t h e s e equations a l s o contain a d d i t i o n a l

terms depending on t h e load on t h e boundary s u r f a c e s and permitting s a t i s f a c t i o n of t h e boundary conditions on them. The r e l a t i o n s (7.5a) ( 7 . 3 ~ ) confirm t h e a p p l i c a b i l i t y of t h e expression of t h e function f ( z ) by t h e equation ( a ) in t h e absence of l o a d s on t h e bounda r y surfaces of t h e s h e l l . Even i n t h i s case, however, t h e components 033 cannot be equated t o zero.

3 . Reduction by Determining t h-~ e C o e f f i c i e n t s of t h e Expansion of t h e D i s -

placement Vector Compcnmt.s __ _____i n S e r i e s , i n S p e c i a l Functions of t h e z Goordinate /172

1.N.Vekua (Bibl.18) formulated t h e boundary problems o f t h e theory of s h e l l s o f v a r i a b l e thickness, solvable by c a l c u l a t i n g t h e c o e f f i c i e n t s of t h e expansion of t h e e l a s t i c d i s lacement components i n s e r i e s i n Legendre polynomials of t h e coordinates = 2 . The content of Sect.24 of t h e p r e s e n t book also belongs t o this trend.

I n Sect.24 we gave a method of determining t h e c o e f f i c i e n t s of t h e Fourier

170

expansions of displacement v e c t o r components in trigonometric s e r i e s over t h e segment (-h, + h ) of a normal t o t h e b a s i c s u r f a c e of t h e s h e l l . Here we used Obviously, t h i s general equation pert h e general equation of dynamics (15.1). mits t h e construction of equations f o r determining t h e c o e f f i c i e n t s of t h e expansion of t h e required q u a n t i t i e s i n s e r i e s i n any s p e c i a l function, and t o f i n d t h e general formulations of t h e corresponding boundary problems.

4.

Generalized Fonrmlatfons of t h e Dynamic- .Problems of t h e Theory of P l a t e s and S h e l l s

I n t h e l a s t decade, a new d i r e c t i o n has developed i n t h e dynamics of p l a t e s a n d s h e l l s , with a c h a r a c t e r i s t i c d e p a r t u r e from t h e c l a s s i c a l formulat i o n of t h e corresponding boundary problem and t h e use of r e f i n e d equations. The problems t h a t encouraged the development of this l i n e of i n v e s t i g a t i o n were mentioned in Subsection 1 of t h i s Section.
W e s h a l l n o t analyze t h e numerous i n v e s t i g a t i o n s by Soviet and f o r e i g n a u t h o r s i n t h i s f i e l d of applied theory of e l a s t i c i t y . These s t u d i e s were characterized by t h e d e s i r e t o o b t a i n an approximate mathematical d e s c r i p t i o n o f c e r t a i n r e s t r i c t e d c l a s s e s of dynamic processes i n s h e l l s , which with s u f f i c i e n t accuracy r e f l e c t t h e experimental f a c t s and t h e conclusions from c e r t a i n exact s o l u t i o n s of t h r e e d i m e n s i o n a l dynamic problems.
A s an example of t h e s t u d i e s belonging t o t h i s trend, we might c i t e o t h e r a u t h o r s (Bib1.29,32) who used t h e method given by u s (Bib1.23a,b) and extended it t o t h e dynamic problem of t h e theory of p l a t e s and c y l i n d r i c a l s h e l l s . The The work of general theory of s h e l l s was n o t touched i n t h e s e studie+. (Bib1.32) d e r i v e s an approximate theory of wave processes i n p l a t e s and /173 shells which s a t i s f a c t o r i l y r e p r e s e n t s t h e experimental r e s u l t s and t h e conc l u s i o n s from t h e s o l u t i o n s of three-dimensional problems. Section 31. Comparison of Various Methods ~ of - Reduction -

I n conclusion, we s h a l l give a b r i e f comparison of t h e various methods o f reduction, demanding again optimum s a t i s f a c t i o n of t h e equations of t h e mathematical theory of e l a s t i c i t y by t h e s o l u t i o n s found from t h e equations of s h e l l theory.

It i s well known t h a t exact s o l u t i o n s of t h e boundary problems of elast i c i t y theory must s a t i s f y t h e equations of motion, t h e Saint-Venant compatib i l i t y conditions, t h e r e l a t i o n s r e s u l t i n g from Hookets law, and t h e boundarg

* With

r e s p e c t t o t h e work of (Bib1.29) we must make two statements. a ) It i s impossible t o c o n s t r u c t an approximate theory f r e e from hypotheses??, s i n c e any method of formulating a n apprdximate theory w i l l contain some a p r i o r i p o s t u l a t e , f o r example t h e p o s t u l a t e t h a t it i s p o s s i b l e t o base t h e theory on a f i n i t e segment of a Taylor s e r i e s expressing t h e f u n c t i o n t h a t i s Thus, w e can speak only of t h e r e l a t i v e accuracy of an apt o b e dete&ed. proximate theory advanced. b ) The author r e f e r s t o t h e work by Kennard and t o another monograph (Bibl.3b), ignoring e a r l i e r i n v e s t i g a t i o n s , although t h e method given t h e r e i s d i r e c t l y connected with t h e content of t h e a u t h o r t s own work.

i n

and i n i t i a l conditions, L e t u s now consider t h e approximate systems o f equat i o n s of t h e dynamics of s h e l l s proposed by US.

1 . Equations Obtained by Use of B p a n s i o n in Tensor S e r i e s

~~ ~ ~

The b a s i s o f d i s c u s s i o n here i s t h e series expansion of t h e displacement v e c t o r components, I n terms of t h e c o e f f i c i e n t s o f t h e s e expansions we exp r e s s t h e c o e f f i c i e n t s of t h e expansions i n series of t h e strain and stress t e n s o r components, For this reason we s a t i s f y : a ) t h e Saint-Venant compatib i l i t y equations; b ) Hookets l a w ; c ) t h e c o n d i t i o n s on t h e boundary s u r f a c e s o f t h e shell. The equations o f motions are s a t i s f i e d approximately, s i n c e we used them i n t h e first v e r s i o n only t o detelmine t h e c o e f f i c i e n t s o f t h e expansion i n t e n s o r series of t h e displacement v e c t o r components, and subsequently used only f i n i t e segments of t h e s e series. In t h e second version, which i s c l o s e t o t h e c l a s s i c d theory, t h e equations of motion are used i n t h e i n t e g r a l form.

2 , Equations Resulting f r o m t h e DtAlembert-Lagrange P r i n c i p l e

B y analogy t o t h e preceding, here we s a t i s f y t h e Saint-Venant compatibili-

t y conditions and Hookets law, The conditions on t h e boundary surface a r e i n cluded i n t h e equations of motion. The equations of motion are s a t i s f i e d i n t e g r a l l y and approximately i n consequence of t h e r e s t r i c t i o n of t h e n W e r of . degrees of freedom o f t h e s h e l l i n t h e d i r e c t i o n of t h e coordinates x? = Z For comparison, we r e c a l l t h a t t h e s o l u t i o n s of t h e equations o f t h e /17L s t a t i c s of p l a t e s given by E.Reissner s a t i s f y t h e equations of equilibrium o f t h e theory of e l a s t i c i t y , Hookets law, t h e boundary conditions on t h e boundary s u r f a c e s of t h e p l a t e , and i n t e g r a l l y (approximately) t h e compatibility condit i o n s (B i b l ,203)

Consequently, each method of reduction permits u s t o f i n d only an approximate s o l u t i o n of t h e t h r e e d i m e n s i o n a l problem of t h e theory of e l a s t i c i t y ,

w i t h t h e c h a r a c t e r of t h e e r r o r depending s u b s t a n t i a l l y on t h e reduction method, There i s no method of reduction t h a t does n o t involve some assumptions expressed in geometrical o r a n a l y t i c form,

CHAPTER I V
APPROXIMATELY E Q U I V A ~ N TSYSTEN.5
Section 1. Introductory Remarks The question of reducing t h e three-dimensional problem of t h e e l a s t i c i t y t o a two-dimensional problem of t h e theory of s h e l l s is case of t h e more general problem of t h e a p p r o b t e replacement of system by another which i s c l o s e t o t h e f i r s t one by some d e f i n i t e

/175

theory of a special. one m a t e r i a l criterion.

For example, i n reducing t h e three-dimensional problem of t h e theory of e l a s t i c i t y t o a two-dimensional problem, we replaced t h e three-dimensional e l a s t i c continuum-shell by a c e r t a i n medium, having t h e p r o p e r t i e s of cont i n u i t y i n two-dimensional space and a f i n i t e number of degrees of freedom i n t h e t h i r d dimension. Thus t h e s h e l l , i n V.Z.Vlasovts terminology, i s a discrete-continuum system (Bibl. 3b). This system approximately r e p l a c e s t h e three-dimensional e l a s t i c body. Clearly, t h e problem of constructing approximately equivalent systems i s considerably broader t h a n t h e problem of reduction.
The construction of approximately equivalent m a t e r i a l systems i s r e l a t e d i n i t s meaning t o t h e determination of a system of functions approximately representing another p r e s c r i b e d system of functions. For t h i s reason it i s n a t u r a l t o use t h e a n a l y t i c apparatus of t h e t h e o r y of approximation i n solving t h e problem of constructing approximately equivalent systems. I n t h i s Chapter we s h a l l consider t h e a p p l i c a t i o n of c e r t a i n methods of t h e theory of approximation functions t o t h e f i n d i n g of approximate a n a l y t i c statements of t h e dynamic boundary problems of t h e theory of s h e l l s . Section 2. First Nethod o E i n e a r Approximation of t h e Components of t h e S t r e s s Tensor and t h e Finite-Deformation Tensor
1. O n t h e Construction of a n I s o t r o p i c , Approximately E a d v a l e n t , E l a s t i c Body

I n t h e theory of small deformations i t i s assumed t h a t t h e nonlinear terms e n t e r i n g i n t o t h e composition of t h e components of t h e finite-deformation tensor ( 1 1 , 2.11) can be neglected, without introducing a substant.ial e r r o r /176 i n t o t h e f i e l d of s t r e s s e s . Here we shall consider t h e l i n e a r approximation of t h e components of t h e finite-deformation t e n s o r by t h e components of t h e small-deformation tensor. Such an approximation is a consequence of t h e construction of a n e l a s t i c medium approximately representing t h e motion of t h e corresponding elements of t h e body considered i n t h e i n i t i a l formulation of some nonlinear problem of elastoe obtain a b e t t e r f i e l d dynamics, by t h e motion of i t s elements. A s a r e s u l t w of s t r e s s e s than i n t h e theory of small deformations.

Let us imagine t h a t t w o e l a s t i c bodies with non-coinciding e l a s t i c cons t a n t s have, i n t h e undeformed s t a t e , t h e same g m m e t r i c a l form and dimensions and a r e r e f e r r e d t o i d e n t i c a l systems of Lagrangian coordinates xi. Let t h e deformed s t a t e of t h e f i r s t body be c h a r a c t e r i z e d by t h e finite-deformation t e n s o r D i k , and t h e deformed s t a t e of t h e second body by t h e small-deformation t e n s o r 6 1 k y which e n t e r s i n t o t h e l i n e a r p a r t of t h e components of t h e tensor Di,

W e s h a l l consider t h e s e bodies as approximately equivalent m a t e r i a l systems i f t h e e l a s t i c c o n s t a n t s of t h e second bod? a r e such t h a t t h e y satisfy t h e condition of t h e l e a s t - s q u a r e d e v i a t i o n of t h e i r s p e c i f i c p o t e n t i a l energies of ; of v a r i a t i o n of t h e t e n s o r components $ ) k i , defined deformation i n some region C elsewhere (11, 2.5). W e r e c a l l t h a t t h e t e n s o r components Dik and a r e cons t r u c t e d from t h e t e n s o r components h i ,as we have shown i n Cnapter 1 1 . The p o i n t s o f t h e region R a r e i n d i v i d u a l i z e d by t h e coordinates

Consequently, tne r e g i o n ;2 i s a nine-dimensional space. Let us assume t h a t each of t h e coordinates a k v a r i e s from zero t o some p o s i t i v e and negative q u a n t i t y a i k , which may be selected, f o r example, on t h e b a s i s of t h e requirement k h a t the f i r s t body s h a l l have no p l a s t i c deformations. It i s a l s o poss i b l e t o use o t h e r methods f o r s e l e c t i n g t h e q u a n t i t i e s a i k , based on kinematic considerations connected w i t h t h e r e s t r i c t i o n s imposed on t h e components of n i k W e s h a l l make use of t h e kinematic r e s t r i c determined previously (11, 2.7). t i o n s i n determining t h e bound.aries of t h e region R i n t h e following subsect i o n , b u t here we will use t h e condition of t h e absence of p l a s t i c deformations. Let us assume f i r s t t h a t t n e boundary of t h e r e g i o n R i s known. Making use of (11, 11.2a), and assuming t h a t both bodies a r e i s o t r o p i c , we f i n d t h e s p e c i f i c p o t e n t i a l energy of deformation of t h e f i r s t body:

The s p e c i f i c p o t e n t i a l energy of t h e second body i s expressed similarly:

/177-

Making use of e q s . ( I I ,

2.5),

(11, 2.6),

(11, 2.11) and (2.1),

we f i n d

174

and a f t e r elementary transformations we o b t a i n

Consider t h e i n t e g r a l

Let us f i n d t h e e l a s t i c components condition t h a t I s h a l l b e minimum.

X.i$

and

~9

of t h e second body from t h e

Equating t h e d e r i v a t i v e s

% * and

31 , t o apt

zero, we o b t a i n t h e following system of l i n e a r a l g e b r a i c equations

(2.6

where

It can be shown t h a t t h e determinant of t h e system of equations (2.6) i s e can convince ourselves of this by determining t h e e l a s t i c connonzero. W s t a n t s 1%-and p* from t h e system (2.6).
Since t h e l i n e a r i z a t i o n performed here r e l a t e s t o t h e s t r e s s e d - s t r a i n e d s t a t e of one of t h e elements of a n e l a s t i c body, i . e . , s i n c e i t i s l o c a l , l e t us introduce a l o c a l Cartesianrectilinearsystem of coordinates connected w i t h this element. I n t h i s system, t h e components of t h e metric t e n s o r a r e expressed by t h e well-known equations : gi = 1; gik = 0 ; (i # k)

To simplify t h e s o l u t i o n of this problem without c o n f l i c t i n g with i t s physical content, l e t us s u b s t i t u t e f o r t h e region of i n t e g r a t i o n I ; t h e extended region W, assuming t h a t each of t h e nine coordinates Xik v a r i e s from -a t o +a, where a i s t h e g r e a t e s t of t h e a b s o l u t e values assumed by t h e coord i n a t e s xik on t h e boundary of t h e region s1. Under t h i s condition, t h e region R Will b e included i n Si. Let us find, under these assumptions, t h e ex(2.7b). pressions f o r t h e c o e f f i c i e n t s bi, and b, determined by eqs.(2.7a)

From eq.(2.3a) we o b t a i n
3 3 3

i-1 3 3

i = l k-1

(2.8a)

r-1 j - 1
3 3

(2.8b)
3

i-1 k - l j - 1

(2.8~)
3 3

i-1 k = l j - 1 q = l

(2.8d)
17b

I I

1111I

I I

W e put

Consequently,

I n c a l c u l a t i n g bi, a l l terms containing odd powers of t h e v a r i a b l e s Si1, m u s t be excluded i n advance, s i n c e t h e limits of i n t e g r a t i o n a r e symmetric. W e find

1-1 k-1 j - 1

(2.10a)

(2."

To c a l c u l a t e t h e nine-fold i n t e g r a l s e n t e r i n g i n t o t h e expression f o r t h e c o e f f i c i e n t s of eqs. (2.6), i t i s a p p r o p r i a t e t o use approximation formulas. The a n a l y t i c p r o p e r t i e s of t h e i n t e g r a l expressions are v e r y simple and permit t h e use of formulas approximately expressing double and t r i p l e integrals*.
C f . , f o r instance, Sh. Ye. a e l a d z e , Numerical Methods of Mathematical Analysis, Gostekhizdat, 1953, p.507.

Performing t h e c a l c u l a t i o n s , we f i n d

z 4847,0a13; bI1 665,7a13; bI2= b,, z 1161,0a13; b,, s

bl

= (732,9A+ 561,2pj

u15; b,

(2890,Oh

+2245,Cp)

(2.11.a)

(2.11b)

From t h e system of equations (2.6) we o b t a i n

A* = A (1

p*=

+allu2)+ pa,2u2; laz,a* + p (1 + ap2a2).

= 0,4485.
0,061 1,

(2.12a)

(2.12b)

where

a], - 0,1057, aI2

Q ~ G ,

0,5709,

The terms i n a2 approximately determine t h e e f f e c t exerted b y t h e nonlinear terms, e n t e r i n g i n t o t h e components of t h e finite-deformation tensor, on t h e s t r e s s - t e n s o r components. Tnese terms are equivalent t o a c e r t a i n i n c r e a s e i n t h e Lame' constants.

TrJe note t h a t i n t h e case of t h e v a r i a t i o n of xik over nonsymmetric i n t e r v a l s , terms l i n e a r i n a would e n t e r i n t o eqs.(2.12a) (2.12b).

Now, on t h e b a s i s of (11, 4.5b), we can w r i t e t h e following r e l a t i o n s :

Hence, we f i n d

or

where

( 2 . 1 5 ~ )determine t h e r e q u i r e d l i n e a r approxEquations (2.1.l+), (2.15a) imations of t h e s t r e s s - t e n s o r components and of t h e f i n i t e - d e f o r m a t i o n tensor.

178
I

Several remarks will be given on t h e results. I n t h e c o n s t r u c t i o n of a n e l a s t i c body, approximately equivalent t o a body with f i n i t e deformations, we assumed t h a t t h e body t o b e constructed w a s i s o t r o p i c . If we abandon this assumption i t would be p o s s i b l e t o dispose of a l a r g e r number of e l a s t i c cons t a n t s and decrease t h e mean-square d e v i a t i o n of t h e s p e c i f i c p o t e n t i a l energy i ! from t h e p o t e n t i a l energy iI+ i n t h e region (2. Consequently, t h e cons t r u c t i o n of a body approximately equivalent i n t h e e n e r g e t i c c r i t e r i o n t o a body w i t h f i n i t e deformations W i l l l e a d t o t h e c o n s i d e r a t i o n of an a n i s o t r o p i c e l a s t i c medium.

2.

Connection with t h e Theory of Optimum Systems

The method given above f o r t h e c o n s t r u c t i o n of an e l a s t i c body which i s m e r g e t i c a l l y approximately equivalent, i s c l o s e l y r e l a t e d i n meaning t o t h e c o n s f r u c t i o n of what i s c a l l e d an optimum system, which i s known from t h e theory of automatic control+.
I n c e r t a i n problems connncted with t h e theory of noise, a l i n e a r func- /181 t i o n i s separated from t h e random f u n c t i o n d e s c r i b i n g a dynamic process i n cludin,? white noise. This s e p a r a t i o n i s based on t h e c o n d i t i o n of t h e mini" of t h e correspondinq mean-square e r r o r .

This problem i s analogous t o t h e above problem, which l e a d s t o a separat i o n of t h e l i n e a r f u n c t i o n s of t h e t e n s o r components Q i k from t h e components of t h e finite-deformation tensor. Here we have i n mind not only a n e x t e r n d . s i m i l a r i t y , b u t a more profound analogy of problems whose p h y s i c a l content i s d i f f e r e q t . Indeed, t h e process of v a r i a t i o n of t h e s t r e s s e d - s t r a i n e d s t a t e of a n e l a s t i c body has under a c t u a l conditions a random c h a r a c t e r and belongs i n The s e p a r a t i o n t h e f i e l d of problems s t u d i e d by p r o b a b i l i t y methods (Bibl.2b). of t h e l i n e a r p a r t of t h e s t r e s s and s t r a i n t e n s o r components, based on t h e requirement of a minimum of t h e corresponding mean-square d e v i a t i o n of t h e pot e n t i a l enerqies I I and Z+:may, i n t h i s connection, b e regarded as a p r a c t i c a l l y j u s t i f i e d s i m p l i f i c a t i o n of t h e mathematical d e s c r i p t i o n of a complex phenomenon, permittin2 a s e p a r a t i o n of t h e " p r i n c i p a l p a r t s " of t h e q u a n t i t i e s under study.

It i s easy t o e s t a b l i s h a d i r e c t c o r r e l a t i o n between t h e above-described method f o r t h e c o n s t r u c t i o n of a system approximately equivalent as t o t h e e n e r g e t i c c r i t e r i o n , and t h e methods of p r o b a b i l i t y . This, however, would go beyond t h e scope of t h e p r e s e n t i n v e s t i g a t i o n .

3.

Determination of t h e Parameter a

A s a l r e a d y s t a t e d , t h e parameter iz can be determined from v a r i o u s physical requirements imposed on t h e components of t h e t e n s o r s Dik and kk.

'de s h a l l s t a r t out from t h e Huber-Uses p l a s t i c i t y condition. According t o this c o n d i t i o n and t o t h e connection between t h e i n t e n s i t i e s of stresses
%

Cf. V.S. Pugachev, Theory of Random Fuiictions and i t s Application t o Probl e m s of Automatic Control, Chapter 16. fizmatgiz, 1960

179

and those of t h e s t r a i n s , l e t t h e region of e l a s t i c deformations of t h e materi a l be determined by t h e condition imposed on t h e i n t e n s i t y of t h e deformat i o n s . This condition, i n t h e Cartesian system of pectangular coordinates, has t h e following form:*

where A i s a c e r t a i n physical constant*. To determine t h e boundaries of t h e region n*, l e t us set a l l t h e coor- /182 d i n a t e s xlk, except one, as e q d t o zero and t h e n l e t us f i n d t h e values of t h e nonzero coordinate on t h e boundaries of t h e e l a s t i c i t y region from From t h e r e s u l t a n t values of IxikIn, l e t us s e l e c t t h e g r e a t e s t and eq.(2.16). assume t h a t a i s equal t o tihis quantity. Let
XI1

# 0; x,, =

. .. =

x32

= 0;

Then,

Let us put f u r t h e r

Then,

D,,. = 2
From eq.(2.16),
we f i n d

D,,= x ~ , ; D,,= ... = D,,= 0, 2

(4

Cf. A.P.Iltyushin, P l a s t i c i t y , Gostekhizdat, 1948; L.M.Kachanov, of t h e Theory of P l a s t i c i t y , Gostekhizdat, 1956.

Foundations

tensor i n t o t h e condition (2.16) i s controversial, s i n c e th,e condition (2.16) belongs t o t h e theory of small e l a s t o - p l a s t i c deformations, which, i n p a r t i c u l a r , i s The transnoted i n t h e book by R.Hill llMathematical Theory of i t i o n i n t h e conditions (2.16) t o t h e components of t h e small-deformation tensor introduces no s u b s t a n t i a l changes i n t h e conclusions of this Subsection.

M- The i n t r o d u c t i o n of t h e components of t h e finite-deformation

180

Comparing eqs. ( c ) and ( f ) we f i n d t h a t CL i s expressed by t h e equations

For s u f f i c i e n t l y small values of A we may approximately p u t

O f course, t h i s method of determining t h e region 0 5 i s c o t p e r f e c t . Est h e region fi s e n t i a l l y we have confined ourselves t o t h e r e s u l t s of lrsoundingrr only i n t h e d i r e c t i o n of t h e axes of t h e multi-dimensional coordinate system )Lik and, i n a d d i t i o n , we have used highly s i m p l i f i e d concepts as t o t h e s t r u c t u r e of t h e region n. The method of "soundingr1 used here does not ref l e c t t 3 e i n f l u e n c e of t h e components of t h e antisymmetric t e n s o r R i k on t h e nonlinear terms contained i n t h e components of t h e finite-deformation tensor D i k . W e have l i k e w i s e considered t h a t t h e p o s i t i v e and negative s i g n s f o r t h e coordinates xik were equally probable. T h i s l e d us, i n p a r t i c u l a r t o t h e conclusion t h a t A$5 > A and +* > p.

These conclusions may be i l l u s t r a t e d by an elementary one-dimensional sxample. From t h e r e l a t i o n

i t follows t h a t , f o r 0, E % > E, while f o r c L 1 < 0, E+* < E, and t h a t t h e a b s o l u t e value of t h e d i f f e r e n c e E+>- E i s t h e same i n t h e s e cases.l However, i n apgroximation on t h e symmetric i n t e r v a l of t h e p o t e n t i a l energy -zE ( G + ~ + )" by t h e energy E * we g e t t h e r e s u l t t h a t not always E * E.

The one-dimensional case d i f f e r s from t h e o t h e r s p r e c i s e l y i n t h a t a l l t h e components of P i , vanish here. This confirms t h e conclusion t h a t our deductions a r e i n s u f f i c i e n t , because of t h e f a c t t h a t they a r e based on a simpl i f i e d concept of t h e s t r u c t u r e of t h e region n and on t h e use of approximat i o n over t h e symmetric i n t e r v a l . These shortcomings may prove s u b s t a n t i a l for t h e case of s h e l l s , s i n c e considerable displacements and r o t a t i o n s of t h e elements may take p l a c e t h e r e , wiLhout t h e appearance of p l a s t i c zones. L e t us, t h e r e f o r e , consider a diff e r e n t choice of independent v a r i a b l e s and replace t h e region of approximate r e p r e s e n t a t i o n of t h e s p e c i f i c p o t e n t i a l energy of f i n i t e deformations I7 by

181

t h e s p e c i f i c p o t e n t i a l energy of small deformations ii*. Section

3.

Second Method of Linear Appro-&aion of t h e Components of t h e S t r e s s Tensor and of t h e Finite-Deformation Tensor

L e t us r e t u r n t o eq.(II, 9.1):
1
D i k = E l k f y g

(&irekj

irQkj

kj9ir

&irQ&j)-

This equation shows t h a t i t i s p o s s i b l e t o u t i l i z e t h e nine q u a n t i t i e s c i k and filk d i r e c t l y as coordinates of t h e region ;2. Tnere i s no need, however, f o r r e p e a t i n g a l l t h e c a l c u l a t i o n s given i n t h e l a s t Section. Let us f i r s t f i x t h e components S G k , considering them as c e r t a i n parame t e r s . This i s e q u i v a l e n t t o a s e p a r a t i o n , i n t h e nine-dimensional space Si, of a six-dimensiord space of deformations w. To c a l c u l a t e t h e i n t e g r a l s on t h e right-hand s i d e s of eqs.(2.6), it i s s u f f i c i e n t t o consider t h e transformation of coordinates according t o t h e formulas

Let

where max Isik I ;, i s t h e g r e a t e s t a b s o l u t e magnitude, among t h e s e t of V a l - /184 ues, taken by t h e component c i k on t h e boundary of t h e region w. The quantit y a i s d e t e r d q e d from t h e conditio? (2.16). Hereafter, i n c a l c u l a t i n g t h e i n t e g r a l s e n t e r i n g i n t o t h e expressions 1% and I.&+:-,we shall consider t h e extended region &%- (Sect.2) determined by t h e q u a n t i t y a.

L e t us now put

The v a r i a b l e s lllk may vary over a r b i t r a r y i n t e r v a l s l y i n g w i t h i n t h e range (-1, +1). Bearing i n mind t h e conclusions of our study of approximation over a symm e t r i c range, drawn i n t h e l a s t Section, l e t us assume t h a t all t h e quantit i e s Tjlk vary over a n i n t e r v a l (cy, S ) where I c y 1 and 131 are proper f r a c t i o n s . The choice of t h e t o t a l range of v a r i a t i o n f o r a l l t h e v a r i a b l e s 711, i s a subs t a n t i a l s i n p l i f i c a t i o n of the problem. I n performing s p e c i f i c c a l c u l a t i o n s we s h a l l most o f t e n assume t h a t cu = 0, $ = 1 o r cy = -1, i3 = 0. Xowever, most of t h e conclusions drawn below do not depend on t h e v a l u e s of cr and p. L e t us

182

r e t u r n t o eqs.(2.7a) (2.7d) and (2.8a) (2.8d) i n order t o f i n d new expresBearing i n mind eqs.(3.3), s i o n s f o r t h e c o e f f i c i e n t s bi, and bi of eqs.(2.6). we o b t a i n i n s t e a d of eq.(2.8a)

Equations (2.7a)

( 2 . P ) now t a k e t h e f o l h i n g form:

I n c a l c u l a t i n g t h e q u a n t i t y b, l e t us m.ke use of t h e v a r i a b l e s xik. Applying eqs.(3.1), (3.3) and making use of eqs.(2.7c) - (2.7d), w e find

From eqs.(2.7c)

(2.7d), {2.8a)

(2.8b), and (3.3) w e obtain

where

Calculating t h e summands i n t h e right-hand s i d e of eq.(b), we f i n d , on t h e b a s i s of eqs.(3.1), a f t e r s e v e r a l transformations and a f t e r proofs t h a t t h e sums l i n e a r l y containing i l i k vanish*, t h e following

W e have, f u r t h e r ,

The t r i p l e sums e n t e r i n g i n t o eq.(3.7c) t h e b a s i s of t h e following equation:

3. 3

may b e calculated, f o r example, on

Here, of course, one must remember t h a t t h e q u a n t i t i e s t h e i n d i c e s and t h a t t h e Gik are antisymmetric.

Eik

are symmetric i n

From t h e transformations leading t o eq.(3.'7d) i t is c l e a r t h a t t h e s u m /186 of terms with a n odd dimension r e l a t i v e t o t h e components of t h e t e n s o r G k must vanish. Let us introduce t h e n o t a t i o n

Q -1

Now, bearing i n mind eqs.(3.3),

we f i n d

i-1 k - 1
3 3 3

i-1 k - 1 j - 1
3 3 3
\ 2

form.

The right-hand s i d e of eq.(3.9> can be represented i n a somewhat d i f f e r e n t Let us introduce t h e n o t a t i o n

3

(3.10)
j-1

Then,

'B2 = c 2

+a
3

/187
Bikeik

Bikqik

Ga2
i-1 k-1
3 3

-k

1-1 k-1
3

From eqs.(3.6a) - (3.6b), (3.7a), (3.9) eral representations of the quantities bi :

- (3.11) result the following gen-

A ( clod4
r

+c,,A2a+cl,A2d +cI3a3 +c14a4) +

3 3

In these equations, the coefficients Cik, dik, a i k , $ik, yik, b i k do not depend o n the parameter a n o r the components of the tensor C i k . These coefficients are expressed by the folladng sextuple integrals over the region u, arithetized by the coordinates Gk :

186

/188

!Jqder t i e above assumptions as t o t h e v a r i a t i o n i n t e r v i l

P
1 . .

d U = J . .
a

s
P

: ;1

:,ik

, we

have

. . . d ~ ,.,. . dqZ3.

(3.W

D.

;>,le s t i l l have t o determine t h e q u a n t i t y a. Let us t u r n again t o t h e ccnd i t i o n of p l a s t i c i L y (2.16), r e p l a c i n g i n i t , according t o t h e theory of small e l a s t o - p l a s t i c defornations, t h e components of t h e Z5nite-deforrnation t e n s o r by t h e components of t h e small-deformation ten-,or.
T h e boundary of t h e p l a s t i c i t y region i s a sur.fac:: of t h e second order i n the six-dimensional space of q u a n t i t i e s E l k . Let u s f i n d t h e o i n k s of i n t e r s e c t i o n of t h i s s u r f a c e with t h e axes 8 i k and l e t us f i q d l E i k inen, according t o eq.(3.2), we shall f i n d a.

P)(.

m7

Let us p u t f i r s t obtain

ll

0;

... - ca3
-

e.

Then, from eq.(2.16),

we

PTittin? c I 2

0 and

... - c23

= 0,

we f i n d

187

Comparing eqs ( c) and ( d ) , we conclude t h a t

where E i s Youigls modulus, v i s Poisson's constant, and points.

i s the yield

The d i f f e r e n c e between t h e value of a found here and t h a t found i n t h e preceding Section, as w a s t o b e expected, i s i n s i g n i f i c a n t .
If we r e t a i n t h e components of t h e small-deformation t e n s o r i n t h e condit i o n s (2.16), then t h e q u a n t i t y a Will n o t b e connected with t h e quantit i e s ' dik These values, however, a r e s t i l l r e s t r i c t e d by c e r t a i n conditions, -to b e mentioned below.

Again, eqs.(2.6) permit u s t o f i n d A* and P*.

W e obtain

where t h e q u a n t i t i e s Aik are f u n c t i o n s of t h e q u a n t i t i e s a, cy, $ and of t h e The form of t h e s e f u n c t i o n s i s d e t e r antisymmetric t e n s o r components i i i k . mined b y t h e composition of t h e c o e f f i c i e n t s b,, and b , . Since i n A s and +>% t h e r e e n t e r t h e parameters r h k , which depend on unknown components of t h e displacement v e c t o r , t h e formulas (3.16a) - (3.16b) ( 2 . 1 5 ~ ) . To concannot b e d i r e c t l y used f o r s u b s t i t u t i o n i n t o eqs.(2.14) s t r u c t t h e f i r s t approximation t o t h e s o l u t i o n of t h e nonlinear problem of t h e theory of e l a s t i c i t y , t h e obtained expressions f o r A* and p-2 must b e averaged over Q i k , and this w i l l b e discussed below.

A comparison of t h e c o e f f i c i e n t s b, , determined by eqs.(2.l0a) - (2.10b) and (3.l2a) - (3.12b), shows t h a t t h e presence of f i n i t e r o t a t i o n s of t h e e l e ments of t h e body has a noticeable e f f e c t on t h e p r o p e r t i e s of t h e approximat i o n s being considered. O f importance i s l i k e w i s e t h e choice of t h e i n t e r v a l (a, 6 ) . A t c e r t a i n values of t h e q u a n t i t i e s s,, a, cy, 3, t h e right-hand s i d e s of eqs.(2.6) may vanish, w h i l e a t o t h e r values of t h e s e q u a n t i t i e s , t h e

X-

See t h e above-cited book by I l l y u s h k i n , pp.98-loO; whence follows eq. (3.15).

we have A = - -3~ ; e . - - =I J/2 ' - 3G'

188

f u n c t i o n s Ai, may t a k e negative values. For b, = 0 , obviously, the. convent i o n a l l i n e a r i z a t i o n i s permissible, c o n s i s t i n g i n a s u b s t i t u t i o n of t h e t e n s o r components Dik by t h e t e n s o r components & i k .

For negative v a l u e s of t h e f u n c t i o n s A , , , t h e r e i s a l o c a l diminution of t h e reduced e l a s t i c constants h+ and F-% b y comparison with h and p , w.hich i s i n t e r p r e t e d as a l o c a l decrease i n t h e r i g i d i t y of t h e m a t e r i a l . For posit i v e A i r , t h e r i z i d i t y of t h e material undergoes an apparent i n c r e a s e . Consequently, t h e n o n l i n e a r i t y of t h e components of t h e t e n s o r D i k l e a d s t o t h e development of a quasi-inhormgeneity of t h e mechanical p r o p e r t i e s of t h e m a t e r i a l , wnich may be c a l l e d kinematic.

O f course, t h e above statements can merely b e regarded as c e r t a i n heurist i c conclusions which r e q u i r e more d e t a i l e d j u s t i f i c a t i o n . For t h i s reason, we s h a l l present a d d i t i o n a l explanations.

1 . Preliminary S e l e c t i o n of t h e Region of Approximate Representations of t h e P o t e n t i a l Energy I: Sy t h e Energy I X

The choice of t h e region w i s of fundamental s i g n i f i c a n c e . This i s known from t h e theory of approximation f u n c t i o n s b u t i s a l s o c l e a r from t h e preceding argument. If we dispose a r b i t r a r i l y of t h e q u a n t i t i e s a, P and ; L i k , we may e v i d e n t l y impart almost any d e s i r e d v a l u e s t o t h e f u n c t i o n s A,, and reduce t h e s o l u t i o n of t h e problem t o a physical a b s u r d i t y . Concrete problems of mechanics d i s c l o s e r e l a t i o n s between t h e quantiFor t h i s reason, b y p r e s c r i b i n g t h e i n t e r v a l (a, 8 ) over and 0 ties 1 , . which t h e q u a n t i t i e s c i k vary, we a l s o impose c e r t a i n r e s t r i c t i o n s on t h e r e ,?ion of v a r i a t i o n of t h e q u a n t i t i e s z i k . The d i f f i c u l L y i s t h a t t h e s e res t r i c t i o n s a r e not p r e s c r i b e d i n advance i n t h e form of e x p l i c i t a n a l y t i c r e l a t i o n s . I n o t h e r words, t h e r e e x i s t s a c o r r e l a t i o n o f Lhe q u a n t i t i e s 1 , and . - i l k , b u t t h e 1imi';s of v a r i a t i o n of t h e c o r r e l a t i o n f a c t o r , d i f f e r i n g from t h e t r i v i a l cases of zero and u n i t y , a r e unknown.'?. 'de can only a s s e r t t h a t , as t h e region of v a r i a t i o n of 1, arld iiik i s sxpanded, t h e p r o b a b i l i t y inc r e a s e s f o r p o i n t s zorresponding t o t h e p h y s i c a l r e l a t i o n be'sween s i k and nlk t o f a l l w i t h i n t h i s region. The reader may r a i s e t h e q u e s t i o n whether i t is l e g i t i m a t e t o i n t r o d u c e t h e concept of c o r r e l a t i o n here. Indeed, i n s o l v i n g concrete boundary probl e m s of t h e e l a s t i c i t y theory, a d i r e c t c o n n e c t i v i t y i s e s t a b l i s h e d between Lhe values of c i k and nik, i . e . , i n t h i s case t h e c o e f f i c i e n t of c o r r e l a t i o n i s ullity. But t h e v e r y essence of t h e problem of approximation under conside r a t i o n here i s p r e c i s e l y t h a t t h i s approximation does not r e l y on t h e solut,iorl of any p a r t i a l problem. Imagine t h e s e t of p o s s i b l e deformed states of a body described b y t h e t e n s o r s c i k and ; 2 1 k . To each s t a t e t h e r e corresponds a p o i n t i n t h e six-dimensional r e g i o n of t h e q u a n t i t y 1, and t h e t h r e e dimensional r e g i o n of t h e components of i - d i k . I f we do not know t h e a n a l y t i c connection between t h e p o i n t s i n t h e s e spaces, t h e n t h e c o r r e l a t i o n a l connecti v i t y comes i n t o f o r c e . The c o e f f i c i e n t of c o r r e l a t i o n c h a r a c t e r i z e s t h e proba b i l i t y of t h e p h y s i c a l correspondence of p o i n t A of t h e f i r s t space t o p o i n t B of t h e second space.
9

On t h e basis of t h e above statements, we s h a l l not cansider approximation over a n a r b i t r a r i l y s m a l l i n t e r v a l . Ne s h a l l i n s t e a d consider a f i n i t e region i n t h e six-dimensional space of t h e q u a q t i t i e s ' I C k , bound5d by t h e condition:

For example, as s t a t e d above, l e t u s put b

1, tu = 0 o r B = 0, CY = 1 .

/l9l

The choice of a s u f f i c i e n t l y wide i n t e r v a l of v a r i a t i o n of t h e compone n t s of E i k e l i m i n a t e s t h e need f o r e s t a b l i s h i n g a n exact connectivity between c i k and R i k .

2.

Preliminary Delimitation of t h e Region of V a r i a t i o n of t h e Q u a n t i t i e s sL+lr

A s we know from t h e theory of deformation of t h i n rods, p l a t e s , and s h e l l s , f i n i t e displacements and r o t a t i o n s may simultaneously a r i s e i n t h e i r elements under s m a l l deformations E i k . A s a l r e a d y mentiorled, t h e r e i s a corr e l a t i o n between t h e q u a n t i t i e s e l k and < > i k . Since t h e c o r r e l a t i o n c o e f f i c i e n t i s unknowg, a r e g i o n of v a r i a t i o n of t h e q u a n t i t i e s x i k i s assigned a r b i t r a r i l y a t f i r s t ; f o r e,yample, i t , i s assumed t h a t t h e s e q u a n t i t i e s vary from zero t o - b where t h e q u a n t i t y b i s a t f i r s t a r b i t r a r i l y prescribed. Then we i n v e s t i g a t e A?? and k s a t v a r i o u s values of E l k l y i n g on t h e i n t e r v a l (aa, Ba) and v a r i o u s v a l u e s of : i l k , l y i n g on t h e i n t e r v a l (-b, +b). The v a l u e of t h e parameter b i s r e s t r i c t e d by t h e requirement t h a t A+:- and IL+* s h a l l b e p o s i t i v e and by t h e requirement t h a t t h e Poisso? constants vf:- s h a l l b e included i n t h e i n t e r v a l (0; 0.5).
O n preliminary determination of t h e region of v a r i a t i o n of t h e quantit i e s Lik, we may use experimental d a t a , for example t h e r e s u l t s of a study on t h e deformation o f s h e l l s under g r e a t displacements and angles of r o t a t i o n i n t h e supercritical stage.

The study of t h e v a r i a t i o n of I>-::- and ++:-i n t h e fow-dimeqsional region (aa, Sa; C q k ) i s i n i t s e l f a means of t h e q u a l i t a t i v e a n a l y s i s of s p e c i a l problems.

3.

Determination of t'ie Mean Values of

AX-

and p+*
i;ik,

E s t a b l i s h i n g f i r s t t h e v a r i a t i o n a l region Q , of t h e components us average t h e q u a n t i t i e s A+:- and p+:-i n t h i s region. Xe f i n d

let

. "

This averaging may b e done with t h e weignt p ( 2 l k ) i f i t i s p o s s i b l e t o i;idic a t s a :unction p(i-hk) on t h e b a s i s of experiments or t h e o r e t i c a l cons i d er a t i o n s . Tllen, eqs.(3.18a) - (3.l8b) are replaced by t h e following:

/192

driik can always b e rednced t o ) i n t e g r a l s w i t h i n limits l y i n g i n s i d e t h e i n t e r v a l (-1,1 ) . (Fb

By a chaqgc, i n s c a l e , t h e i n t e q - a l s 1'

...

The q u a n t i t i e s h?: and p+ are introduced i n t o eqs.(2.1/+) - (Z.l5c), and t,hereb;T we complete t h s ~o l u t i o n of t h e problem of l i n e a r approximaticn, i n f i r s t approximation. S e c t i o n 3a. Further Developnent, of t h e Method of Linear Appro.xb"aion
~~

A s noted i n Sect. 3, t h e determination of t h e region of v a r i a t i o n of t h e t e n s w components, over which t h e approximate r e p r e s e n t a t i o n extends, i s of fundameqtal importaqce i n performi?: t h e approximation of components of t h e filii-te-deformation t e r s o r by components of t'le small-deformation tensor.
i d L h t h e coordinates

I n Sect.3, we assumed t h a t t h e b a s i c region w a s a six-dimensional space E i k , a d . as an a7Lxiliar-y region t h e three-dimensional space with t h e coordinates i i i k . Ne s h a l l now supplem2nt t h e above.

If we know i n advance, from t h e conditions of t h e prablem, t h a t c e r t a i n components of t h e t e n s o r E i k o r ; i i k are zero, then t h e number of dimensibns of t h e b a s i c and auxiliary r e g i o n s i s correspondingly decreased. This c l e a r l y l e a d s t o o b v i m s changes i n t h e m u l t i p l i c i t y of the i n t e g r a l s e n t e r i n g i n t o t h e formulas of Sect.3.

(cya,

W e assumed t h a t a l l t h e q u a n t i t i e s c i k vary over t h e t o t a l i n t e r v a l + b ) . On t h e b a s i s of t h e contents of Sect.2, we may consider a more gene r a l case, i n d i v i d u a l i z i n g t h e v a r i a t i o n a l i n t e r v a l for each q u a n t i t y of E l k .

L e t us p l a c e t h e components

E, ?,

G~~~

33,

Cl2, 23,

and c g l i n corre-

191

spondence with t h e numbers 1, 2 , 3, 4, 5, 6. L e t t h e components of G i k vary over t h e i n t e r v a l s (aJ ail,, Bjaik) where t h e symbols j correspond t o p a i r s of numbers i, k i n t h e above-discussed manner. I n s t e a d of eqs.(3.3) we i n t r o - /193 duce t h e r e l a t i o n s

The transformation of t h e b a s i c r e l a t i o n s described i n Sect.3 i s obvious i n t h i s case, and we W i l l not r e p e a t them here. W e note t h a t t h e i n t r o d u c t i o n of i n d i v i d u a l v a r i a t i o n a l i n t e g r a l s of t h e components of e l k i s p o s s i b l e only i n those s p e c i a l cases where t h e condition of t h e problem of mechanics permits such i p d i v i d u a l i z a t i o n . I n e x a c t l y t h e same way, s e p a r a t e v a r i a t i o n a l i n t e r vals of t h e components of t h e a n t i s p e f u r i c t e n s o r & k may b e introduced.

L e t us continue our consideyation of t h e f u r t h e r development of t h e proposed method. Assume t h a t we have solved tke q u a s i - l i n e a r dynamic problem of t h e theory of e l a s t i c i t y w i t h t h e constants A-X- and ps. This quasi-linear solut i o n g i v e s a f i r s t approximation t o t h e s o l u t i o n of t h e nonlinear problem i n displacements. men, t h e t e n s o r components i ; k y i n f i r s t approximation, w i l l b e known f u n c t i o n s of t h e coordinates x J ( j = 1, 2, 3) of a n e l a s t i c body and of t h e t i m e t . Returnilng t o eqs.(3.16a) - (3.16b), we f i n d A* and ,A+ as f u n c t i o n s of XJ and t. S u b s t i t u t i n g t h e r e s u l t a n t v a l u e s of A"*. and 9-2 i n t o eqs.(2.14) and determining t h e t e n s o r components E i k from t h e f i r s t approximation, we f i n d t h e f i r s t approximation for t h e f i e l d of s t r e s s e s . The f i r s t approximation f o r t h e s t r e s s e s nil1 c o n t a i n nonlinear terms depending on C i i k .
O n t h e b a s i s of t h e f i r s t approximation, t h e t o t a l v a r i a t i o n a l i n t e r v a l of t h e q u a n t i t i e s t?ik can a l s o b e refined. However, a considerable decrease i n ?:?e diameter of t h e region LC% may l e a d t o t h e c o n t r a d i c t i o n s mentioned above. W e r e c a l l now t h a t t h e l i n e a r i z a t i o n of t h e components of D i k s t i l l does not, r e s u l t i n a complete l i n e a r i z a t i o n of t h e equations of elastodynamics, s i n c e o t h e r sources of nonlinear equations considered i n Chapter I1 a r e of considerable s i g n i f i c a n c e here. I n t h i s connection, we must a g a i n i n v e s t i g a t e t h e approximate method of l i n e a r i z a t i o n i n a somewhat more complex form, and then go on t o o b t a i n i n g f u r t h e r approfixnation. S e c t i o n lr. L i n e a r i z a t i o n i n a n Element of t h e S h e l l

The r e s u l t s of t h e preceding Sections permit a n approximate elimination of t h e nonlinear terms e n t e r i n g i n t o t h e equations expressing Hooke's l a w on i n t r o d u c t i o n of t h e components of t h e finite-deformation t e n s o r i n t o t h e s e equations. This, however, does not l e a d t o l i n e a r equations of motion of a n element of t h e s h e l l y 8 s i n c e t h e components of t h e s t r a i n t e n s o r e n t e r i n t o t h e f o r t h e volume of t h e defnvmed element. expression $%xldx'dx W e a r e now confronted by t h e following a l t e r n a t i v e : e i t h e r t o r e t a i n t h e approximate l i n e a r expressions (2.14) obtained above f o r t h e stress t e n s o r /I94 components connected with t h e s p e c i f i c energy of deformation and not t o c a r r y t h e l i n e a r i z a t i o n t o completion, o r e l s e t o consider t h e quasi-specific energy
192

of deformation determined by t h e e q u a t i o n

and, by introducing v a r i a t i o n s i n t o t h e r e s u l t s above, o b t a i n a complete l i n e a r i z a t i o n , and i n t h i s case t o e n t e r i n t o formal c o n t r a d i c t i o n with t h e w e l l known e n e r g e t i c p r i n c i p l e s of Hookes l a w .

It seems permissible t o make use of iqs.(4.1), s i n c e every approx5mate s o l u t i o n of a physical problem contains e r r o r s c o n t r a d i c t i n g t h e exact solutions.
Consider t h e i n t e g r a l

Making use of (11, 6.3) and r e t a i n i g g i n t h e i n t e g r a l expression a l l terms ) ~ o,b t a i n of t h e order of ( ~ ~ k we

From eqs.(2.3b),

i t follows t h a t

Let us r e p r e s e n t I i n expanded form:

O n comparing eos.(4.Lc) and (2.5) we conclude t h a t only t h e ri$ht-hand /195 s i d e s of eqs.(2.6) a r e c:?anged. The new right-hand s i d e s of eqs.(2.6) now have

It i s c l e a r f r o m eqs.(4.i+) t h a t t h e ci a r e expressed as follows:

where

.is abov?, w? now p a s s t9 a l o c a l system of r e c t a n g l d a r Cartesian coordinates, takin;?

g..I [ 1. 9
gjk=o

(i # k ) .

(b 1

To transform t h e expression ci, we must make use of tine r e l a t i o n s employed i n transforming t h e expressions f o r 4, and cjz i n t h e l a s t Section, and of t h e formula
. 3
I -

We f i n d

where
dw=
dqll

... d q 2 3 .

After transformations, Y , and Y 2 t a k e t n e following form:

/196

B y analogy t o eqs.(3.12a)

(3.12b),

we o b t a i n

c, = -- as i . (e,,tzeA2+ el3a3+el,u4)+2p ( / l , a z ~ z +

1 95

1111I

1-11

1 1 1 1 11111

. I 1m.111

I I, I ,

I . , . I

I.

I,

. ,

. ..... . ..... ..

The c o e f f i c i e n t s e n t e r i n g i n t o t h e s e fornulas are expressed by t h e following sextuple integrals :

196

A s i n Ssct.3, we may s u b s t i t u t e i n r e l a t i o n s (4.9a) r e s u l t a n t expressions (4.12a) - (4.12d):

(4.9b) and i n t h e

J . .. d w = s . .
a

.. . d q , , . .. dqg,.
/I98

I n some cases, as noted i n Sect.3a, t h e i n t e r v a l s of v a r i a t i o n of t h e v a r i a b l e s Tiik must be individualized. Then,

(w)

. .. dw =

r ... J
8,

P.

. . . dyI1. .. dqZ5.
4

. I

A n example of approximation over various v a r i a t i o n a l i n t e r v a l s of i n Sect.6. I n s t e a d of eqs.(3.l6a)

vi,

i s given

(3.16b),

we o b t a i n

197

111111 11111111 I1

1111111.1

I . II

..., ._. .__

~~-

The p r o p e r t i e s of t h e c o e f f i c i e n t s cik a r e analogous t o t h o s e of t h e coe f f i c i e n t s A,, i n eqs.(3,16a) - (3.16b) considered i n Sect.3.

'vie g i v e bel2w a n e x a p l e i l l u s t r a t i n g t h e contents of Sect.4. Section 5 . On t h e R e l a t i o n between Linear Approximation of t-he Components of t h e Finite-Deformation Tensor and t h e Method of Equivalent Lineari z a t i o n and t h e P r o b a b i l i t y MethodsI Further Stgges of Successive Approximations
~~~~

The above method of l i n e a r i z a t i o n f o r t h e s t r z s s tensor components and t h e fidbe-deforrration tensor components a r e c l o s e t o t h e methods of equivalent l i n e a r i z a t i o n , which we know from t h e n o d i n e a r mechanics of systems with a f i n i t e number of dsgrees of freedon+?.

The method applied above i s a l s o c l o s e l y r e l a t e d t o t h e p r o b a b i l i t y method of s o l v i n g t h e problems of mechanics. I n f a c t , t h e region Iri, c o n s i s t i n g of t h e regions I): and G, i n which t h e approximation i s performed, i s t h e region of llprobable statesll of a n e l a s t i c body. The method of approximation adopted by u s i s equivalent t o t h e hypothesis t h a t t h e s e s t a t e s a r e e q u p r o b a b l e , which. only approximately carresponds t o r e a l i t y .
I n i n v e s t i g a t i n g .lamow c l a s s e s of problems of t h e mechanics of s h e l l s , one must b e a r i n mind t h e r e s u l t s of experiments p e r m i t t i n g us t o c o n s t r u c t functions c h a r a c t e r i z i n g t h e p r o b a b i l i t y d i s t r i b u t i o n of t h e .appearance of cert a i n v a l u e s of t h e q u a n t i t i e s forming t h e r e g i o n s w and ;'*. By using prob- /199 a b i l i t y d i s t r i b u t i o n curves, we must introduce f u q c t i o n s of weight i n t o t h e i n t e g r a l s I of t h e preceding Sections, i.e., we must consider t b e weighted-square d.eviations. The quaqtities ai, entering i n t o e q s . ( 3 a . l ) a r e s p e c i d f w m s of t h e weighting functions. W e do not dispose of t h e necessary experimental d a t a and were t h e r e f o r e compelled to a b s t a i n from t h e use of weighted-square deviations. I n t h e development of S e c t i o n 3
%

-4

we g i v e a surmnary of t h e f i r s t s t a g e s

Various v e r s i o n s of t h e method of equivalent l i n e a r i z a t i o n can b e found by t h e readar i n t!ie following books: N.N.Bogolyubov, Yu. A. K i t r o p o l l s k i y , Asymptotic Methods i n t h e Theory of Nonlinear Vibrations, Fizmatgiz, 1958; S.Krendel1, Random V i b r a t i o n s of Systems w i t h Nonlinear Tiestoring Forces, Transactions of t h e Symposium on Nonlinear Vibrations, Kiev, 1961; Ya.G.Panovko, Action.of Periodic Impacts on a Nonlinear E l a s t i c System with One Degree of Freedom, Trudy In-ta F'iziki AN Latv. SSR, No.5, 1953

of t h e construction of approximate s o l u t i o n s f o r t h e nonlinear boundary probl e m s of t h e theory of s h e l l s , based on t h e use o f t h e l i n e a r approximation mekiod of t h e zomponents of t h e finite-deformation t e n s o r . L e t us consider only those problems i n which t h e n o n l i n e a r i t y i s connected with t h e efisbence of f i n i t e displacements and a n g l e s of r o t a t i o n . The components of ';ne s t r a i n t e n s o r E i k will b e regarded as q u a n t i t i e s small i n comp a r i s o n with u n i t y . Turning t o t h e general equation of dynamics (111,15.1), we note t h a t t h e l i n e a r approximation, with a n i n t r o d u c t i o n of averaged e l a s t i c constants and E-: considered i n t h e preceding Sections, makes i t p o s s i b l e t o l i n e a r i z e t h e o p e r a t o r s e n t e r i n g i n t o t h e elementary work of t h e i n t e r n a l f o r c e s 6A. The terms expressing t h e work of t h e body and s u r f a c e f o r c e s and of t h e i n e r t i a l f o r c e s W i l l , as beyore, s t i l l c o n t a i n nonlinear summands. But t h e s e nonlinear terms will not c o n t a i n components of Q k b u t w i l l depend i n s t e a d on t h e s c a l a r a ' k s i k i f we d i s r e g a r d a l l the second-order terms i n Lhe composition of D i k , which a f t e r m u l t i p l i c a t i o n by t h e i n e r t i a l f w c e s W i l l y i e l d terms w i t h a homoqeneity iqdex equal t o t h r e e with r e s p e c t t o t h e displacement components and t ' l e i r d e r i v a t i v e s . I f , i n agreement with most i n v e s t i g a t o r s , we n e g l e c t Che influence of t'le volumetric expansion g i k E i I ( on t h e v i r t u a l work of t h e i n e r t i a l forces, t h e body f o r c e s and t h e surface f o r c e s , then t h e approxh"aion r i v e n i n Sects.3-li will permit a l i n e a r i z a t i o n of Lhe system of equations of motion-::. The r e s u l t a n t system, i n accordance with Sects.3-3aY Will be denoted e note t h a t this system of as a system of equations of P i r s t approximation. W equations will d i f f e r from t h e systems of l i n e a r equations of Chapter 111, obt a i n e d by d i r e c t r e j e c t i o n of a l l nonlinear terms. The d i f f e r e n c e w i l l depend on ?'?e term p * e r E e n t e r i n g i n t o t h e composition of D i k according t o eq. (2.15b ).

x%-

W e W i l l noc f u r t h e r d i s c u s s t h e construction of t h i s system, s i n c e i t s method of d e r i v a t i o n does r l o t d i f f e r i n p r i n c i p l e from t h a t discussed i n Chapter 111. 4 f t e r deciding t o l i n e a r i z e t h e boundary problem i n displacements, we w i l l o b t a i n t%e f i e l d of s t r e s s e s , makin,? use of eqs.(2.14) and (4.13a) (4.13b), /200 t,o,?ether with eqs. (2.6) and t h e i r c o e f f i c i e n t s , found i n Sect .3-Ic.

A s a l r e a d y noted i n Sect.3aY t h e r e s u l t a n t expressions f o r t h e s t r e s s tensor components Will c o n t a i n nonlinear terms depending on t h e components of t h e Supplementing our remarks i n Sect.3aY i t will b e reantisymmetric t e n s w C i i k . c a l l e d t h a t t h e expressions for A+ and p L 0 i n Sects.2 - 4 were obtained i n a l o c a l Cartesian coordinate system. For t h i s reason, in considering t h e f i e l d of stresses, we m u s t transfarm t h e components of i-21, i n t o t h e l o c a l C a r t e s i a n system of coordinates, and o n l y then determine t h e q u a n t i t i e s of A* and IL*
: The terms of t h e ord2r 9f 6 i k and of higher o r d e r s i n t h e expression f o r t h e v i r t u a l work of t h e f o r c e s of i n e r t i a and t h e l i v i n g f o r c e s are neglected i n

Returning t o t h e questions considt h e equations of t h e monograph (Bibl.12). ered here, we note t h a t t h e appro.uimation given i n Sect.3 can be a p p l i e d i n t h i s case.

A s a result, we o b t a i n t h e first approximation f o r t h e s t r e s s t e n s o r conponents. To o b t a i n t h e second approximation, l e t us substitute i n t o eqs.(2.14) t h e values of A?:- and p$t found from t h e f i r s t approximation. W e recall that t h e s e values f o r A% and @+t i n t h e problems of s h e l l mechanics are f u n c t i o n s of the coordinates xi of t h e p o i n t s of i t s b a s i c surface and a l s o of t h e a coord i n a t e s . .&en s u b s t i t u t i n g , i n t h e v a r i a t i o n a l equation, all nonlinear terms which had been neglected i n o b t a i n i n g t h e first approximation, by q u a n t i t i e s found from t h e s o l u t i o n s of t h e equations of f i r s t approximation, we o b t a i n t h e system of l i n e a r equations of second approximation. Such a system of l i n e a r equations W i l l have v a r i a b l e c o e f f i c i e n t s . This complicates t h e s o l u t i o n of t h e problems, b u t , e v i d e n t l y permits a q u a l i t a t i v e a n a l y s i s of t h e s o l u t i o n s of t n e nonlinear prgblem, t h a t i s more profound t h a n a n analysis based on t h e equations of second approximation obtained b y t h e ordinary methods of t h e elasn example of t h e s e methods i s t h e use of t h e nonlinear Lame' t,ici",y theory. A e q u a t i o r s (11, 7.5a) - (11, 7.5b), where a d d i t i o n a l body f o r c e s g1 a r e d e t e r mined from t h e f i r s t , approximation. W e have a l r e a d y noted t h e disadvantages of such methods (11, Sect.8).
To f a c i l i t a t e t h e i n t r o d u c t i o n of t h e methods considered i n Sects.3-4, we present a n i l l u s t r a t i v e example i n Lhe following Section". This example 1All a l s o permit u s t o supplement t h e general c h a r a c t e r i z a t i o n of t h e s i g 2 i f i c a n c e of t h e method. Section 6.
On k i s y m n e t r i c Deformations and t h e Z l a s t i c S t a b i l i t y of a Cir:

c u l a r Tube Subejected t o t h e Action Forces

Df

Longitudinal Compressive

The heading of this S e c t i o n coincides with t h e t i t l e of another work Yere we make use of c e r t a i n results of t h a t paper, with t h e ob(Bib1.23d). j e c t of t h e i r f u r t h e r development, on t h e b a s i s of t h e method given i n Sects.3-4. A t t h e same t i m e , t h e method of approximation under consideration i s given /201 a n elementary i l l u s t r a t i o n , showing i t s promise.
W e consider t h e well-known problem of t h e s t a b i l i t y of a c i r c u l a r cylind r i c a l tube, compressed by l o n g i t u d i n a l f o r c e s uniformly d i s t r i b u t e d along t h e contour of t h e b a s i c (middle) surface i n t h e f a c e s e c t i o n s of t h e cylinder.

W e s h a l l study only t h e case of axisymmetric deformations, although t h e experimenkal and t h e o r e t i c a l results obtained i n t h e l a s t 15 y e a r s convincingly prove t h e s i g n i f i c a n c e not of p u r e l y &symmetric deformations, b u t of deforma0 ) . The assumption of t i o n s w i t h a c y c l i c symmetry about t h e a x i s (Bibl.4, 1 t h e p o s s i b l e existence, i n t h i s case, of axisymmetric forms of deformation i n t h e t r a n s c r i t i c a l s t a g e i s l i k e w i s e confirmed by experiments described i n o l d e r r e p o r t s . Evidently, i n t h i s work r e l a t i v e l y t h i c k s h e l l s were i n v e s t i g a t e d , and t h e symmetric forms of deformation w e r e accompanied b y s t r e s s e s exceeding t h e y i e l d p o i n t of t h e material-:+. Thus, we s h a l l confine ourselves t o a conThis example c a m o t b e the o b j e c t of t h e s t u d i e s i n Sects.3-4, s i n c e t h e extreme s i m p l i c i t y of t h e problem admits many more s o l u t i o n s than t h a t given below. .:-XCf. I.V.Gekkeler, S t a t i c s of a n E l a s t i c Body, ONTI, 1934, pp.271-276; S.P.Thoshenko, S t a b i l i t y of X l a s t i c Systems, OGIZ, 1946, pp.388-392
% -

200

s i d e r a t i o n of t h e adsymmetric deformations of a closed c i r c u l a r c y l i n d r i c a l s h e l l . Following t h e n o t a t i o n adopted earlier (Bibl,23d), we put

where t h e , c o o r d i n a t e x i s t h e d i s t a n c e from one of t h e f a c e contours of t h e undeformed middle s u r f a c e measured along t h e generatrix, and s i s t h e l e n g t h of a r c of t h e d i r e c t r i x measured from some i n i t i a l point, while t h e z coordinate i s measured along a normal t o t h e undeformed middle s u r f a c e i n t h e d i r e c t i o n toward t h e axis of t h e tube. I n l o n g a l coordinates (6.1) on t h e undeformed middle surface, we have

1 1 1 , Making use of, t h e method of successive approximations given i n ( Sects.9-10), we confine ourselves here t o t h e f i r s t approximation. Accordingl y , we put

where u, v, w a r e t h e displacement v e c t o r components of a p o i n t of t h e middle s u r f a c e of t h e s h e l l .

A s we know from Chapter 111, t h e expression ( 6 . 3 ) corresponds i n accuracy t o t h e reduction formulas r e s u l t i n g from t h e Kirchhoff-Love hypotheses. An i n crease i n t h e accuracy of reduction i s not necessary s i n c e t h i s example o n l y has an i l l u s t r a t i v e purpose.

I n &symmetric deformations, t h e displacement v e c t o r wmponents of a p o i n t of t h e middle surface a r e expressed as follows:

u = u ( x ) ; u = 0; w = rer(x).

/202

(a>

Using (III,9.1)y (111, 9.2),

we f i n d

where k = R-' i s t h e curvature of a s e c t i o n of t h e middle surface normal t o i t s generatrix. At t h e accuracy f o r constructing t h e equations adopted by u s y t h e s e same r e l a t i o n s are equivalent t o t h e conditions

201

e31 = e23 = 0;

=~ ( 1 ) .

IJsiqg (111, 10.la)

(111, 1 0 . 3 ~ ) w ~ e find

6.6)
a y If we remain w i t h i n t h e limits of t h e accuracy adopted by us, then we m approximately p u t

Thus, of t h e s i x s t r a i n t e n s o r components, t h r e e vanisn i r i t h i s case, and t!le component 53, n e g l e c t i n ? t h e summands of o r d e r z i n i t s composition,is expressed i n terms of c 1 and 6 2 2 .

Consequently, t h e l i n e a r approximation of t h e components of t h e f i n i t e deformation t e n s o r , i n t h e problen under consideration, must b e performed i n O f course, t h e ret n e two-dimensional space of t h e q u a n t i t i e s e;1 and s t r i c t i o n i n t h e ;?umber of space dimensioiis of p o s s i b l e deformations w a s obtained here as a r e s u l t of v e r y gross s i m p l i f i c a t i o n s . W e have given above t h e motivation f o r t h i s appToach t o t h e problem. O n t h e b a s i s of eqs.(6.3), we f i n d

(6.8a)
o:n, b e a r i n g i n mind. eqs.(6.4),

(6.8b) Hereafter we s h a l l c m s i d e r o n l y t h e mean v a l u e of ~5~ over t h e thickof .:he s h e l l , expressed 5;- t h e f o r n u l a

283s

From eq.(3.%),

we f i n d
At

/203

2 (v "31
202

)2.

(6.10)

Equations (318a)

(3.8b) y i e l d

B,, = - (9,1)2;
The o t h e r f u n c t i o n s Bik vanish.

1 4

B,, = -Further,
fQ3J4.

1 (Q3J2. 4

(6.11a)

1 c*=32 -

(6.11b) and (6.10)

Using eqs.(3.12a)
find

(3.Ub), (4.lla)

- (&.llb),

(6.11b),

w2

The q u a n t i t i e s A% and p% determine t h e required l i n e a r approximation i n t h e s p e c i a l case under consideration. I n c a 1 c d a t i n g t h e c o e f f i c i e n t s q k and n i k , eqs.(3.13a) - (3.13d) and (4.12a) - (4.12d) must b e used, b e a r i n g i n mind eqs.(6.5) - (6.7) and passi n g i n t o t h e two-dimensional region of i n t e g r a t i o n according t o t h e indicatLons c c i v e n i n Sect.3a.

The radial displacements w of t h e middle surface e s t a b l i s h a f i e l d of flexural s t r e s s e s . Therefore, t h e s h e l l i s divided i n t o two zones over i t s t h i c k n e s s , a zone i n which t h e t e n s i l e s t r e s s e s dominate and a zone i n which t h e compressive s t r e s s e s o11 dominate. I n t h e zone of t e n s i l e s t r e s s e s , t h e v a r i a b l e Tll i n t h e i n t e g r a l s (3.13a) - (3.13d) and ( 4 . 1 2 a ) ' - (4.12d) varies over t h e i n t e r v a l (0,l). I n t h e zone of compressive s t r e s s e s , t h i s v a r i a b l e W e assume t h a t t h e v a r i a b l e Tb2, correspondv a r i e s over t h e i n t e r v a l (-1,O). i n g t o t h e component e Z 2 of t h e s t r a i n tensor, varies over t h e symmetric i n t e r val ( 1 , l ) .
The numerical. v a l u e s of t h e c o e f f i c i e n t s q k and nik h e r e depend on A and p, T h i s r e l a t i o n i s due t o t h e f a c t t h a t 633 i s expressed i n terms of cI3 Consequently, and ea2 by eq. (6.7).

The r a t i o A : ( A + i s independent of Young(s modulus E, b u t depends only on Poissoncs constant v. Table 1 shows t h e p h y s i c a l c o n s t a n t s and parae have assumed t h a t t h e conmeters a f o r s t e e l , aluminum, and duralumin. W same, and t herefore the coefficients q k s t a n t s v f o r these materials a r e the shown i n Table 2 , a r e a l s o t h e same. f o r t h e s e materials, and ni, /x)4

a)

TABLE 1

2 3

STEEL 30KhGSA ALUhlINUh! AMG DURALUMIN D17

. .. . .

12700

0,805.10-2

1,54

2060 2350

0.396.I O ' 0,439 I O '

6.38
5,19

TABLE 2

I n t h e columns w i t h a (i)sign, t h e numerical values of t h e c o e f f i c i e n t s a r e given f o r t h e r e g i o n of tension. I n t h e region of compression these coe f f i c i e n t s have t h e opposite sign.

It will be seen from eqs.(6.12a)

i n t h e zone I n t h e zone a l s o occur, ;one, s i n c e

(6.1%) t h a t t h e q u a n t i t i e s A* and p* of t e n s i l e s t r e s s e s a r e always g r e a t e r , r e s p e c t i v e l y , than A and p . of compressive stresses t h i s i n c r e a s e of A* and p* over A and 1-1 may b u t t h e d i f f e r e n c e s A* A and W I.L a r e smaller than i n t h e f i r s t t h e q u a n t i t i e s i n odd powers of t h e parameter a a r e negative.

If t h e parameter a and t h e components of

a r e such t h a t t h e t r i n o m i a l s

i n t h e zone of compression a r e negative, t h e n i n this zone t h e d i f f e r e n c e s A*-A

and p*

- I.L may a l s o b e

negative.

I n any cass, r e l a t i o n s (6.12a) (6.12b) show t h a t t h e presence of nonl i n e a r terms among t h e s t r a i n t e n s o r components D , , strengthens t h e asymmetry i n t h e d i s t r i b u t i o n of s t r e s s e s over t h e thickness of t h e s h e l l , i n connection with t h e appearance of r a d i a l displacements w. I n t h e zone of tension, t h e s t r e s s e s i n c r e a s e more r a p i d l y than would follow from t h e l i n e a r theory, while i n t h e zone of compression they i n c r e a s e more slowly. I n t h i s connection, t h e stresses i n t h e t e n s i l e zone reach t h e y i e l d p o i n t e a r l i e r than i n t h e compressive zone, t h e m a t e r i a l of t h e s h e l l i n t h e t e n s i l e zone l o s e s i t s load-carrying capacity, and t h e a c t i v e load i s transmit',ed t o t h e m a t e r i a l i n t h e i n i t i a l zone of compression. O f course, i n t h i s case t h e r e i s a r e d i s t r i b u t i o n of s t r e s s e s , and t h e very concept of i n i t i a l zone of compression l o s e s i t s meaning. /205 To o b t a i n a q u a n t i t a t i v e evaluation of t h e e f f e c t of n o n l i n e a r i t y on t h e deformation process, l e t us use one of t h e s o l u t i o n s of t h e problem i n i t s l i n e a r postulation, given e a r l i e r (Bibl. 23d). Coxsider t h e case of t h e equilibrium of a tube f r e e l y r e s t i n g on t h e f a c e The function w(x) i n t h i s case i s defined i n contours of t n e middle surface*. t h e following manier + S F :

Here, R and G a r e , r e s p e c t i v e l y , t h e r a d i u s and length of th,e tube, and D and 3 a r e expressed by t h e formulas

where E i s Young's modulus, v Poisson*s constant, and E and v a r e connected The f o r c e w i t h t h e Lame/ constants by t h e r e l a t i o n s (11, 4.2a) - (11, 4.2b). compressin,; t h e tube i s denoted by T. Ne s h a l l now make a statement of subs t a n t i a l importance f o r what follows. The parameter 4 i n eq.(6.13) may a l s o be understood as a q u a n t i t y connec5ed with t h e l e n g t h of t h e t u b e by t h e r e l a t i o n

I,,

= kl,

where k i s a whole number. I n this case, t h e f u n c t i o n w(x) determined by eq. (6.13) will s a t i s f y t h e boundary conditions and t h e fundamental d i f f e r e n t i a l
.#.

Freely r e s t i n g i s considered by u s as equivalent t o hingedly r e s t i n g .

C f . [Bib1.23d,

eq. (2.3)l.

205

equation of t h e problem, s i n c e t h a t equation does not contain t h e parameter 4 Tnus, we have t h e r i g h t t o a t t r i b u t e t o t h e parameter 4, a d e f i n i t e (Bib1.23d). meaning predetermining t h e s o l u t i o n (6.13). W e shall g i v e t h i s p r e d e t e d n a t i on b elow. Tne c r i t i c a l value of T [ t h e upper c r i t i c a l value according t o convent i o n a l terminology (Bibl.!+, l o ) ] satisfies t h e equation

It can be e s t a b l i s h e d t h a t t h e minimum c r i t i c a l value T corresponds t o t h e following approadmate r e l a t i o n :

(6.16) Under condition (6.16), w e f i n d from eq.(6.15) t h e well-known formula

/206

''

= =

4E h2 . v3(l-vz) R

(6.17)

Using eq.(6.15) i o n f o r w(x):

and r e l a t i o n (6.16), w e f i n d t h e following approximate express-

(6.18)
The term r e t a i n e d here determines w(x) for a half-wave corresponding t o t h e p r i n c i p a l form of l o s s of s t a b i l i t y . Therefore, & i s here not t h e t o t a l This l e n g t h i s indeterminl e n g t h of t h e tube b u t t h e l e n g t h of a half-wave. a t e . For i t s determination we must use experimental d a t a . Making use of eq.(6.9), w e f i n d from eq.(6.18):

P u t t i n g v = 0.3,

we o b t a i n

! . ? 3 1

zz

- 0,727 --

I1

rr,,
T

1,28 cos 1

(Rh):

206

W e now a l s o give expressions f o r

I,.x,
-

(4,
1

and

(a, )tax :
(6.X)a)

I F31 ImaX= 0.727

I1

5r-1'

Equation (6.20a) determines t h e limits of v a r i a t i o n of (3.18b), w e have mine k andL+, according t o eqs.(3.18a)

,.

To deter-

b
(!!31)2 -

1-s 20
--b

1 d!!31 = - (!l3,)iaX;
-

1 -

-b

(6.2%)

(6.22b)

These equations permit only a r a t h e r rough e v a l u a t i o n of t h e e f f e c t s connected with t h e presence of nonlinear terms among t h e components of t h e f i n i t e deformation t e n s o r Dik

Consider now t h e p r i n c i p a l conclusion r e s u l t i n g from t h e above.

1 .

Ebaluation of t h e E f f e c t of t h e Component Sar on t h e Stressed S t a t e of

a S h e l l Depending on t h e Value of t h e R a t i o s Tcr : T and h: G

It i s c l e a r from eqs.(6.20b), (6.21), and (6.22a) (6.22%) t h a t t h e eff e c t of t h e nonlinear terms on t h e c o e f f i c i e n t s A* and p* s u b s t a n t i a l l y depends on t h e r a t i o s T,, : T and h: 4. Let us introduce t h e n o t a t i o n
(6.23a) Then,

As can be concluded f r om eqs.(6.22a) (6.22b), t h e e f f e c t of t h e nonline a r terms among t h e components of Dik decreases with decreasing r a t i o h: 4 and i n c r e a s e s as T approaches T c r .
To d i s c l o s e more d i s t i n c t l y t h e s i g n i f i c a n c e of t h e nonlinear terms, w e present below c e r t a i n numerical c a l c u l a t i o n s of t h e value of t h e q u a n t i t i e s Of entering i n t o eqs.(6.2%) (6.22b), based on t h e s o l u t i o n (6.18). /x)8 course, t h e r e s u l t s s o obtained can be regarded merely as rough and i n d i c a t i v e .

Taking as before, v = 0.3 and, consequently, Let us s e p a r a t e terms i n eqs.(6.22a) (6.22b). eqs.(6.23a) (6.23b)

p, l e t us consider t h e earing i n mind

Then, eqs.(6.22a)

(6.22d) can be represented as follows:

208

I n t h e r e l a t i o n (6.22a) (6.22b) we neglected term2 of r e l a t i v e order r e t a i n i n g terms of o r d e r s a-f(&l)Eax, and a-2(&l)tax and took i n t o consideration t h e values of t h e c o e f f i c i e n t s q k and q k given i n Table 2. A p o s i t i v e s i g n before terms w i t h i n parentheses corresponds, as above, t o a region of tension.

It i s c l e a r from eqs.(6.25a) (6.25b) and from Table 2 t h a t t h e normal s t r e s s e s i n t h e zone of t e n s i o n a r e g r e a t e r t h a n those determined by t h e l i n e a r theory, while i n t h e zone of compression they a r e correspondingly l e s s . It i s c l e a r t h a t t h e shearing s t r e s s e s i n t h e zone of dominating t e n s i l e s t r e s s e s , i n areas i n c l i n e d a t an angle of t o t h e g e n e r a t r i x , will be g r e a t e r than those

determined by t h e l i n e a r theory. To evaluate t h e degree of d e v i a t i o n of t h e nonlinear theory f r o m t h e l i n e a r , l e t us t u r n t o Tables 3 and 4 . These Tables give t h e v a l u e s of xi and yl and t h e i r corresponding values of w.

It will be seen f r o m Tables 3 and 4 and eqs.(6.25a) - (6.2%) t h a t , even a t small values of u), t h e reduced e l a s t i c constants %* and G+-- can d e v i a t e cons i d e r a b l y f r o m A and CL, b u t this, according t o eqs.(2.14),involves a d e v i a t i o n
of t h e f i e l d of s t r e s s e s from t h a t found by t h e l i n e a r theory.

/209

;le c a l l a t t e n t i o n , f o r i n s t a n c e , t o t h e f i g u r e s marked by an a s t e r i s k (*) Table 3 i n d i c a t e s that, f o r s t e e l , a t u , 3 0.09, i n t h e zone Df dominant t e n s i l e s t r e s s e s x t i s more than 3% g r e a t e r t h a n A , whereas i n t h e zone of dominant compressive s t r e s s e s i t i s more than 35% smaller. For duralumin, t h e s e changes i n x * a l r e a d y occur a t w = 0.07.

i n Tables 3 and 4 .

TABLE 3

STEEL JOKHGSA

-- - I
0.0265 0,0374 0,0530
0.0648

I
0,05 0.10 0.20

WRALUMIN DL7

0,05

0,lO
0,20 0.30

0,0900+ 0,107 0,127 0.141

0.05

0.10 0,20

...-

0,0664* 0,0790

0 . 3 0 0;40

0;0749 O@Ku+

0 . 0 9 3 9

0.05 0.10 0.20

0 . 3 0

0;40
0.50

0.0195 0.0276 0.0391 0,0478 0,0552 U,0618+

209

TABLE 4

I - I
0.05 0.10 0,20

STEEL 30KHGSA
. Y,
0

Q l

__.__

0,40 0.30

430

0,0914 0.109 0,1292 0,143* 0,154 0,162

0.05 0.10 0,20 0.30 0.40

O.i42*
0;201 0.284 0,348 0,402 0,449

0.05

0,lO
0,20

0,30
0.40

0 . 3 0

0 . 5 0

0.0674 0,0802 0,0953 0,106* 0.1 13 0,120

0.05 0,i0 0.20

0,105' 0.148

430
0.40 0.50

0,209 0.256 0;2!36 t?#331

The reduced e l a s t i c constant l i k e w i s e d i f f e r s appreciably from p a t r e l a t i v e l y small values of w. Thus, f o r example, it Will be c l e a r from Table 4 t h a t t h e value 10 f o r s t e e l , a t w E 0.14, i s about 35% g r e a t e r than p, i n t h e zone of dominant t e n s i l e s t r e s s e s , and 25% i n t h e zone of compressive occur already a t w = 0.11. s t r e s s e s . For duralumin, t h e s e charges of

z9

W e s h a l l now determine whether t h e values of w given above a r e possible., I n t h e case under consideration, t h e i r existence i s ensured by t h e factor

(% - 1 ) '

on t h e right-hand s i d e of eq.(6.23a).
h

Table 5 g i v e s t h e

values of w,

t h e ratios i and t h e corresponding values of t h e r a t i o T : T c r .

.e

TABLE 5

0,06
0,08

0.02 0.04

0,750

0.600
0.500 0.800

0,06
0.02 0.04 0.06 0.02 0.04

0.10
0,12

0 . 0 6
Oi857 0,750 0,667

Since t h e value of t h e ratio h : 4 i s unknown, l e t us t u r n t o t h e ex210

/ a 0

perimental d a t a given i n t h e Vol fmirmonograph (Bibl.4). W e r e l i e d on t h e s o l u t i o n of t h e axisymmetric problem. However, as shown by experiments, t h e shapes i n which t h i n s h e l l s buckle a r e not adsymmetric. Two o r t h r e e systems of depressions a r e formed, which can s t i l l be considered as a c e r t a i n equivalent of t h e systems of half-waves of t h e &symmetric form of deformation. Using Figs.8.2 and 8.3 of t h e V o l f m i rmonograph (Bibl.k), we can evaluate the ratio 4 , : R i f we assume, a s i n d i c a t e d above, t h a t t h e parameter 4 i s t h e T h i s evaluation shows t h a t t h e r a t i o 4 : R f o r specil e n g t h of a half-wave. mens shown i n Figs.8.2 and 8.3 (Bibl.4) can have values of 0.15 0 . 2 .

Now, from eq.(6.16),

f o r n = 0, i . e . ,

f o r one half-wave,

w e find (6.26a)

For v

= 0.3,

w e obtain

h
--

z=

I
>dl (0.15;

0,168 -

(6.26b)

Hence, w e f i n d t h a t t o t h e v a r i a t i o n of t h e r a t i o 4, : R over th5 i n t e r 0.20) t h e r e corresponds a v a r i a t i o n of t h e r a t i o h : G over t h e i n T h i s c a l c u l a t i o n confirms t h e a d v i s a b i l i t y of s e l e c t t e r v a l (0.025; 0.034). i n g t h e v a r i a t i o n a l i n t e r v a l of t h e r a t i o h : G given i n Table 5. Let us now f i n d t h e r a t i o 2h : R. From t h e data presented here i t follows t h a t t h i s r a t i o , f o r t h e specimens shown i n Figs.8.2 and 8.3 of t h e Vo1"ir vonograph,

v a r i e s over t h e i n t e r v a l

(ik! $j)

The determination of t n e r a d 0 2h : R

i s presented for v e r i f i c a t i o n . Judging f r o m t h e content of t h e monograph (Bibl.L), t h e values of t h e obtained r a t i o 2h : R approximately correspond t o t h e geometrical c h a r a c t e r i s t i c s of t h e t e s t specimens, s i n c e t h e r a t i o R : 2h= = 1 0 0 - 180 i n t h e experiments described i n t h a t study. Of course, bearing eq.(6.26a), i n mind, t h e a n a l y s i s might be conducted without f i r s t having recourse t o experimental d a t a .

From eq.(6.26a),

i t follows t h a t

/211
(6.26~)

L e t , f o r example, 2h : R = 1 : 100.

Then,
h :I

I : R = 0,165;

= 0,029.

This corresponds t o t h e v a r i a t i o n a l i n t e r v a l of t h e r a t i o s 4 : R and h : 4 found from t h e experiments given i n t h e above monograph (Bibl.4).
To summarize we may say that our assignment of eqs.(6.13) i s s u f f i c i e n t l y motivated.
t o one half-wave

Returning t o t h e qu_estion - of t h e connection between t h e values of t h e reduced e l a s t i c constants A*, p z and t h e r a t i o T : T , , , we f i n d from Table 5 t h a t t h e differences between A* and and A, p, which reach 30% f 5% i n ab, , over t h e i n t e r s o l u t e value occur during t h e v a r i a t i o n of t h e r a t i o T : T I n t h i s case, t h e constants %++ and found for duralumin v a l (0.75; 0.85)*. a r e more s e n s i t i v e t o t h e v a r i a t i o n s of t h e r a t i o T : T,, than t h e s e same cons t a n t s given f o r s t e e l , To go deeper i n t o t h e meaning of t h e s e conclusions, l e t us consider the s i m p l i f i e d expression f o r w(x) d i f f e r i n g from eq.(6.18), and l e t us d r a w several supplementary conclusions.

TI*

c*

If we again use eq.(6.15) t o determine T,, Let us r e t u r n t o eq.(6.13). and h e r e a f t e r take T,, t o mean i t s minimum value expressed by eq.(6.17), then from eq.(6.13) we can f i n d :

Howeve * , (6.28)

Consequently,

w Gz - __ v ( Tcr 7 - 1) -1
2R

x (2-x).

It should be noted t h a t eq.(6.29) i s approximate, since i n s t e a d of t h e roots of eq.(6.15), which a r e functions of n, we introduced i n t o eq.(6.13) only t h e m i n i " value of t h e root, which was independent of n. Moreover, /212 eq.(6.29) does not satisfy a l l t h e boundary conditions of t h e problem, although t h e expressions (6.27) obtained from t h i s r e l a t i o n do s a t i s f y t h e boundary cond i t i o n s . This f a c t i s connected with t h e well-known p r o p e r t i e s of t h e expans i o n s of functions i n Fourier s e r i e s .
A d i r e c t comparison of t h e r e l a t i v e accuracy of eqs.(6.18 and

(6.29) i s

* This interval.

i s s t a t e d as a rough approximation.
Definite I n t e g r a l s and Fourier Series,

wt Cf., f o r instance, L.V.Kantorovich, Leningrad University, 1940

212

Equation (6.29) h a s t h e advand i f f i c u l t and r e q u i r e s s p e c i a l i n v e s t i g a t i o n . It a l s o r e f l e c t s t h e influence of t h e terms tage of s i m p l i c i t y over eq.(6.18). o f t h e s e r i e s t h a t a r e r e j e c t e d i n deriving eq.(6.18). Obviously a r e l a t i v e l y small e r r o r i n eqs.(6.18) and (6.29) may have a s u b s t a n t i a l e f f e c t on t h e r e s u l t s , s i n c e one must operate w i t h t h e f o u r t h and second powers of t h e component 7b1. Since t h e paraLet us now d i s c u s s t h e p h y s i c a l meaning of eq.(6.29). meter n d i d not s n t e r i n eq.(6.29), t h e q u a n t i t y & no longer has any d e f i n i t e meaning i n t h i s equation. Here, 4 may be taken t o mean a segment, varying from t h e l e n g t h of a half-wave determined by t h e sinusoid according t o e assume, as before, t h a t 4 , eq.(6.18), up t o t h e e n t i r e l e n g t h of t h e tube. W i s t h e l e n g t h of a half-wave. The p r o p e r t i e s of Fourier s e r i e s permit eq. (6.29) t o be d i f f e r e n t i a t e d twice, and t h e r e s u l t a n t d e r i v a t i v e W i l l have meaning everywhere over t h e unclosed i n t e r v a l ( 0 , 4). Consequently, s e t t i n g v = 0.3, we f i n d

where

and f u r t h e r

Instead of eqs.(6.&a)

(6.24d), we obtain

Equations (6.22a)

(6.22b) now t a k e t h e following form:

The d a t a given i n Tables 3 and 4 permit us t o f i n d R f o r a s p e c i f i e d Si and T i r . For t h i s , i t i s sufficient-to m u l t i p l y t h e values of w by a WransiT h i s f a c t o r can a l s o be f o w d f o r t i o n f a c t o r " equal t o 0.727 : 0.15 = 4.85. a n a r b i t r a r y value of Poisson's constant v , i f we compare eqs.(6.19a! and (6.30). W e have

instead of eq. (6.26a).

A comparison of t h e "exact" r e l a t i o n (6.26a) with t h e approximate r e l a t i o n ( b ) shows t h a t t h e use of eqs.(6.19a) and (6.30) involves a considerable e r r o r i n t h e r e s u l t s , a n e r r o r of t h e order of 20%. This undoubtedly i s due t o t h e f a c t t h a t eqs.(6.19a) and (6.30) were obtained by d i f f e r e n t i a t i o n of t h e approximate expressions w( x)
If we use t h e same t r a n s i t i o n f a c t o r from w t o i 2 and from h : 4 t o 4 : R, , , d e t e d n e d by eq.(6.26a) o r r e l a t i o n (b), then t h e values of t h e r a t i o T : T will be independent of t h e cnoice of t h e t r a n s i t i o n f a c t o r , s i n c e i n t h i s case t h e following equation W i l l be t r u e :

T :T , ,= W : (fW

+3)
-

=9

:( Q

+i ) .

But s i n c e t h e a r a c y of e q s . ( 6 . 3 ) (6.30) may be g r e a t e r than t h e (6.19a), we s h a l l consider t h e conclusions obtained accuracy of eqs.(6.1 from eqs.(6.30) (6 ,b) independently of t h e conclusions obtained from eq.(6.18) and i t s conse'quences. For t h i s purpose, l e t us make use of Tables 3 and 4 w i t h t h e t r a n s i t i o n f a c t o r ( b ) of 4.85 f o r v = 0.3 and consider t h e

quantity

4 as -

a n independent parameter, which i s equivalent t o adopting t h e

r e l a t i o n (6.26a) o r (6.26b).

It i s obvious t h a t , under t h e s e conditions, t h e reduced constants A* and F* will vary more r a p i d l y as T approaches T,, than they would according t o t h e c a l c u l a t i o n s based on eq.(6.18). I n f a c t , a t values of t h e r a t i o 4 : R = = 0.15 - 0.20, t h e changes i n t h e reduced constant E* r e l a t i v e t o p , which i n absolute value reach 30% f 5$, occur while t h e r a t i o T,, : T varies over t h e I n t h i s case, h+c v a r i e s more r a p i d l y than;*. i n t e r v a l (0.67, 0.79).

11111111111mImllll

I1 11111.111l11

II 111 I

I I I

11111111

I IIII 11111IIIIII I1111111111

-'

.
C r i t e r i o n of i n s t a b i l i t x . W e mentioned above t h a t , i n t h e zone of 424 dominant compressive stresses, t h e parameter A-S decreases and E* i n c r e a s e s . * i n t h e zone of This p o i n t s t o a decrease i n t h e reduced Poisson constant T dominant compressive s t r e s s e s . While decreasing, t h e constant <+ may become equal t o zero and then become negative. Above, we d i d not consider t h e meani n g of t h e change i n s i g n of t h e reduced e l a s t i c constants. 'Let us f i l l t h i s gap somewhat.

It i s easy t o prove t h a t t h e s p e c i f i c work of deformation W i l l have a p o s i t i v e d e f i n i t e q u a d r a t i c form i f t h e Poisson constant varies over t h e
interval
(-1,

+).

Here, as i s w e l l known, t h e proof of t h e Kirchhoff theo-

r e m i s v a l i d and t h e s o l u t i o n of t h e boundary problems of t h e s t a t i c s of l i n early deformed b o d i e s i s unique, i . e . , t h e s t a t e of equilibrium of such bod i e s i s s t a b l e . Yowever, no negative v a l u e s of t h e Poisson constant have been found i n a c t u a l i s o t r o p i c bodies. Evidently, t h i s experimental f a c t i s not random b u t depends on t h e a c t u a l p r o p e r t i e s of matter which a r e n o t r e f l e c t e d by t h e s i m p l i f i e d scheme of t h e continuous medium and, consequently, a l s o not by t h e a n a l y t i c s t r u c t u r e of t h s s p e c i f i c energy of deformation.
Most a u t h o r s assume t h a t t h e Poisson constant i s always p o s i t i v e . It w a s s t a t e d by E.Trefftz, erroneously, t h a t t h e p o s i t i v e n a t u r e of Poisson's cons t a n t results f r o n t h e requirement of t h e p o s i t i v e determinacy of t h e s p e c i f i c energy of deformation-:+. Let us assume, a t f i r s t , according t o experiments, t h a t i n a r e a l i s o - t r o p i c body t h e Poisson constant i s a l w a y s p o s i t i v e and t h a t t h e c r i t e r i o n of i n s t a b i l i t y % + + i s t h e change of s i g n of t h e reduced Poisson constant V-3. Let us f i n d t h e m i n i m u m value of t h e compressive f o r c e T a t which t h e Poisson constant, i n t h e zone of dominant compressive stresses, becomes negative. Equating X-X t o zero, we f i n d , according t o eq.(6.25),

I XI I+ x ,

3 -- =o. 2

(6.35)

Making use of eqs.(6.24.a) - (6.24.b) and t h e d a t a given i n Table 2 f o r duralumin, we o b t a i n a? equation b i q u a d r a t i c i n u. The smallest p o s i t i v e r e a l root of t h i s equation i s
-

wc-= - 0,098.

(6.36)

It will be seen from Table 5 t h a t t o t h i s value of w,, and t o t h e ra',io h : 4, = 0.03 corresponds t h e following value of t h e r a t i o T : T,, :

/215

: E .Trefftz,
% +

Mathematical Theory of E l a s t i c i t y ONTI, 1934, pp.39-40.

The use of t h i s i n s t a b i l i t y c r i t e r i o n i s not mandatory. more n a t u r a l approach i s p o s s i b l e . See below, eqs.(6.63a)

A d i f f e r e n t and

(6.63b) etc..

215

where T= i s t h e hew c r i t i c a l value of t h e compressive f o r c e T. and g r e a t e r r o o t of t h e equation

The second

I x, 1 --x,

3 - -=o, 2

e s t a b l i s h e d f o r t h e zone of dominant t e n s i l e s t r e s s e s , i s

To t h i s r o o t , a t t h e r a t i o h : 4 = O.G3, t h e r e corresponds a value of t h e r a t i o T : T,, close t o u n i t y . The first r o o t carresponds t o t h e a r i t h m e t i c mean value of t h e c y i t i c a l load, and t h e second t o t h e upper value. According t o d a t a given i n theVo1"irmonograph (Bibl.4), t h e experimental r a t i o T : T,, ranges from 0.384 t o 0.700.
Thus, t h e a r i t h m e t i c mean of the c r i t i c a l l o a d found by us approaches t h e upper limit of t h e experimental data. U p t o now w e have been r e l y i n g on cq.(6.35), derived from t h e approximate expression (6.18). Let us now make use of t h e equation

It, I
r e s u l t i n g from eq.(6.3/+a). result, w e obtain

+E*

-2 =0,

(6.40)

Will be obtained from t h e s o l u t i o n (6.36) a f t e r multiplying i t by 4.85.

llCr - =0,475.
Assuming t h a t t h e r a t i o 4 : R r 0.18, we f i n d t h a t

Obviously, t h e required s o l u t i o n of t h i s equation As a

Consequently, here too t h e c r i t i c a l value of t h e l o a d i s outside t h e v a r i a t i o n a l region of t h e experimental data.

T w o p o s s i b l e causes of t h e c o n t r a d i c t i o n between t h e r e s u l t s obtained by us f o r t h e approximate determination of t h e c r i t i c a l value of t h e load and t h e experimental d a t a could be given.

(Fel ) ,

Tne f i r s t i s t h a t we based our c a l c u l a t i o n s on t h e averaged value of , i n s t e a d of on (Sal ,)

The second cause of t h e u n s a t i s f a c t o r y r e s u l t s of t h e t h e o r e t i c a l invest i g a t i o n s undertaken by us l i e s i n t h e choice of t h e parameter a . W e have /216 already noted repeatedly t h a t t h e proper s e l e c t i o n of t h e region of approximat i o n s i s of s u b s t a n t i a l importance. It i s t h i s choice t h a t determines t h e vale ue of the constant G, e n t e r i n g i n t o t h e formulas determining A* and !A*. W have assumed t h a t t h e s t r e s s e s i n t h e s h e l l reach t h e y i e l d point, and from -these conditions we have d e t e r s n e d t h e value of a . But f o r s u f f i c i e n t l y l o n g and t h i n s h e l l s , t h e l o s s of s t a b i l i t y may occur before t h e s t r e s s e s reach t h e y i e l d point. Therefore, t h e above conclusions a r e t r u e only f o r s u f f i c i e n t l y <hick and s h o . r t s h e l l s , i n which t h e r e a r e no l o s s e s of s t a b i l i t y under s t r e s s e s approaching t h e y i e l d p o i n t .

For o t h e r shells, one must decrease t h e values of t h e parameter U t o below those shown i n Table l. This l e a d s t o an increased i n f l u e n c e of t h e nonlinear term. But t h e d i f f i c u l t y here c o n s i s t s i n t h e determination of

a.

Let us f i n d t h e value of t h e r a t i o T,, : T,, s t a r t i n g from t h e unaveraged values of A+ and p++ corresponding t o t h e Ftudy of l o c a l i n s t a b i l i t y i n a medium approximately equivalent t o a s h e l l . Instead of eqs.(6.2f+a)

( 6 . a d ) we w i l l then have

= 5x, -- 0 , 2 7 3 ~3~ m,,

(- + n l o )
2

w4,

GI =5y, = 0 , 2 7 3 (~3~m,,+ n,,)

w4,

y, = 3y2= 0 , 5 2 3 ~ - * 3 nz21-k i t z l ) w3.

(-2

Further, we f i n d

q2

3 - I x1,G,),

Equating

A-3

t o zero, we o b t a i n a n equation biquadratic i n

u ) :

217

3 I XI I TX, -=0.

--

(6.45)

The smallest p o s i t i v e r e a l r o o t of t h i s equation, determined f o r duralumin, i s as follows:

wcr= -

0,058.

( 6 -46)

, A d i r e c t calculatiol?, t o g e t h e r with Table 5, shows t h a t t h i s value of w and t h e r a t i o h : 4 , = 0.03 correspond t o

-m

According t o t h e experiments described i n t n e Vo1"ir monograph (Bibl.4), t h e mean v a l u e of t h e r a t i o T : To, i s --.0.61-%. The v a l u e of T : Tc, found by xs l i e s roughly a t t h e c e n t e r of t h e i n t e r v a l between t h e experimental mean value of T : T , , and t h e upper boundary of t h e experimental data, which i s 0.7. However, we derived eq.(6.4?) by r a k i n g use of t h e sequels of eq.(6.18). Let us consider t h e conclusi.ms based on eq.(6.29). By analogy with eq.(6.41), we f i n d

Assuming as above t h a t t h e r a t i o 4 : R = 0.18, we have

Tcr - : T,, r 0,59.

(6.49)

This value i s somew5at l e s s t h a n t h e mean experimental v a l u e given by V o l f m i r (Bibl.)!), wnich i s 0.61. For 30KhGSA s t e e l [eq.(6.48)], we g e t

Accordingly,

36

The value of Volrmir's dimensionless parameter p (Bibl.4) corresponding t o T : Tc, = 0.66 i s 0.396. The mean value according t o experiments i s 0.364.

a 8

It will b e c l e a r from t h i s t h a t a c a l c u l a t i o n , w i n g t h e maximum values of t h e components of RS1, y i e l d s s a t i s f a c t o r y r e s u l t s .

Now l e t us consider t h e parameter %. A s above, we s h a l l confine ours e l v e s t o an approximate d e t e r m i n a t i m of t h i s q u a n t i t y .
W e s h a l l b a s e our determination of t h e parameter a on eq.(6.29), which permits us t o f i n d t h e madmum bending s t r e s s e s i n t5e s h e l l . This makes i t p o s s i b l e t o f i n d t h e v a r i a t i o n a l limits of t h e s t r a i n t e n s o r components o r t h e region of l i n e a r approximation of +,he components of t h e finite-deformation tensor or, i n o t h e r words, t h e parameter 3.

Double d i f f e r e n t i a t i o n of khe equivalent equations (6.27) and (6.29) i s permissible, s i n c e t h e trigonometric s e r i e s obtained on d i f f e r e n t i a t i o n of eq. (6.27)converges t o t h e d e r i v a t i v e of t h e right-hand s i d e of eq. (6.29). S t a r t i z q from eq.(6.29), we f i n d from our paper (Bib1.23d) t h e m a x i m u m absoluLe value of t h e normal s t r e s s due , t o f l e x u r e and campression:

I Jmax I =

T
2a

Ev

-3

It

- I)-'.

(6.52) Let
US

Hence, we f i n d t h e new value of t h e parameter U . value a+*. Making use of eq.(6.17), we f i n d

call this

Equation (6.53a) can be put i n t o t!le following form:

Further, b e a r i n g i n mind eqs.(6.30), (6.32) and considering t h e unaveraged values of A+: and CL*, we f i r d , by analogy t o eqs.(6.33a) - (6.33b), t h e r e l a tions

It goes without saying t h a t t h e s e r e l a t i o n s are t r u e o n l y f o r v

s i n c e we have already assumed t h a t h =

0.3,

T,qer e t a i n e d here t h e l i t e r a l 2 designation v o n l y t o make t h e transformation more d i s t i n c t . W e have f u r t h e r

2 &.

The equation f o r determining t h e c r i t i c a l values of form :

:i

here t a k e s t h e following

I 1 / + E ?

. _

--

=o.

(6.56)

This i s a n equation of t h e f o u r t h degree i n 3. L e t u s denote i t s smallest 'de f i n d , without solvin,? eq.(6.56), t h e approximate posi5ive r o o t by DSr. value of ; k r , makizg use of t h e previous c a l c u l a t i o n s . Now, comparing eqs.(6.%a) - (6.54b) with eqs.(6.33a) - (6.33b) and /a9 b e a r i n g i n mind t h a t t h e d i f f e r e n c e i n t h e numerical f a c t o r s i s connected w i t h the f a c t t h a t eqs.(6.5/+a) (6.54b) y i e l d t h e v a l u e s of A%- and ~ 9 , correspondi n s t e a d of t h e mean V a l in,.: t o t h e maximum v a l u e s of t h e components of ues, A%- and Go, we g e t

This equation can be p u t i n t o t h e following form:

where

220

Consequently, the required v a l u e of i & r is

(6.60)
while t h e corresponding v a l u e of t h e r a t i o T : Tc,

is
(6.61)

Thus, t h e r a t i o T,, : T,, does not depend e x p l i c i t l y on t h e r a t i o h : R. : T,, on h : R i s connected with t h e The i m p l i c i t dependencF of t h e r a t i o Tu presence of t h e q u a n t i t y L f L i n eq.(6.58). This r e s u l t i s a p p a r e n t l y confirmed by t h e experimentally establis'rled f a c t of t h e weak dependence of t h e r a t i o Tu : T , , on t h e r a t i o h : RX-. Equation (6.61) does not confirm i n a n e x p l i c i t form tile e s t a b l i s h e d experimental tendency of t h e r a t i o T,, - :T , , to decrease w i t h decreasing r a t i o

3%. R

Consider now t i e numerical value of t h e r a t i o T,, :T , , for duralumin, /220 p u t t i n g v = 0.3; a-1 = 2.28 x 10";A = = 0.257 [ t h e v a l u e of i& i s taken from eq. (6.14.8) 1. W e have
Tcr - : 7'cr
E Z

0,654.

(6.62)

Consequently, t h e methods of determining t h e parameter a i n c r e a s e t h e d e v i a t i o n of t h e r a t i o T,, : T,, from t h e mean experimental value, as compared with t h e previous method, which l e d t o eq.(6.49). W e note a l s o t h a t eqs.(6.57) (6.58) can be used a t small values of t h e r a t i o h : R, i . e . , f o r v e r y t h i n s h e l l s , s i n c e a l r e a d y a t 2h : R = 0.01 t h e normal stress l o m a x l c a l c u l a t e d from eq.(6.52) exceeds t h e y i e l d p o i n t for duralumin D17.
g -

Cf.

(Bibl./+, p.322).

% + Ibid.

W e have used v a r i o u s methods, based on t h s g e n e r a l method of l i n e a r i z a t i o n , i n i n v e s t i s a t i n g t h e values of t h e r a t i o To, : T,, corresponding . t o t h e vanishing of t h e Poisson constant v+: of a l i n e a r T y deformable medium, approxi m a t e l y e q u i v a l e n t t o a m e d i u m w i t h f i n i t e deformations.

Despite t h e d i f f e r e n c e i n t h e a n a l y t i c expressions of t h e r a d i a l displacement w used by us and d e s p i t e t h e employment of b o t h averaged and maximum a b s o l u t e values f o r t h e component Sol, t h e values of t h e r a t i o T,, : T , , were found t o b e r a t h e r s t a b l e under v a r i a t i o n s of t h e i r determination'methods. This shows t h a t t h e conclusions obtained here cannot b e due t o random agreement of t h e numerical results. Rather, t h i s must b e a r e f l e c t i o n of a c t u a l processes t a k i n g p l a c e on any loss o f s t a b i l i t y of t h e s h e l l . L e t u s now r e t u r n t o determination of t h e r a t i o T,, : T c p . Above, we have used t h e change i n s i g n of v+c o r & o d y as a c r i F e r i o n of i n s t a b i l i t y .

A d i f f e r e n t approach t o determination of t h e r a t i o T,, : T,, i s p o s s i b l e , based on t h e use of eq.(6.17), i n which t h e e l a s t i c constants E and v must b e substituted by E + : and v9, corresponding t o o r %- and v s found from Using eqs. (6.17) and (6.23a), we o btain !,:?e average v a l u e s of

I.

Imax,

and, on using t h e averaged v a l u e s of

]ralII.,

I f we express E* and v+ i n t e r n s of A% and L*, and use eqs.(6.!+lca)-(6.&L+b) t h e n eq.(6.63a) i s transformed i n t o a n equation a l g e b r a i c i n LC. A s i m i l a r /221 equation C a r l be obtained from eq.(6.63b).

To f i n d t h e smallest r e a l p o s i t i v e r o o t s of t h e s e equations, corresponding - : T,,, we employed a g r a p h i c a l method. t o t h e minimum values of t h e r a t i o T,, Let US p u t

assuming v = 0.3;

0.03.

Let u s i n t r o d u c e t h e coordinates (&, f c ) ( s e e graph).

222

978
968

45
'(13

002

404

4064 408

9106 412 W

m o t t i n g on t h e plane (a, f c ) o r (E, fc) t h e f u n c t i o n s f(w), +(u) and ~ ( u J )we , f i n d t h e r o o t s sought. It Will b e seen from t h e graph t h a t t h e value of T : Tcr found fbom eq.(6.63a) i s 0.68, while t h a t c a l c u l a t e d from eq.(6.63b) 1 s 0.78. Tne graph a l s o snows t h a t t h e values of v* and v*, corresponding t o t h e 1 , c r i t i c a l values of t i e load, l i s i n t h e i n t e r v a l (-1, 0). For W equal t o t h e q u a n t i t y E* vanisnes, which e v i d e n t l y corresponds t o a complete l o s s of = 1 , similarly vanishes. t h e load-carrying capacity of t h e s h e l l . A t

T *

z *

Thus t h e region of i n s t a b i l i t y corresponds t o t h e v a r i a t i o n of v* andT* m e r t n e i n t e r v a l (-1, 0). These regions a r e shown by hatc'ning i n t h e graph. W e note Turther t h a t tile p o i n t s of t h e planes (u), f@) o r (u), f c ) cor-, /222 , , always l i e i n t h e i n responding t o t h e c r i t i c a l v a l u s s of tine r a t i o T : T s t a b i l i t y regions between t h e t h e o r e t i c a l upper values and t h e t h e o r e t i c a l lowe r value of this r a t i o .

If we use eq.(6.63a), t h e n t h e value of Tcr : Tcr = 0.68 found by US i s very c l o s e t o t h e a r i t h m e t i c mean of t h e u p p e r a n d lower values of t h i s r a t i o found by t h e t h e o r e t i c a l e n e r g e t i c method. The a r i t h m e t i c mean of t h e upper A l l this confirms t h e and lower c r i t i c a l values of t h e r a t i o T : T,, i s 0.65. expedience of using t h e method we have considered i n t h e problems of s t a b i l i t y of a s h e l l . Tnis method permits a n appro>xGnate determination of t h e mean values of t h e c r i t i c a l load i f we f i n d i t s upper value from t h e s o l u t i o n of t h e problem i n l i n e a r formulation. It i s c l e a r t3at t h e same method makes i t a l s o p o s s i b l e t o evaluate t h e lower c r i t i c a l value of t h e load. For example, find, , : T,, equals 0.68, we determine t h e h w e r c r i t i c a l value of this ing that T r a t i o as 0 3 8 (100 0.68) = 0.36. I n t h e Voltmir monograph (Bibl.4) it i s shown t h a t t h e lower c r i t i c a l value of t h e r a t i o T : Tcr i s about 0.33.

W ' e s'lall not consider t h e r e s u l t s of t h e a p p l i c a t i o n of eq.(6.29) ale can here o b t a i n a though, as w i l l be seen from t h e preceding discussion, w

223

value of t h e r a t i o T : T,, menbal value. Section 7.

s t i l l c l o s e r t o t h e mean t h e o r e t i c a l and experi-

Brief Conclusions

Tne contents of t h e preceding Section confirm t h e usefulness of applying t h e metLlod of l i n e a r i z a t i o n of t h e components of t h e finite-deformation tensor developed by us a t t h e beginning of t h i s Chapter. The d i f f e r e n c e between t h i s method and other approximate methods of solving nonlinear problems, f o r example t h e method of Galerkin and R i t z , i s t h a t w e did not s t a r t out from t h e content of s p e c i a l problems suggesting t h e form of t h e approximation function, but t r i e d t o derive t h e general p r i n c i p l e s of construction of t h e approximate s o l u t i o n s u i t a b l e f o r extensive c l a s s e s of boundary problems. S t r i c t l y speaking, our method i s s u i t a b l e f o r t h e approximate s o l u t i o n of a v nonlinear boundary problem of s h e l l theory. This method, i n i t s concept, i s close t o t h e well-known methods of a n a l y t i c s y n t h e s i s of mechanisms according t o P.L.Chebyshev and was t h e r e f o r e included by us i n t h e p r i n c i p l e s of t h e ana l y t i c mechanics of s h e l l s .

We s h a l l make concluding remarks on t h e c o r r e l a t i o n between t h e consequences of t h e method of l i n e a r approximation of t h e components of t h e f i n i t e deformation tensor and t h e theory of s t a b i l i t y of s h e l l s .

1 . O n t h e Mechanism of Development of a Local Equilibrium and Motion of

I n s t a b i l i t y of a S h e l l

L 2 a

A study of t h e p r o p e r t i e s of a continuous medium w i t h small deformations, approximatelgT replacing a s h e l l with f i n i t e deformations, y i e l d s a preliminarg i d e a as t o t h e course of t h e development of a l o c a l equilibrium or motion i n s t a b i l i t y of a s h e l l .
When t h e load approaches t h e upper c r i t i c a l value, t h e r e i s a n i n c r e a s e of t h e asymmetry of s t r e s s d i s t r i b u t i o n over t h e thickness of t h e s h e l l . I n t h e zone of dominant t e n s i l e s t r e s s e s , t h e y i e l d p o i n t i s reached considerably before t h e load i x r e a s e s t o t h e upper c r i t i c a l l i m i t . After loss of t h e load-carrying capacity, by t h e m a t e r i a l i n t h e zone of tension, t h e l o a d i s transmitted t o t h e m a t e r i a l i n t h e zone of dominant compressive s t r e s s es, i f t h e m a t e r i a l i n t h a t zone i s i n t h e s t a b l e s t a t e . On t h e b a s i s of t h e p r o p e r t i e s of t h e medium approximately equivalent t o t h e m a t e r i a l of t h e s h e l l , i t can only b e s a i d t h a t , i n t h e zone of compressi v e s t r e s s e s , processes t a k e place t h a t approadmate t h e s t a t e of t h i s s u b s t i tuted medium t o t h e unstable s t a t e . T h i s i s characterized by a change i n s i g n of Poisson's constant v+:- and i t s passage through zero. The i n s t a b i l i t y of t h e medium i s accompanied by a motion of i t s elements, r e p r e s e n t i n s t h e elements of t h e s h e l l , i n accordance with t h e Gauss p r i n c i p l e of l e a s t constraint, or w i t h t h e Le Chatelier-Brown principles-. ?:-The Le Chatelier-Brown p r i n c i p l e i s as follows: If any stress or f o r c e i s brouzht t o bear upon a system i n equilibrium, t h e equilibrium is (contfd)

224

These p r i n c i p l e s a r e i n p a r t i c u l a r d i r e c t l y connected w i t h t h e methods of studying t h e theory of s t a b i l i t y of s h e l l s r e c e n t l y proposed by t h e author A.V Pogorelov*.

2 . The Role of Random Imperfections cf Shape

The l o s s e s of s t a b i l i t y of a s h e l l c o n s t i t u t e a nonstationary dynamic wave process which o r i g i n a l l y a r i s e s as a r e s u l t of v a r i o u s random sources, even i f t h e load i s far from t h e upper c r i t i c a l value and t h e r e are noie of t h e f l e x u r a l deformations considered by us i n t h e preceding Subsection>s-. /;?4 These random f a c t o r s i n c l u d e t h e i n i t i a l imperfections ef shape, which m a y be i n t e r p r e t e d as t h e e x i s t e n c e of i n i t i a l f i n i t e displacements of p o i n t s of t h e middle surface.

'

To t h i s i n i t i a l displacement correspond components of t h e antisymmetric t e n s o r i % k . Returning t o t h e above axisymmetric problem, l e t us assume that t h e d e v i a t i o n from t h e c y l i n d r i c a l shape of t h e b a s i c s u r f a c e i s determined by t h e f u n c t i o n w O ( ~ ) . To t h i s f u n c t i o n corresponds a f i n i t e component A s w i l l b e seen from eqs.(6.22a) (6.22b), t h e e x i s t e n c e of t h e component l e a d s t o a g r e a t e r v a r i a t i o n of t h e f i e l d of s t r e s s e s u d e r 'oad-

a ~(r&l)o )~.
nil)

i n g of t h e s h e l l even if t h e terms
+

5 2 (i = 1 , 2) a r e of t h e order of a" , i.e.,

1 (2

qo

n1o)

(G, 1 : ;

and

3 2 very s m a l l i n magari-

1 (2q ? +

tude

These q u a n t i t i e s m a y e x e r t a g r e a t i n f l u e n c e on A* and A-% and, consea y decrease t h e value of t h e c r i t i c a l l o a d below t h a t found above. quently, m

3. Regions of S t a t i c ' I n s t a b i l i t y
I f ,we make use of eq.(6.19a) and s u b s t i t u t e Tbl i n t o eqs.(b.lza) (6.12b), then XX- and p* Will be p e r i o d i c f u n c t i o n s of t h e x-coordinate. To o b t a i n t h e next approximation, l e t us use eqs.(2.14), t o f i n d t h e s t r e s s t e n s o r compone n t s . Then, applying one of t h e methods of reduction, we shall-obtain the system of equations of equilibrium of a c y l i n d r i c a l s h e l l , b u t this system of l i n e a r equations, as a l r e a d y noted i n Sect.5, will have v a r i a b l e c o e f f i c i e n t s

- in

t h i s case p e r i o d i c c o e f f i c i e n t s -which depend on t h e parameter

-(% l)-I
,.

( c o n t f d ) d i s p l a c e d i n a d i r e c t i o n which tends t o d i r h x k h t h e i n t e n s i t y of t h e s t r e s s o r f o r c e . (Cf. L.Landau and Ye.Lifshits, S t a t i s t i c a l Pnysics, Gostekhizdat, 19/&0)

+ Cf. A.V.Pogorelov,
,
. /

, .

'

Contribution t o t h e Theory of E l a s t i c S h e l l s i n t h e T r a n s c r i t i c a l Stage, Kharkov University, 1960. The i s o m e t r i c deformations of t h e middle s u r f a c e of t h e s h e l l on which his theory i s based a r e d i r e c t l y connected with t h e s e p r i n c i p l e s . Various a s p e c t s of t h i s i d e a may b e found i n t h e monographs of V.V.Bolotin (BibLZb, 2c). 225

i.'

W-

Such equations show t h e p o s s i b i l i t y o f t h e existence of i n s t a b i l i t y r e g i o n s r e p l a c i n g t h e i s o l a t e d c r i t i c a l v a l u e s of T obtained from t h e q u a s i l i n e a r equations of t h e f i r s t approximation. These same remarks apply t o t h e problems of d y m L c s .
A complete i n v e s t i g a t i o n of t h e questions touched upon here would be beyond t h e scope of k h i s book. W e note a g a i n t h a t a number of t h e above propert i e s f o r n o n l i n e a r l y deformed s h e l l s can be found by o t h e r methods, without t h e us2 of t h e l i n e a r approximation developed b y u s f o r t h e components of t h e finite-deformation t e n s o r .

S e c t i o n 8.

Construction of a FIomogeneous I s o t r o p i c S h e l l Approximately Equivalent t o a Layered S h e l l

/225

I n Sects. 26-27 of Chapter 1 1 1 , we considered t h e equations of motion of a two-layered s h e l l . Retaining twelve degrees of freedom on t h e normal t o t h e b a s i c s u r f a c e of t h e shell, we obtained a system o f e q u a t i o n s ofmotion of t h e t h i r t y - s i x t h order. Clearly, we m u s t seek methods f o r obtaining a mathematical f o r n u l a t i o n of the p x b l e m tha-t would make i t solvable i n p r a c t i c e .
W e s h a l l here consider t h e method of solving t h e problem of t h e motion of a l a y e r e d s h e l l , based on t h e approximate replacement of t h i s s h e l l by a homoseneous s h e l l . S t u d i e s i n t h i s d i r e c t i o n , and a study of a s i m p l i f i e d system of equations by means of t h e s e l e c t i o n of a b a s i c s u r f a c e of t h e l a y e r e d s h e l l , have been performed b y E.L.Aksel*rad, E.L.Grigolpk, and V.I.Korolev ( B i b l . l S a , b, 21, 2 ~ ) . I n c o n t r a s t t o t h e s e i n v e s t i g a t o r s , we s h a l l here apply methods of approximation f u n c t i o n s connected with t h e requirement of t h e least-square e r r o r i n c o n s t r u c t i n g t h e Lagrange flinction L3:- of a homogeneous s h e l l approximately equivalent of t h e layered shell. W e s h a l l here i n d i c a t e t h r e e methods of solving t h i s problem. The f i r s t i s based on t h e consideration of a n incompatible system of a l g e b r a i c equations e s t a b l i s n e d independently of t h e p r o p e r t i e s of t h e v a r i a b l e s e n t e r i n g i n t o t h e Lagrange function. The second method i s connected w i t h a general e v a l u a t i o n of t h e magnitudes of t h e s e variables. The t h i r d method r e l i e s on a prelimi n a r y s o l u t i o n of s p e c i f i c problems of t h e dynamics of homogeneous shells. W e 1 1 1 , 26.8 - 2 6 . 1 1 ) , s i n c e t h e presence of t h e s h a l l not base our work here on ( covariant d e r i v a t i v e s Vi uJ ( O ) i n t h e expressions f o r t h e c o e f f i c i e n t s V j ( O and VJ() makes t h e s e equations u n s u i t a b l e f o r s o l u t i o n of t h e problem w i t h which we a r e now corlcerned. W e will make u s e of a n i d e a whic5 i s t h e inverse of t h a t L.A.Molotkov i n 3ne of h i s papers on e l a s t i c waves i n l a y e r e d s i d e r s a medium, in!!omogeneous i n t h e d i r e c t i o n of one of t h e tle l i m i t i n g case of a layered medium. W e shall consider t h e
X-

advanced by e conmedia%. H coordinates, as l a y e r e d s h e l l as

L.A.Kolotkov. Engineering Zquations of Vibrations of P l a t e s with a Layered S t r u c t u r e . Questions of t h e Dynamic Theory of t h e Propagation of Seismic Waves, Vol. V, Leningrad otd. I n s t . m t e m . AN SSSR, 1961

a s p e c i a l case of a s h e l l inhomogeneous i n t h e d i r e c t i o n of t h e normal t o i t s b a s i c surface. This method of studying t h e mechanics o f l a y e r e d s h e l l s of b e discussed below. course involves tine d i f f i c u l t i e s which will,

I. Application of a n Incompatible System of Algebraic Equations

W e shall confine ourselves t o a study of t h e question i n i t s l i n e a r f o r mulation. Consider a s h e l l inhomogeneous i n t h e d i r e c t i o n of %he normal t o t h e b a s i c surface. Let us s e l e c t t h e b a s i c s u r f a c e as i n d i c a t e d in. (111), Sect.25). The object of t h e approximation will be t h e Lagrange f u n c t i o n LdS of a p r i s m a t i c elenent of a s h e l l of height 2h with t h e base area dS:

1226

where TdS and RdS a r e , r e s p e c t i v e l y , t h e k i n e t i c energy and t h e s t r a i n energy of this element of t h e s h e l l . The f u n c t i o n s L, T and II a r e t h e r e s p e c t i v e d e n s i t i e s of t h e Lagrange function, of t h e k i n e t i c energy, and of t h e s t r a i n energy. Hereafter, f o r b r e v i t y , we s h a l l o f t e n omit t h e term t ~ d e n s i t y t and ~ c a l l L, T and II r e s p e c t i v e l y t h e Lagrange function, t h e k i n e t i c energy and t h e straTn energy.

L e t us construct a homogeneous s h e l l of thickness 2h*, of d e n s i t y p% and vdth t h e e l a s t i c Lam6 constants A* and p*, assuming t h a t t h e b a s i c surface of t h e inhomogeneous s h e l l and t h e b a s i c s u r f a c e of t h e homogeneous s h e l l , approximately equivalent t o i t , c o i n c i d e , s t a r t i n g out from t h e condition of least error

ai l L - L*,
which a r i s e s i n t h e above-indicated s u b s t i t u t i o n .

(8.21
, _ , -

me required q u a n t i t i e s p*, A*, @ and h* must be found from t h e condit i o n s of optimum approximate r e p r e s e n t a t i o n s of t h e f m c t i o n L by t h e function L * .
The number of a v a i l a b l e q u a n t i t i e s i n c r e a s e s t o - s i x i f we abandon t h e preliminary s e l e c t i o n of t h e b a s i c s u r f a c e s i n t h e l a y e r e d s h e l l and t h e s h e l l equivalent t o it. The i d e a of s e l e c t i n g t h e b a s i c s u r f a c e so as t o introduce s i m p l i f i c a t i o n s i n t o t h e system o f e q u a t i o n s o f t h e t h e o r y of l a y e r e d s h e l l s . i s discussed elsewhere (Bibl.15, 2 4 ) . Since w e i n t e n d i n t h e following t o d i s c u s s only t h e p r i n c i p l e s of t h e The meaning of t h e requirement of l t o p t i below.
227

r e p r e s e n t a t i o n t 1 will be explained

proposed method, we will almcrst everywhere confine ourselves t o an a r b i t r a r y s e l e c t i o n of t h e f o u r abovp parameters, W e s h a l l confine ourselves t o t h e problem of determining t h e s e q u a n t i t i e s , A, , pi , hi f o r t h e layered s h e l l , dispensing with t h e study of t h e i r connection with t h e metric of t h e s h e l l . I n t h a t case, we may pass t o t h e l o c a l Cartesian system of rectangular coordinates with t h e altis OZ d i r e c t e d d o n ? t h e normal t o t h e b a s i c s u r f a c e i n s i d e t h e s h e l l , and t h e x ' l o c a t e d i n a plane tangent t o t h e b a s i c surface. I n t h i s case, g i i = axes O = 8 " = 1, -.lk = 0 (i # k ) : an element of volume i s expressed by t h e derivat i v e dxl and t h e element of a r e a dS Will b e equal t o dxldxa.
o n l y as f u m t i o n s pi

&'&,

e must express L and L s i n t h e same v a r i a b l e s . I n /227 I n ccnsidering Ci w t h e choice of t h e s e v a r i a b l e s we s h a l l s t a r t out from t h e well-known propert i e s of t h e f i e l d s gf displacements, s t r a i n s , and s t r e s s e s i n a l a y e r e d s h e l l .

It has been shown i n Chapter 1 1 1 , Sects.25-27, t h a t , i n a layered s h e l l , t h e componenf,s of t h e displacement vector, t h e components Eik(i, k = 1, 2) of t h e s t r a i n tensor, and t h e components T r 3 ( i = 1, 2) of t h e stress t e n s o r are continuous.
X e shall base our work on t h e assumption t h a t t h e f i e l d s of t h e s e quantit i e s coincide i n t h e layered s h e l l and t h e equivalent homogeneous s h e l l t o within $he limits of t h e prismatic element mentioned above.

It i s well known t h a t a l l corkinuous functions of t h e coordinate z can b e aFproxk"aed by polynomials of zs Since we shall determine L and L s with an accuracy t o t e r m s of t h e order h and h*$, and i n t e n d t o give here o n l y t h e general p r i n c i p l e s of t h e method preposed, l e t us put

Further, from Zooke's law, w e find

where

228

(8.7)
Equations (8.5) (8.6) show that the representation of the components of the displacements by (111, 15.5) is inapplicable to the problems of the v i brations of a layered shell.

In fact, as w i l l be clear from eqs.(8.5) - (8.6), the continuwx coefficients of Z. in the formulas determining the components uJ are expressed in terms of piecewise-continuous functions of z, constant on those segments of the 02 a x i s included within the layers. This dependence of the coefficients / Z 8 m on z is not reflected in explicit form by (111, 15.5). of z
Let us a l s o approximate the contiinums components sor by the polynomials
t

cik

of the strain ten-

( $ . $ a )

in this connection
2

(8.8b)
i-1

Now let us express the discontimotls components of the strain tensor and the stress tensor in terns of continuous compoxents. We have

2 2 9

Since we are i n v e s t i g a t i n g t h e l o c a l p r o p e r t i e s of t h e Zunction L, without going outside t h e boundaries of t h e p r i s m a t i c element of t h e s h e l l defined above, we must consider the derivatives o, (i = I, 2) of t h e c o e f f i c i e n t s of t h e polynomials here introduced as new, l o c a l l y independent, q u a n t i t i e s . This, m D r e p a r t i c u l a r l y , e x p l a i n s t h e a p p l i c a t i o n of t h e independent r e p r e s e n t a t i o n s , b y t h e polynomials, of t h e wmponents c i k without i n v e r s i o n t o eqs.(8.5)-(8.6), s i n c e t h i s would n o t l e a d t o a decrease i n t h e nanber of l o c a l l y independent q u a n t i t i e s introduced by us.

All t h e l o c a l l y independent q u a n t i t i e s e n t e r i n g i n t o eqs. (8.3) - (8.10b) belong t o t h e v a r i a b l e f i e l d i n terms of which t h e Lagrange f u n c t i o n L of a n element of t h e continuous medium i s expressed. They a r e a g e n e r a l i z a t i o n of hhe generalized coordinates and generalized v e l o c i t i e s k n o m from c l a s s i c a l mechanics+. W e r e c a l l t h a t i r ? the c l a s s i c a l Lagrange function, t h e generalized /229 v e l o c i t i e s 9j and t % e generalized coordinates 9j are considered as independent q m n t i t i e s i n s e t t i n g ' ~ p t h e equations of notion. Their i n t e r r e l a t i o n i s taken i n t o account a f t e r s e t t i n - ? -1p t h e equations of motion based on t h e
J = da_l elementary equations o

dt'

I n t h e problem of i n t e r e s t t o us, t h e i n t e r -

r e l a t i o n of t h e v a r i a b l e f i e l d s i s expressed b y more complex r e l a t i o n s res u l t i n 5 from t h e equations of e l a s t i c i t y theory. For example, from eqs.(8.4) arld (8.6) we may f i n d

(8.11)
Squation (6.7) must b e a s s o c i a t e d w i t h t h e s e r e l a t i o n s .
k number of r e l a t i o n s result from eqs.(8.6) - (8.1Ob), b u t we s h a l l not consider them here, s i n c e we are not s e t t i n g up a system of equations of mot i o n of a l a y e r e d s h e l l bg t h e methods of c l a s s i c a l a n a l y t i c a l mechanics, b u t propose t o make u s e of t h e Lagrange f u n c t i o n L as t h e fundamental q u a n t i t y i n Lhe problem of c o n s t r u c t i n g a homogeneous s h e l l approximately equivalent t o a layered s h e l l .

It follows from eqs.(8.3) - (8.lOb) t h a t t h e number of v a r i a b l e f i e l d s It i s easy t o e s t a b l i s h e n t e r i n g i n t o t h e Lagrange f u n c t i o n i s t,hirt;y-one. t h a t t h i s number does not depend on t h e number of l a y e r s n i n t h e s h e l l if n >1. For n = 1, t n e nmber of variables of t h e f i e l d can be reduced t o twenty-s eve n.
face. Consider now t h e k i n e t i c energy T related t o u n i t area of t h e b a s i c surI n t h e l o c a l system of r e c t a n g u l a r C a r t e s i a n coordinates, we have

(8.12)

4 :

Cf.,

f o r example, J.Leach,

C l a s s i c a l Mechanics, Chapter

IX, IL, 1961.

Equation (8.12) i s a p p l i c a b l e t o an inhomogeneous s h e l l . a layered s h e l l , p u t t i n g :

n-I

Let us p a s s t o

k=l
n-1

k-I

n-ll

where Apk , Ah , & a r e t h e changes i n p , A, CL on t r a n s i t i o n f r o m t h e kth l a y e r t o t h e (k + l\tn l a y e r , n i s t h e number of l a y e r s , uo i s t h e u n i t Heaviside function, and zl, a r e t h e coordinates of t h e i n t e r f a c e s of t h e l a y e r s . It /230 i s assumed t h a t t i e l a y e r s a r e p a r a l l e l , i . e . , t h a t t h e coordinates z, a r e const a n t s . f t h e form To shorten t h e formulas we s h a l l also make use of r e l a t i o n s G of eqs.(8.13a) - ( 8 . 1 3 ~ ) i n considering t h e functions p , A, P.

Let us a l s o introduce t h e notation:

and r e t a i n i n eqs.(8.5) - (8.6) o n l y t h e terms containing t h e f a c t o r s z. t o 'z inclusive. This makessit p o s s i b l e t o f i n d T approximately t o terms With f a c I n t h i s case, T Will c o n t a i n o n l y seventeen v a r i a b l e t o r s of t h e order of h f i e l d s . W e have

231

( 8.16a )

S u b s t i t u t i n g t h e s e equations i n t o eq.(8.12), (8.1/+), (8.15), we f i n d

and making use of eqs.(8.7),

23 2

I , ,

.I

I1 I

I1

I111111

Further, we have 2rI =

[Tll%l+

T&t

+ w 3 3

2 ('1,E12+

Tl3E13

w 2 3 ) J

dz.

(8.18)

Confining ourselves t o t h e r e l a t i v e accuracy adopted in c a l c u l a t i n g T, and (8.1Ob), we f i n d making use of eqs.(8.8a)

(8.19)

Here t h e q u a n t i t i e s 6 (d and c i $)

are connected b y t h e r e l a t i o n (8.8b).

Equations (8.17) and (8.19) complete t h e c o n s t r u c t i o n of t h e f u n c t i o n L f o r a l a y e r e d s h e l l with t h e accuracy adopted by us. A s a s p e c i a l case, t h e s e equations y i e l d t h e ejrpression of t h e Lagrange f u n c t i o n L%,for a homogeneous ( s i n g l e - l a y e r ) s h e l l i n t h e v a r i a b l e f i e l d s s e l e c t e d by us.

For a s i n e - l a y e r s h e l l , eq.(8.15)

t a k e s t h e f o l l o d n g form : (8.20)

+ Fd i r e c t l y from eqs. (8.17a)Using t h i s notation, we can o b t a i n 2TX- and Z (8.13) and ( S . 1 9 ) , b u t s i n c e a l l this reduces down t o s u b s t i t u t i n g t h e operfor A,(f), a t o r s &-(f+:-)

we s h a l l not write out t h e expressions f o r 2T?F and ;?I:*.

Let us rpturn t o the problem of t h e approximation of t h e Lagrange func* . t i o n L by t h e f u n c t i o n L

L e t u s consider t h e d i f f e r e n c e A . As will b e s s e n from eqs.(6.2) and t h e p r o p e r t i e s of T, II, W-, E%-, t h i s d i f f e r e n c e in t u r n i s a Lagrange f u n c t i o n with t h e coefficients

(8.Z)
I f a f u n c t i o n L+F e d s t e d equal t o L, t h e n a l l t h e d i f f e r e n c e s 4 , and t h u s a l s o t h e d i f f e r e n c e A , would vanish. This vanishigg p f A would mean t h e e x i s t e n c e of a single-layer s h e l l equivalent i n t h i s r e s p e h t t o a multi-layer s h e l l . But we have a v a i l a b l e only f o u r q u a n t i t i e s c h a r a c t e r i z i n g t h e propert i e s of t h e single-layer s h e l l : p+, A-%, k-x- and %$. Equating a l l t h e i l , to /233 6 . 1 9 ) , a system of zero, we o b t a i n , as w i l l b e seen from eqs.(6.1%) and ( twenty-five equations i n f o u r udmowns:

A (f) = 0.
m

(8.22).

The System (8.22) i s inconpatibIe. Consequently, i t i s impossible t o cons-i;ruct a homogeneous s h e l l equivalent t o a multi-layer shell*. 5Je
Carl

speak only of approximate equivalents.

It i s w d l known t h a t t h e r e a r e s e v e r a l methods of c o n s t r u c t i n g s o l u t i o n s
3:-

This conclusion was o b v h u s i n advance, s i n c e , i f t h e f u n c t i o n s L and L* were e x a c t l y equal, and t h e e x t e r n a l loads and boundary c o n d i t i o n s of a single-layer s h e l l coincided with t h a t of a multi-layer s h e l l , t h e single-layer s h e l l would e x a c t l y imitate t h e motion of t h e multi-layer s h e l l .

234

approximately s a t i s f y i n g a system of incompatible equations. Let us applz t h e method based on t h e requirement of minimizing t h e sum o f 4 . Let us consider t h e sum i n t h e expanded form, making use of eqs.(8.17b), (8.19) and (8.20):
the squares of

S O = ~ A= ; [A,(?) -2/~p*]*+[A,

(p) --(2/P)2p*]2+ 1 2

235

LI

II

II

I l l

II

I I I I I I

11111

II

I l l

I I

II1111 I l l

IIIII

I I I I I 111 I I

111 I

II I

1111I

II

II.II

111 I

11111

Let us i n t r o d u c e t h e n o t a t i o n

To determine t h e unknown parameters c h a r a c t e r i z i n g t h e homogeneous s h e l l , l e t u s s e t up t h e equations:

From t h e system of n o d i n e a r equations (8.25) we determine t h e parameters GS, ! I * and h*, a f t e r which t h e c o n s t r u c t i o n of t h e homogeneous s h e l l , a p p r o b t e l y r e p l a c i n g t h e l a y e r e d s h e l l , W i l l b e completed. Several remarks must b e made on t h e method proposed here f o r t h e cons t r u c t i o n of an equivalent homogeneous s h e l l .

a. me system of nonlinear equations (8.25) f o r s u f f i c i e n t l y small Ap,, Ah, and &,, obviously has a t l e a s t one system of real s o l u t i o n s , s i n c e for A p , , Ah, and aiii equal t o zero, we o b t a i n t h e s o l u t i o n p s = pl, A* = A , , FLX- = cL1Y a ; ; = h, = 2h.
b ) The incompatible system of equations (8.22) and t h e r e l a t e d funct i o n So were coilsidered by u s apart from t h e d i f f e r e n c e L L*. W e certainly had t h e r i g h t t o proceed i n t h i s way, b u t t h e s o l u t i o n proposed involves t h e i n p l i c i t assumption t h a t t h o s e f u n c t i o n s of t h e v a r i a b l e f i e l d i n t h e expression f o r L L* whose c o e f f i c i e n t s a r e t h e d i f f e r e n c e s & on t h e left-hand s i d e s of t h e incompatible equations (8.22) a l l have t h e same physical s i g n i f i cance. This i s undoubtedly t h e vulneFable p o i n t of t h e method. Evidently, i n s t e a d of t h e f u n c t i o n So, we should consider t h e f u n c t i o n of a more g e n e r a l type :

s, = SC,A2,,

(8.26)

where c, i s t h e weight of t h e term 4' c h a r a c t e r i z i n g t h e physical s i g n i f i L*c a x e of t h e corresponding term++ i n t h e expression f o r L

The p r i n c i p a l d i f f i c u l t y here l i e s i n t h e determination of t h e numbers W e s h a l l f i r s t give a n elementary example of t h e choice of t h e c o e f f i c i e n t s s , based on t h e c l a s s i c a l t h e o r y of s h e l l s .

s.
/235

The Kirchtnoff-Love hypotheses l e a d t o t h e conclusion t h a t , i n t h e Lagrange functions L and L++, we may neglect a l l terms containing v a r i a b l e f i e l d s connected with t h e components e 1 2 of t h e s t r a i n tensor and ei3 of t h e s t r e s s tensor. Consequently, i n eq.(8.26) we should equate t h e c o e f f i c i e n t s c, of t h e corresponding A, t o zero. If we p u t t h e remaining c o e f f i c i e n t s c , as equal t o unity, t h e n we f i n d
r)

+[A,(

1:;J-T

(2h*)3 2p*h* ] ' + [ A , (2p) - (2h*)2p+12+ 'A .* 2p *

A, (2p) -- 1 (2/~.')~2p" 2

A , (2p) - - ( 2 / ~ * ) ~ 2 p ~ .* 3

1 .

I'

I n c o n t r a s t t o eq.(8.23), we introduced t h e c o e f f i c i e n t ;E" i n c e r t a i n 5erms of eq.(8.27), bearing i n mind t h e numerical c o e f f i c i e n t s of t h e corresponding terms i n t h e expressions f o r 2T and Zn. T h i s again corresponds t o t h e assumption of t h e same physical s i g n i f i c a n c e of t h e functions of t h e v a r i a b l e f i e l d s f o r a l l c o e f f i c i e n t s of t h e form a A , ( f ) where a i n this case equals u n i t y o r two.
-

3:.

C f . , f o r example, V.Ya.Goncharov, Functions, O N Z , 1934, p.161

Theory of I n t e r p o l a t i o n and Approximation 237

-..

..

. .. ._.

W e now consider t h e case when t h e components c,,(O)(i,k = 1,2) vanish on deformation of t h e s h e l l . I n t h i s case, we cannot n e g l e c t t h e terms containing of t h e stress tensor. t h e components Ti@
The determination of t h e required parameters, c h a r a c t e r i z i n g t h e propert i e s of a homogeneous s h e l l approximately equivalent t o a layered s h e l l , i s again reduced t o t h e s o l u t i o n of t h e system of equations: (8.28) The methods based on t h e consideration of t h e sms S, and SI,, defined /236 by eq. (8.27) y i e l d t h e r e s u l t of averaging incompatible val!es of t h e required unknowns obtained from t h e system of equations (8.22).
m i t s constructing of an equivalent s h e l l with rough approximation,

This r e s u l t of averaging, as will be seen from t h e above discussion, persince the physical meaning of t h e q u a n t i t i e s entering i n t o T and n, which a r e f a c t o r s e must t h e r e f o r e consider i n g r e a t e r d e t a i l t h e of 4 , d i f f e r s u b s t a n t i a l l y . W c o e f f i c i e n t s ci of eq.(8.26).

2. Evaluation of t h e ~Weights - ci
~

W e admit t h a t t h e technique proposed below f o r evaluating t h e weights i n eq.(8.26) i s q u i t e imperfect. However, i t permits a n introduction, i n t o t h e c a l c u l a t i o n s , of q u a n t i t i e s approximately c h a r a c t e r i z i n g t h e s i g n i f i c a n c e and physical p r o p e r t i e s of various groups of terms e n t e r i n g i n t o t h e funcL+F. Here, as before, i t i s necessary t o r e f i n e t h e region of v a r i a tion L t i o n of t h e v a r i a b l e f i e l d , s i n c e t h i s region i s a t t h e same time t h e region of approximation of t h e f u n c t i o n L b y t h e f u n c t i o n L.Y.

W e s h a l l assume t h a t t h e s h e l l undergoes s t a t i o n a r y v i b r a t i o n s a t a frequency w l y i n g i n t h e i n t e r v a l (w, , w, ). The q u a n t i t i e s wI and ur, a r e assumed t o be known. To determine t h e frequency w1 we may use any approximation methThe upper value of t h e frequency w may be od, f o r example t h e R i t z method. e will show t h e influence of t h i s choice. Let us asselected a r b i t r a r i l y . W sume, f o r i n s t a n c e , t h a t t h e displacements u, a r e expressed by t h e equations
ui =

ui sin w t
and (8.19) and t h e n avero),; 0,

Let us s u b s t i t u t e eq. (8.29) i n t o eqs.(8.1%) age t h e r e s u l t s over t h e two-dimensional region b e a r i n mind t h e equations

( wl,

").
W

L e t us

2x -

cos w f sin w t d t =0;

2x w

2rr

cosz w f d t d w

L e t us temporarily introduce i n t o t h e consideration t h e v a r i a b l e If, def i n e d by t h e r e l a t i o n

The v a r i a b l e s Y f a r e defined by t h e values taken by t h e r a t i o ( a , f ) : ( f ) on t h e b a s i c surface, where f i s t h e general symbol f o r t h e f u n c t i o n s charact e r i z i n g t h e s t r e s s - s t r a i n s t a t e of t h e s h e l l . Obviously t h e v a r i a b l e s Y f a r e i n p a r t i c u l a r connected w i t h t h e v a r i a t i o n a l i n d i c e s of t h e f u n c t i o n f (Bibl.5) I n considering s p e c i f i c problems on t h e v i b r a t i o n s of layered s h e l l s , we can sometimes determine i n advance t h e approximate limits c1 and Q w i t h i n which t h e values of t h e v a r i a b l e s If w i l l l i e , from t h e known s o l u t i o n s of similar problems f o r homogeneous s h e l l s , and then average over Y f t h e difference L - L+ on t h e i n t e r v a l s ( c l , % ) .
A s a r e s u l t , t h e constants c ( , ) ,

defined by t h e equation

enter

t h e equations.

Since t h e q u a n t i t i e s c(,)have a d e f i n i t e meaning only f o r very narrow c l a s s e s of problems, we s h a l l below apply various methods permitting t h e i r exclusion from t h e equations solved. L e t us continue our study of t h e v a r i a t i o n a l i n t e g r a l s of t h e variable. f i e l d s e n t e r i n g i n t o t h e Lagrange functions of homogeneous and l a y e r e d s h e l l s , assuming t h a t t h e components uJ (j = 1, 2, 3 ) of displacement v a r y over t h e i n t e r v a l s (-2h, +2h). It i s w e l l known t h a t , i f t h i s i n t e r v a l is f u r t h e r extended, t h e l i n e a r theory becomes unsuitable.

239

The g r e a t e s t d i f f i c u l t i e s a r e connected with i n d i c a t i n g t h e v a r i a t i o n a l i n t e r v a l s of t h e q u a n t i t i e s BJ;) and T::) (i, k = 1, 2; m = 0, 1 , 2). These i n t e r v a l s depend l a r g e l y on t h e s p e c i a l p r o p e r t i e s of t h e s o l u t i o n s of d e f i n i t e c l a s s e s of boundary problems. W e s h a l l s t a r t o u t from t h e hypotheses t h a t y i e l d r e s u l t s i n t h e general form, admitting of f u r t h e r s i m p l i f i c a t i o n and connected with t h e s p e c i a l p r o p e r t i e s of s p e c i f i c problems.

Assume t h a t t h e components ci(f) (i, k = 1 , 2) of t h e s t r a i n t e n s o r v a r y over t h e i n t e r v a l (-a, + a ) where a i s t h e g r e a t e s t value taken by this quan- /238 t i t y ( c f . Sect.2) i n t h e various layers. The exceptions a r e t h e cases i n which we know i n advance t h a t t h e e x ' are small o r zero, when t h a t range of v a r i a t i o n contracts t o a point.
of t h e s t r e s s tensor vary over Assume f u r t h e r t h a t t h e components T ! : ) t h e i n t e r v a l (-%, +%). I f t h e r e i s reason t o suppose t h a t t h e s t r e s s e d s t a t e of t h e s h e l l i s a l m o s t momentless, then t h e q u a n t i t y co must be dettirn:iric:? from t h e r e l a t i o n
C,

=2~,k,,,

i n accordance w i t h t h e i n t e r v a l of v a r i a t i o n c i k y i e l d point of t h e m a t e r i a l s of t h e l a y e r s and p r i n c i p a l curvature of t h e ' b a s i c surface.

(0)

Here a' i s t h e m a x i m u m t h e g r e a t e s t value of t h e

I n purely f l e x u r a l deformations of t h e s h e l l , taking place i n t h e absence of l o a d s on i t s boundary surfaces, t h e absolute values of t h e components T{!) (i = 1 , 2, 3 ; m = 0, 1, 2) are small. I n t h i s case, t o o b t a i n approximate s o l u t i o n s , t h e i n t e r v a l (-cay co) must b e contracted t o a point. Assume f u r t h e r t h a t a l l t h e nonzero q u a n t i t i e s cf:' ( i , 1, 2; m = 1,2) vary over t h e range [-(-2h)'a, (2h)'aI while t h e u a n t i t i e s T~~ ( i = 1, 2, 3; m = 1 , 2) vary over t h e i n t e r v a l from [-(2h).c,, ?2h)' c(s1 T h i s i s equivalent t o t h e assumption t h a t t h e terms a c t u a l l y e n t e r i n g i n t o t h e approximation polynomials of t h e form (8.3) are of t h e same r e l a t i v e order.

4 . 7

All the above compels t h e conclusion t h a t weighted quadratic approximat i o n s should b e introduced a f t e r s u f f i c i e n t l y complete c o n c r e t i z a t i o n of t h e content of t h e problem of s h e l l mechanics.
Let us t u r n now t o a consideration of t h e s p e c i a l case of t h e determining of t h e e l a s t i c constants of t h e homogeneous s h e l l , approximately equivalent t o a layered s h e l l .

3. Application of t h e Weighted Quadratic Approximation

u m S,, assuming t h a t among t h e terms of t h e d i f f e r e n c e Let us form t h e s L - L* t h e r ? a r e no zeroes and t h a t a n averaging has been made over t h e above i n t e r v a l s of v a r i a t i o n of t h e v a r i a b l e f i e l d s and over t h e i r time d e r i v a t i v e s e n t e r i n g i n t o t h e composition of t h e k i n e t i c energy.

I n considering t h e terms i n t h e coordinates x i , w i t h a comma according t o eq.(d), followed by t a i n t h e constants ccml, defined

2T, which contain d e r i v a t i v e s with r e s p e c t t o i n t h e indices, l e t us use t h e v a r i a b l e Yf

averaging. by eq.(e). Thus t h e f i n a l r e s u l t w i l l con-

Let us i n t r o d u c e t h e n o t a t i o n

/239

Now, based on e q s . ( 8 . 1 n ) and (8.19), l e t us form t h e sum SI of t h e squares of t h e d e v i a t i o n s from zero of t h e averaged values of t h e independent summands e n t e r i n g i n t o t h e d i f f e r e n c e L L*.

As before, l e t u s consider t h e v a r i a b l e f i e l d s introduced by us and t h e i r e have d e r i v a t i v e s as independent variables. W

S, =

c,AL = (Aw), 16h" [A, (p) - 2/z* p*]'

+64 /i4c(,) [ A , (p) 9

--

-j- a,* [A, (2p) - (2he)(2pL*)I2+16he4a4A, (2p) - 1 (2/~*)~ 2p3:

The following s h o r t remarks apply t o t h e sum S,, ments i n Subsection 2 of t h i s Section.

supplementing t h e state-

a ) The s u m Sl contains a smaller number of summands t h a n t h e sum S,, since, on averaging over symmetric i n t e r v a l s , t h e terms containing odd powers of t h e v a r i a b l e f i e l d s w i l l cancel out. b ) As a l r e a d y noted, a t e r m containing a f a c t o r c ( ~, ) can r e c e i v e a d e f i n i t e meaning o n l y f o r d i s t i n c t l y r e s t r i c t e d c l a s s e s of dynamic problems.
I,, and t h e constant cc2)

I n s t a t i c problems i t Will not be necessary t o i n v e s t i g a t e t h e v a r i a b l e s Will not e n t e r i n t o t h e eq,yations. I n t h e remain- /240 i n g cases an attempt must be made t o e l i m i n a t e t h e terms with t h e f a c t o r cc2) The most g e n e r a l method of e l i m i n a t i n g one of t h e terms with such a f a c t o r i s t o choose t h e b a s i c surface i n t h e equivalent homogeneous s h e l l such t h a t t h e c o e f f i c i e n t W i l l vanish i f t h e f a c t o r cC2) occurs i n t h e s e l e c t e d term.

Equation (8.31) contains only one term with t h e f a c t o r c t 2 ) . Let us denote by h, t h e z coordinate determining t h e new p o s i t i o n of t h e b a s i c s u r f a c e r e l a t i v e t o t h e above-selected surface. To determine h,, l e t u s s e t up t h e f o l l o w i n g equation r e s u l t i n g from eq.(8.31) :

, the posiI f eq.(8.31) s t i l l contains another term with t h e f a c t o r cC2) t i o n of t h e b a s i c s u r f a c e would a l s o have t o be changed i n t h e layered s h e l l .
I n t r o d u c t i o n of t h e frequencies w1 and cu, shows t h a t t h e approximate C replacement of t h e l a y e r e d s h e l l by a homogeneous s h e l l permits i n v e s t i g a t i o n of only a l i m i t e d r e g i o n of t h e frequency spectrum. The higher frequencies cannot be determined by Chis method. d ) Equation (8.31) w a s obtained under c e r t a i n assumptions, which might

f a i l t o correspond t o t h e physical content of i n d i v i d u a l problems. But t h e form of t h e r e l a t i o n (8.31) permits i t s a d a p t a t i o n t o a number of s p e c i a l cases. T h i s has a l r e a d y been mentioned above.
To supplement t h e above we note t h a t i n t h e case of p l a t e s , t h e c o e f f i c i e n t QJ must be t a k e n as zero, and f o r f l a t s h e l l s c l o s e t o zero, i n accordance with eqs.(8.30a). These cases approximate t h e assumptions under which the s u m ,310 was obtained.

If a p l a t e , under c e r t a i n boundary conditions, has no chain s t r e s s e s , t h e n t h e evaluation of t h e v a r i a t i o n a l i n t e r v a l of t h e s t r e s s t e n s o r compone n t s 713 by means of t h e q u a n t i t y co defined by t h e d i f f e r e n c e (8.30a) l o s e s i t s meaning. I n t h e s e cases, one must s t a r t out from t h e s p e c i a l conditions of loading of t h e boundary surfaces of t h e s h e l l . A t considerable s u r f a c e d e n s i t i e s of the load, co may be taken as equal t o a,, i . e . , we may average t h e s t r e s s tensor components 'ri3 over t h e i r n a t u r a l i n t e r v a l . O f course, t h i s a p p l i e s a l s o t o t h e corresponding cases of deformation of s h e l l s .
W e emphasize i n conclusion t h a t t h e expression obtained by us f o r t h e

242

s u m S , should be regarded merely a s an example of t h e a p p l i c a t i o n of t h e gene r a l method. I n solving s p e c i a l problems one must s t r i v e toward a preliminary i n d i v i d u a l i z a t i o n of t h e f i e l d variables, permitting d i f f e r e n t v a r i a t i o n a l int e r v a l s t o be prescribed f o r them and permitting t h e i r r e l a t i v e magnitude t o be estimated, a s noted above. For this reason, t h e f o r d a p p l i c a t i o n of eq. (8.31) t o a r b i t r a r y problems cannot be recommended.
e ) Equation (8.31), as w a s assumed, approximately r e f l e c t s t h e physical s i g n i f i c a n c e of t h e i n d i v i d u a l terms e n t e r i n g i n t o t h e sum S,.
O f course i n a comparative estimate of t h e magnitudes of t h e i n d i v i d u a l summands of t h e sum S, one must, as a l r e a d y noted, s t a r t from t h e physical cont e n t of t h e problems of a d e f i n i t e c l a s s . This permits an i n t r o d u c t i o n of Let us consider one v e r s i o n of t h e s e f u r t h e r s i m p l i f i c a t i o n s i n t o eq.(8.31). s i m p l i f i c a t i o n s , s t a r t i n g out from t h e b a s i c assumptions made i n deriving eq. (8.31).

L &

There, w e assumed t h a t t h e t a n g e n t i a l components of t h e s t r e s s tensor and t h e s t r a i n tensor d i f f e r from zero and t h a t t h e r e l a t i v e magnitudes of t h e f l e x u r a l and chain s t r e s s e s a r e t h e same. Comparing under t h e s e conditions t h e expressions containing t h e operators and &, we f i r s t of a l l note t h a t t h e orders of and A2 h- a r e $he same. Therefore, i n comparing terms contain& and &, t h e f a c t o r h must be a t t r i b u t e d t o t h e t e r p s containing &. ing ; )' Finally, Then, we note t h a t t h e order of t h e r a t i o & : p1 equals a we must bear i n mind t h e numerical c o e f f i c i e n t s of t h e summands entering i n t o t h e sum S,.

W e mentioned above t h a t t h e terms containing t h e f a c t o r c C z ) had t o be excluded from t h e equations by various methods, f o r example, by a r a t i o n a l choice of t h e b a s i c surface i n t h e layered s h e l l and i n t h e approximately equivalent homoqeneous s h e l l . Here, however, w e s h a l l not change t h e p o s i t i o n of t h e b a s i c surface, but s h a l l d i r e c t l y equate t o zero t h e c o e f f i c i e n t of c c z ) i n , (under t h e above assumpThen, r e t a i n i n g t h e dominant terms i n S eq.(8.31). t i o n s ) , we f i n d t h e following simplified expression f o r t h e sum S,:

A s dl1 b e seen from t h e s i m p l i f i e d expression f o r S , , t h e t e r m s depending on t h e chain s t r e s s e s have been dropped f r o m i t . This i s a consequence of t h e assumptions we made on tne region of approximation and t h e neglect of a number of terms e n t e r i n g i n t o eq.(8.31). It i s c l e a r t h a t t h e r e s u l t s obtained from eq. (8.33) a r e only roughly approximate.

For S , t o vanish i t i s s u f f i c i e n t t o equate t o zero t h e expressions i n t h e

243

brackets. Also bearing i n mind t n e condition f o r t h e vanishing of t h e coeffic i e n t of c ( ~ ) i n t h e right-hand s i d e of eq.(8.31), we f i n d t h e following / a 2 simultaneous system of equations:

2h* P* = A, (PI,

- (2/1"j3 p* =A (p), 3
y (2h*)32p* = A , (2p),

(8.344

Hence, w e find

Of cm.rse, i f cl-x and A% a r e determined by eqs.(8.35b)-(8.35~), w e may i n s p e c i a l cases o b t a i n llphysically impossible1! values of t h e Poisson constant V , which may above a l l i n d i c a t e t h e u n s u i t a b i l i t y of t h e simplified express i o n (8.33)- f o r t h e s u m s,. Here, we will not i n v e s t i g a t e t h e question as t o t h e i m p o s s i b i l i t y of determining t h e parameters of t h e homogeneous s h e l l from eq.(8.31).

W e s h a l l l i k e w i s e not i n v e s t i g a t e i n d e t a i l t h e question whether i t i s permissible t o formally apply t h e equations w i t h llphysically impossible1' values of Poissonfs constants and t h e physical meaning of such equations. W e recall merely t h a t negative values of P o i s s o n f s constant correspond t o t h e loss of s t a b i l i t y of t h e s h e l l considered i n Sect.6. The ttphysically impossiblet1 values of Poisson's constant f o r a c t u a l m a t e r i a l s may prove t o be p o s s i b l e f o r a /a3 medium approximately equivalent t o t h e r e a l medium, and may r e f l e c t t h e specifi c p e c u l i a r i t i e s of those problems of mechanics f o r which this medium h a s been constructed

f > Let us consider an example of a p p l i c a t i o n of eqs.(8.35a)-(8.35~)*.

Let

a bimetal s h e l l of thickness 2h c o n s i s t of a l a y e r of aluminum of 2/3h,adjacent

*This example i s merely of a n i l l u s t r a t i v e value. 24.4

44 i n t o t h e b a s i c surface, and a l a y e r of d u r a l u m i n
3
p", A* and I.L>?. thickness. Required, t o f i n d a*, Table 6 gives t h e p r i n c i p a l physical c h a r a c t e r i s t i c s of aluminum and duralumin.

From eq.(8.15)

we have:

4 0,267-10-' +-0,12-10-3 3

(2h) =0,283*10-'(2h), (2h)3= 0,929(2h)3,

0,267-

26 . O,I2. IO-' 81

A 2 (2p)=(;

- 0,521~lOG-/--0,O14~1OG 26 (2h)3=0,178-10G (2/~)~, 81

= 0,763.105

(243.

Making use of eqs.(8.35a)-(8.35~),

we f i n d

2p*

0.544. loGbar;

h* = 0,408. IOG bar

a n
4 -

I4

These r e s u l t s show i n p a r t i c u l a r t h a t t h e values of t h e e l a s t i c constants A* and p* found here Will depend on t h e d e n s i t y of t h e l a y e r material. Comparing t h e obtained values of ;?I* and A* /;1wc with those f o r ;S. and h given i n Table 6, we s e e that they cannot b e termed "averaged", s i n c e t h e y a r e outs i d e t h e v a r i a t i o n a l limits of and h i n t h e mat e r i a l layers. Obviously, t h i s i s p r i m a r i l y a consequence of t h e method of determination adopted here f o r t h e reduced t h i c k n e s s of t h e homogeneous s h e l l , 2h*. This t h i c k n e s s w a s found t o be somewhat l e s s than t h e thickness 2h of t h e layered s h e l l . Besides, t h e exc l u s i o n of t h e terms depending on t h e chain s t r e s s e s a l s o had a considerable influence.

L e t us consider a simpler s t a t i c problem. Of t h e system of equations (8.34a)-(8.34d) t h e r e remain P u t t i n g 2h* = 21,we f i n d eqs.(8.34~)-(8.34d).

I I 111111

1 1 1 1 1 1 1111

.. , .. - . .. .....

.-

_... ..
-

From eqs.(8.36) and t h e d a t a i n Table 6, we f i n d

2p* zz 0,535-lo6 Bar;

i* GZ 0,392. IOG bar.

where t h e v a l u e s of &)t. and h* do not go beyond t h e limits o t h e var,ational i n t e r v a l of 3 and h f o r layer materials almost coinciding w i t h t h e values of t h e s e q u a n t i t i e s f o r duralumin, owing t o t h e f a c t t h a t t h e duralumin l a y e r occupies 2/3 of t h e t h i c k n e s s of t h e s h e l l and t h a t t h e r e a r e no terms dependi n g on i n t h e approximate expression (8.33).
g) W e remarked r e p e a t e d l y t h a t i t i s p o s s i b l e t o s i m p l i f y t h e sum S , by a r a t i o n a l choice of t h e b a s i c s u r f a c e i n t h e layered s h e l l and i n t h e equival e n t shell. W e may, f o r example, s e l e c t t h e b a s i c s u r f a c e i n s p e c i a l cases This choice such t h a t t h e components E#) (i, k = 1, 2) on i t s h a l l vanish.:-. of t h e b a s i c surface, however, i s mandatory i n solving s t a t i c problems, s i n c e i n dynamic problems one must f i r s t e l i m i n a t e t h e terms with t h e c o e f f i c i e n t s c ( , ) , whose meaning i s n o t s u f f i c i e n t l y d e f i n i t e i n t h e general case.

h) It i s easy t o e s t a b l i s h t h a t t h e approximate expression (8.33) f o r t h e sum S, i s a l s o s u i t a b l e i f t h e q u a n t i t i e s E i k (i, k = 1, 2) vanish, s i n c e O f course, i f t h e this expression does n o t c o n t a i n terms depending on E : : components E$:) vanish, t h e accuracy of eq.(8.33) and of t h e r e s u l t a n t con- /245 sequences i n c r e a s e s .

i ) I n t h e g e n e r a l case, t h e q u a n t i t i e s A s , P*, p+*, 2hX- a r e determined from t h e conditions t h a t t h e right-hand s i d e of eq.(8.31) s h a l l b e minimum, i.e., from t h e conditions:

dS,-00; arrt*
where

~-

aG*

-0;

-0; -

as,

aH*

I - 0,

as

ah*

(8.37)

The conditions (8.37) must b e a s s o c i a t e d w i t h t h e condition (8.32) t h a t

$5 SugPestions as t o t h e r a t i o n a l choice of t h e b a s i c s u r f a c e are given elsewhere rBib1.15a, b, 21, 25).

t h e c o e f f i c i e n t of cc2) s h a l l vanish i n eq.(8.31). The system of a l g e b r a i c equations (8.32) and (8.37) i s nonlinear i n t h e required q u a n t i t i e s m++, G+:C, H;C, h*, and hl. I t s s o l u t i o n c l e a r l y i n v o l v e s considerable d i f f i c u l t i e s . W e may, f o r example, use t h e method of successive approximation, s u b s t i t u t i n g i n t h e second-power t e r m s of eqs. (8.32) and (8.37) t h e s o l u t i o n s (8.35a) - ( 8 . 3 5 ~ )of t h e s i m p l i f i e d system and determining t h e next approximation, b u t i n t h i s case t h e r e can be no guarantee t h a t t h e process will converge. Such c a l c u l a t i o n s would b e o u t s i d e t h e scope of this book.

4.

Application of Boundary-Problem S o l u t j o n s of t h e Dynamics of Homopeneous S h e l l s t o t h e Construction o f-a -Homogeneous . S h e l l Approximately Equivalent t o a Layered S h e l l

A s can b e seen from t h e above, t h e p r i n c i p a l d i f f i c u l t y i n s o l v i n g t h e problem of t h e approximation of t h e Lagrange f u n c t i o n L of a l a y e r e d s h e l l by a Lagrange f u n c t i o n L++of a homogeneous s h e l l l i e s i n t h e indeterminacy of t h e v a r i a t i o n a l limits of t h e v a r i a b l e s of t h e f i e l d . This indeterminateness f o r c e s us i n many cases t o consider t h e approximation over Itnaturaltt i n t e r v a l s , determined by t h e requirement of t h e a p p l i c a b i l i t y of t h e l a w s of e l a s t i c i t y theory. Clearly, t h e i n t r o d u c t i o n of t h e n a t u r a l i n t e r v a l s of approximation reduces t h e accuracy of t h e results.

The continuous f i e l d v a r i a b l e s introduced by us y i e l d s t i l l another method of c o n s t r u c t i n g a homogeneous s h e l l approximately equal t o a l a y e r e d s h e l l which i s f r e e from t h e above handicap. Let us assume t h a t we have solved a c e r t a i n dynamic boundary problem of t h e t h e o r y of homogeneous s h e l l s , approximately corresponding i n boundary cond i t i o n s and i n loading c o n d i t i o n s t o t h e problem of v i b r a t i o n s of a l a y e r e d shell. W e s h a l l assume t h a t i n t h e i n i t i a l approximation t h e v a r i a b l e f i e l d s are determined from t h e s o l u t i o n of t h e problem for t h e homogeneous s h e l l . Applying t h e formula of transformation of t h e components of t e n s o r quant i t i e s , we f i n d according t o eqs.(8.17b) - (8.19) t h e approximate expression f o r t h e Lagrange f u n c t i o n L f o r t h e layered s h e l l with v a r i a b l e c o e f f i c /a6 i e n t s f o r t h e operators & ( f ) which a r e f u n c t i o n s of t h e coordinates of t h e b a s i c surf ace. The problem of determining t h e parameters A+, W-, p++, h*, and t h e coordinates of t h e new b a s i c s u r f a c e s i n t h e layered E z n d homogeneous s h e l l s reduce.s t o a c o n s i d e r a t i o n of t h e m i n i " standard d e v i a t i o n

where t h e i n t e r n a l i n t e g r a l extends over t h e a r e a S of t h e b a s i c surfaces, common t o t h e l a y e r e d s h e l l and i t s equivalent homogeneous s h e l l . The meaning of t h e remaining n o t a t i o n has been i n d i c a t e d above. The q u e s t i o n a g a i n reduces

247

t o t h e s o l u t i o n of a system of nonlinear a l g e b r a i c equations:

(8.39)
S e l e c t i n g t h e new b a s i c s u r f a c e s i n t h e l a y e r e d s h e l l and i t s approximately equivarent homogeneous s h e l l , w e can exclude from t h e i n t e g r a l I t h e two dominatillg terms, and then s e t up eqs.(8.39). Let us consider two elementary problems on t h e equilibrium of a c i r c u l a r closed c y l i n d r i c a l s h e l l of r a d i u s R, t o i l l u s t r a t e t h e l a t t e r method.

A s t h e . f i r s t example, l e t us consider t h e s u b c r i t i c a l &symmetric deformation of t h i s s h e l l due t o t h e l o n g i t u d i n a l compressive f o r c e s T uniformly d i s t r i b u t e d over t h e l i n e s of i n t e r s e c t i o n of t h e middle and f a c e surfaces. W e s h a l l assume t h a t t h e boundary conditions do not prevent radial d i l a t a t i o n of t h e tube.
Confining ourselves t o t h e approximation formulas of t h e c l a s s i c a l theory of s h e l l s and t o t h e notations given i n Sect.6, we f i n d (Bib1.23d) t h a t t h e deformed s t a t e of t h e s h e l l i s determined here by only a s i n g l e f u n c t i o n du dx* From t h e condition t h a t t h e annular f o r c e s s h a l l vanish, we o b t a i n

W e r e c a l l t h a t w and

du dx

belong t o t h e f i e l d v a r i a b l e s .

W e therefore

assume t h a t t h e s e q u a n t i t i e s a r e t h e same i n t h e l a y e r e d and homogeneous shells.

For metals such as s t e e l or aluminum, t h e Poisson constant d i f f e r s l i t t l e from 0 . 3 . Let us put v = 0.3 i n eq.(8.40).

I n t h e second elementary problem, l e t us assume t h a t t h e l o n g i t u d i n a l f o r c e s a r e zero. Then, with t h e o t h e r boundary conditions a r b i t r a r y , we f i n d t h a t t h e deformed s t a t e of t h e s h e l l i s described by t h e f u n c t i o n w. Yere,

Since eq.(8.19) i s s e t up i n a l o c a l Cartesian system of coordinates, l e t us consider t h e formulas f o r t h e d i r e c t and i n v e r s e t r a n s i t i o n s between t h e i n t e r n a l coordinates of t h e s h e l l and t h e l o c a l Cartesian system. These formul a s a r e of t h e following form:
X,

= x - x,; x2 = ( R - z ) sin

s - so ; x , = R - ( ( R - Z ) C O S - - - - - - , s--o.

(8.42a)

x' = x = x*

+ x,; xz

= s = so+

Rtan-

lx,; R - ~3

where xi a r e t h e l o c a l Cartesian coordinates and x? t h e i n t e r n a l coordinates of t h e s h e l l . I n this case, t h e i n t e r n a l c o o r d i n a t e s a r e determined by eqs.(6.1). The coordinates x i , and determine t h e p o s i t i o n of t h e o r i g i n of t h e coordinate b a s e s of t h e l o c a l C a r t e s i a n system on t h e middle s u r f a c e of the shell.
#

Making use of e q s . ( I , 5.17), l e t us express t h e displacement v e c t o r components i n t h e l o c a l C a r t e s i a n system of coordinates i n terms of t h e compone n t s u and w. W e obtain (8.42~)
On t h e b a s i s of eqs.(8.L2b) (8.L2c) and p u t t i n g x = & and s = after d i f f e r e n t i a t i o n , we f i n d a l l t h e q u a n t i t i e s c h a r a c t e r i z i n g t h e axisymmetric s t r e s s - s t r a i n s t a t e of t h e s h e l l e n t e r i n g i n t o eq. (8.19).

Since eqs.(8.40) (8.41) were obtained from t h e c l a s s i c a l theory of s h e l l s , we must make use of them i n determining t h e q u a n t i t i e s e{:) W e obtain

F u r t h e r , on t h e b a s i s of t h e Kirchhoff-Love hypotheses, we put

$1 =0 ( i = l , 2, 3; m=o, 1, 2).

/248

Of course, i t i s a l s o p o s s i b l e t o use equations t h a t do not r e l y on t h e Kirchhoff-Love hypotheses, b u t t h e i l l u s t r a t i v e c h a r a c t e r of t h e examples considered here and t h e general o b j e c t of c o n s t r u c t i n g a n approximate s o l u t i o n do not j u s t i f y t h e complications which t h e s e methods would involve.

Let us consider t h e f i r s t elementary problem. I n &symmetric compress i o n of a c y l i n d r i c a l s h e l l under t h e conditions of f r e e r a d i a l d i l a t a t i o n , t h e

flexural moments a r e zero, s o t h a t t h e d e r i v a t i v e

eqs.(8.63a) and (8.1+0) it follows t h a t

fi
dx2

a l s o vanishes.

From

Making use of eq.(8.19> and noting t h a t , i n s t a t i c problems, t h e Lagrange function L equals -n, we f i n d

Here, i t h a s been assumed t h a t h* = h. then i - % Z 2 2

~ 3 6 .

I f we a l s o assume t h a t v

? 0.3,

T h i s permits us t o f i n d p+f, b y equating t o zero t h e expres-

s i o n i n t h e braces i n eq.(8.45)7y. The equation determining w i l l be l i n - /249 ear. The r e s u l t will not depend on t h e f u n c t i o n u. Rejecting t h e terms t h a t depend on t h e r a t i o 2h : R and approximately s e t t i n g t h e f a c t o r (1 v)" as equal t o 0.5, we f i n d

3;

I n this case I reaches i t s exact lower boundary.

250

F I
The methods of f u r t h e r refinement i n this case a r e s o obvious t h a t we s h a l l not d i s c u s s them here. Equation ( 8 . 4 6 ) determines ~ $ by 6 means of a n ope r a t i o n c l o s e t o a simple averaging of t h e e l a s t i c constant of t h e l a y e r s over the thickness of t h e s h e l l . Consider t h e second elementary case. (111, 10.3a), and (8.43a), we f i n d Making use of e q s . ( I I I ,

10.la)

Here w e confined o u r s e l v e s t o t h e approximations corresponding t o t h e Kirchhoff-Love hypotheses. From eqs.(8.19) and (8.38) we f i n d

- (2h)

(1

+v2) + R '[ A, (2p)

-

- - (2h)ZZp"

I +2
-

A , (2p) -

Here, as above, we have assumed t h a t hz- = h.

I f we p u t

W$z

0.3 and, consequently,

A3c

3 =

IJ.*, and 2

a l s o neglect t h e

1250

t e r m s containing t h e r a t i o w : R, t h e n we f i n d from eq.(8.48), t o zero,

on equating I

W e omit a comparison of t h e numerical v a l u e s of eqs.(8.h8) and (8./+9).

IJ.4:-

determined from can be f u r t h e r r e f i n e d

Tne determination of p+:- on t h e b a s i s of eq.(8.48) by assuming, f o r example, t h a t w i s expressed by tnw 1

w=A,sinand, consequently, t h a t

25 2

Equating I t o zero and t a k i n g vJ+ = 0.3, we f i n d p ; : as a f u n c t i o n of t h e r a t i o m : 4. A t high v a l u e s of m we again a r r i v e a t eq. (8.49). Section 9. Construction of t h e Approximate S o l u t i o n t o Problems of t h e Dynamics of Layered S h e l l s . Application of t h e Method of Pert u r b a t i o n s and Nonremovable Errors
. /251

An analysis of t h e problem of t h e approximate replacement of a l a y e r e d s h e l l b y a homogeneous s h e l l shows t h a t t h i s leads t o considerable d i f f i c u l t i e s . The major d i f f i c u l t y i s t h e absence of means p e r m i t t i n g us t o e s t a b l i s h t h e universal r e g i o n of approximation of t h e Lagrange f u n c t i o n L of t h e layered s h e l l by t h e Lagrange f u n c t i o n L + : of t h e homogeneous s h e l l , which f r e q u e n t l y f o r c e s us t o t u r n t o t h e h a t u r a l l l region of approximation determined by t h e requirement t h a t a l l q u a n t i t i e s s h a l l vary w i t h i n t h e r e g i o n of a p p l i c a b i l i t y of t h e l i n e a r t h e o r y of e l a s t i c i t y . The absence of i n d i v i d u a l i z a t i o n of t h e v a r i a t i o n a l i n t e r v a l s of t h e princ i p a l v a r i a b l e s may decrease t h e accuracy of t h e results, p a r t i c u l a r l y when s p e c i a l problems a r e being considered. Even t h e absence of i n d i v i d u a l i z a t i o n of t h e v a r i a t i o n a l regions of t h e v a r i a b l e s m u s t b e considered as a preliminary hypothesis on t h e s t r e s s e d s t a t e of t h e s h e l l , which we have emphasized i n our d e r i v a t i o n of eqs.(8.31), (8.33) and of formulas (8.35a) (8.36). For this reason, t h e method i n d i c a t e d i n Sect.8.4 i s t n e most j u s t i f i e d . This method i s based on a preliminary c o n s i d e r a t i o n of t h e s o l u t i o n s of concrete boundary problems, although i n a number of cases t h i s method may l e a d t o unwieldy calc u l a t i o n s , s p e c i f i c a l l y i n t h e s o l u t i o n of systems of nonlinear a l g e b r a i c equat i o n s , as noted above.

Let u s consider now t h e g e n e r a l order of t h e approximate s o l u t i o n o f probl e m s of t h e dynamics of l a y e r e d s h e l l s . The f i r s t s t a g e of t h e s o l u t i o n c o n s i s t s i n t h e c o n s t r u c t i o n of a homoceneous s h e l l approximately e q u i v a l e n t t o t h e l a y e r e d s h e l l . Here i t i s most expedient t o s t a r t from t h e s o l u t i o n of t h e boundary problem f o r t h e homogeneous s h e l l . This boundary problem should b e s o s e l e c t e d t h a t , f o r t h e prescribed l o a d s and boundary conditions, i t s h a l l e x a c t l y or approximately correspond t o t h e problem of t h e dynamics of t h e l a y e r e d s h e l l . Then we must use t h e technique g i v e n i n Sect.8.4 f o r determining t h e q u a n t i t i e s c h a r a c t e r i z i n g t h e homogeneous s h e l l .
I f t h i s method involves unnecessary d i f f i c u l t i e s , we must use t h e procedure i n d i c a t e d i n Sect.8.3. I n determining t h e v a r i a t i o n a l i n t e r v a l s of t h e f i e l d v a r i a b l e s , we must a l s o attempt t o match t h e s e and all o t h e r assumptions, necessary f o r d e r i v i n g a weighted s u m of t h e form of eq.(8.31), w i t h t h e cond i t i o n s of t h e concrete problem of mechanics. Obviously, t h e f i r s t s t a g e of s o l u t i o n of t h e problem has p r a c t i c a l meaning only i f t h e d i f f e r e n c e between t h e values of t h e p h y s i c a l c o n s t a n t s of t h e l a y e r materials i s s u b s t a n t i a l . If t h e s e d i f f e r e n c e s are s l i g h t , we may confine o u r s e l v e s t o t h e weighted averagi n g of t h e p h y s i c a l constants over t h e t h i c k n e s s of t h e s h e l l , t a k i n g t h e weights equal t o t h e thicknesses of t h e corresponding s h e l l s . However, our a n a l y s i s shows t h a t such a method of determining t h e p h y s i c a l constants, /252 w h i l e q u i t e l o g i c a l a t f i r s t glance, does n o t correspond t o t h e optimum quad-

253

r a t i c approximation

Df

t h e f u n c t i o n L by t h e f u n c t i o n L * .

The f i r s t s t a g e of t h e approximate s o l u t i o n i s completed by determining, from t h e s o l u t i o n of t h e boundary problem, t h e f i e l d v a r i a b l e s i n t h e homogeneous s h e l l . These v a r i a b l e s , as noted a t t h e beginning of Sect.8, approximatel y r e p r e s e n t t h e f i e l d v a r i a b l e s of t h e l a y e r e d s h e l l . Then, using t h e s e f i e l d v a r i a b l e s , we c o n s t r u c t t h e f i e l d s of displacements, s t r a i n s , and s t r e s s e s i n t h e l a y e r e d s h e l l according t o eqs.(8.3) - (8.1ob). I n this case, we must first e s t a b l i s h t h e r e l a t i o n between t h e coordinate z+t i n t h e homogeneous s h e l l and t h e coordinate z i n t h e l a y e r e d s h e l l . This r e l a t i o n may b e taken i n t h e following form :

Equation (9.1) i s not connected with t h e equations of motion nor w i t h t h e boundary conditions. This equation w a s taken a r b i t r a r i l y by us, as t h e simpe may e v i d e n t l y make use of t h i s l e s t form of t h e r e l a t i o n between z and 2 9 . W a r b i t r a r i n e s s t o improve t h e approximate s o l u t i o n s sought. W e will n o t f u r t h e r i n v e s t i F a t e t h i s q u e s t i o n here. The c o n s t r u c t i o n of t h e f i e l d s of t h e p r i n c i p a l t e n s o r q u a n t i t i e s comp l e t e s t h e second sta5e of t h e approximate s o l u t i o n of t h e problem of t h e dynamics of l a y e r e d s h e l l s . W e r e c a l l t h a t t h e s o l u t i o n under c o n s i d e r a t i o n i s based on a n approximat i o n involving a f i n i t e segment of t h e frequency spectrum. By enlarging t h i s segment, we i n c r e a s e t h e ltweightlt of t h e terms depending on t h e k i n e t i c energy, as i s shown f o r example, by eq.(8.31>, and t h u s w o r s m t h e approximation of t h e q u a n t i t i e s depending on t h e p o t e n t i a l energy. These q u a n t i t i e s are t h e compone n t s o f t h e s t r a i n and s t r e s s t e n s o r s . T'nus, t h e approximate s o l u t i o n cons t r u c t e d by us, a s was t o b e expected, Will have o n l y l i m i t e d value. Let us r e f i n e t h e meaning of t h e approximate s o l u t i o n considered here, by comparing i t w i t h t h e exact s o l u t i o n of t h e l i n e a r theory of e l a s t i c i t y . Assume, for d e f i r d t e n e s s , t h a t t h e problem f o r t h e approximately e q d v a l e n t homogeqeous s h e l l has been solved by t h e f i r s t method of r e d u c t i o n considered i n Chapter 1 1 1 , i . e . , by t h e method of expansion i n tensor s e r i e s i n powers of z. Then, i n t h e homogeneous s h e l l , t h e conditions on t h e boundary surfaces will b e s a t i s f i e d as w e l l as t h e r e l a t i o n s expressing Hookels l a w and Saint-Venant's c o m p a t i b i l i t y conditions. The equations of motion and t h e equat i o n s of t h e contour s u r f a c e s will b e approximately s a t i s f i e d . The r e l a t i v e accuracy of s a t i s f a c t i o n of t h e boundary conditions on t h e contour s u r f a c e s will b e lower t h a n t h e r e l a t i v e accuracy of r e p r e s e n t a t i o n of t h e components of t h e displacement v e c t o r and of t h e s t r a i n t e n s o r by t h e segments of t h e /253 tensor series.
*

The f i e l d s of displacements and s t r e s s e s i n t h e layered s h e l l s , constructed i n t h e second stage, satisfy t h e conditions on t h e boundary s u r f a c e s , t h e conditions on t h e i n t e r f a c e s between t h e l a y e r s , t h e equations r e s u l t i n g from

254

IF

t h e generalized Hooke's l a w , and Saint-Venant's c o m p a t i b i l i t y conditions. The equations of motion and t h e boundary conditions on t h e contour surfaces are approximately s a t i s f i e d . The e r r o r i n t h e s a t i s f a c t i o n of t h e equations of motion and of t h e c o n d i t i o n s on t h e contour s u r f a c e s dl1 depend, i n t h i s case, n o t o n l y on t h e r e j e c t e d terms of t h e s e r i e s r e p r e s e n t i n g t h e displacement vect o r components, b u t a l s o on t h e d i f f e r e n c e s between t h e p h y s i c a l constants of t h e layer materials and on t h e d i f f e r e n c e between t h e t h i c k n e s s of t h e l a y e r e d s h e l l 2h and t h e t h i c k n e s s of t h e e q u i v a l e n t s h e l l 2h+:-. T'le e r r o r can be decreased by applying t h e method of pertuybations. I n f a c t , knowing approximately t h e s t r e s s e s on t h e boundaries of t h e l a y e r s , we ,can now consider s e p a r a t e l y t h e motion of each l a y e r . Each or' t h e l a y e r s performs a motion under t h e a c t i o n of l o a d s on t h e s u r f a c e and of quasi-body f o r c e s , which can b e found by s u b s t i t u t i n ? t h e approximately determined components of t h e displacement vector i n t o t h e equations of motion (11, 5.5a or 5.5b). W e can t h e n apply one of t h e systems of equations of motion of homogeneous s h e l l s , considered i n Chapter 1 1 1 , t o each l a y e r s e p a r a t e l y . This dl1 permit t o e l i m i n a t e p a r t of t h e e r r o r a r i s i n g as a r e s u l t of t h e d i f f e r e n c e s . i n t h e 1ayeTs of t h e s h e l l and the quantibetween t h e q u a n t i t i e s h, P , h , P t i e s h+:, p+, A+:-, pi': i n t h e homogeneous s h e l l . There s t i l l remains, however, t h e e r r o r depending on t h e approximate determination of t h e s u r f a c e f o r c e s on %heboundaries of t h e l a y e r s of t h i s s h e l l . T h i s e r r o r i s connected with t h e f a c t t h a t t h e f u n c t i o n L+ only approximately r e p r e s e n t s t h e f u n c t i o n L. For this reason, t h e "exact1' subdivision of t h e gener.al problem of motion of a l a y e r e d s h e l l i n t o i s o l a t e d problems of motion of i t s layers cannot b e c a r r i e d o u t , and t h e l a t t e r e r r o r w i l l b e irremovable. Further s t u d i e s , going beyond t h e scope of t h i s book, must obviously tent e r on t h e search f o r means of decreasing t h i s irremovable e r r o r . S e c t i o n 10. Application of Optimum Quadratic Approximations t o t h e Problem of Reduction of t h e Three-Dimensional Problem of t h e E l a s t i c i t ; y Theory t o t h e Two-Dimensional Problem

The methods used by u s i n s o l v i n g t h e problem of c o n s t r u c t i n g a system which, according t o some c r i t e r i o n , i s approximately e q u i v a l e n t t o t h e pres c r i b e d system are a l s o a p p l i c a b l e t o t h e problems considered i n Chapter 1 1 1 . W e have s t a t e d above t h a t t h e r e d u c t i o n of t h e three-dimensional problem of t h e theory of e l a s t i c i t y t o a two-dimensional problem of t h e theory of s h e l l s can be regarded as t h e c o n s t r u c t i o n of a system approximately equivalent t o /254 a three-dimensional e l a s t i c body. The o b j e c t s of t h e above-selected approximation w e r e t h e p o t e n t i a l s t r a i n energy and t'le Lagrange f u n c t i o n of t h e e l a s t i c body, using systems of v a r i a b l e s t h a t y i e l d t h e e x p l i c i t a n a l y t i c express i o n s of t h e approximated and approximating f u n c t i o n s . Tne v a r i a b l e s used previously are u n s u i t a b l e f o r s o l u t i o n of t h e r e d u c t i o n problem. This f o r c e s us t o abandon t h e p o t e n t i a l s t r a i n energy and t h e Lagrange f u n c t i o n of an e l a s t i c body as o b j e c t s of approximation. As t h e o b j e c t s of $he approximations, l e t us s e l e c t p r e s c r i b e d body and s u r f a c e f o r c e s and pres c r i b e d displacements on c e r t a i n p o r t i o n s of t h e s u r f a c e of t h e body.
25 5

W e s h a l l confine ourselves t o a d i s c u s s i o n of t h e g e n e r a l p r i n c i p l e s of t h e proposed method and not go i n t o t h e details. W e W i l l s t a r t from t h e nonl i n e a r Lam6 equations i n t h e form of (11, 7.6), assuming f o r s i m p l i c i t y t h a t t h e system of coordinate axes of t h e undeformed b a s i c s u r f a c e of t h e s h e l l coi n c i d e s with i t s l i n e s of curvature. Let us s e l e c t two coordinate l i n e s as t h e l i n e s of o r i g i n of t h e coordie s h a l l d e f i n e t h e p o s i t i o n of an a r b i t r a r y p o i n t M on t h e b a s i c n a t e net. W s u r f a c e by i t s a r c coordinates s i , equal t o t h e a b s o l u t e v a l u e s of t h e d i s - , t a n c e s of t h e p o i n t M from t h e l i n e s of o r i g i n . These d i s t a n c e s are measured along t h e coordinate l i n e s passing through t h e p o i n t M from p o i n t M t o p o i n t s The si N, of t h e i r i n t e r s e c t i o n With t h e o r i g i n l i n e s of t h e coordinate net*. coordinates are connected with t h e XJ coordinates by means of c u r v i l i n e a r i n t e g r a l s taken along t h e coordinate axes.
2

( 1 0 . 1 )
(do not sum over i t ) . With a coordinate system s e l e c t e d i n t h i s manner, t h e metric t e n s o r on t h e b a s i c s u r f a c e W i l l have t h e following components:

(10.2)
where 6 :

i s t h e Kronecker d e l t a .

It f o l l o w s from eq.(II, 7.6) t h a t , i n t h e system of coordinates s e l e c t e d by us, t h e covariant components of t h e body f o r c e s may be r e p r e s e n t e d i n t h e following form :

I n eqs.(l0.3), M P r r ( s J , z ) and z ) a r e f u n c t i o n s of t h e coor/255 d i n a t e s sJ of t h e undktormed b a s i c s u r f a c e and of t h e coordinate z, depending on t h e C h r i s t o f f e l symbols ril, and on t h e i r f i r s t d e r i v a t i v e s w i t h r e s p e c t t o si and z ( j = 1 , 2). The f u n c t i o n s @ i are components of t h e a d d i t i o n a l body f o r c e s considered These functions, as r e s u l t s from (11, 7.4), have t h e followi n (11, Sect.7).

q.l(sJ,

X- For d e t a i l s , s e e (Bibl.6,

pp.101-102).

256

i n g composition:

( 1 0 . 4 )
Here, all i n d i c e s except j t a k e t h e values 1, 2, 3 . W e s h a l l not w r i t e o u t t h e expressions for t h e s e f u n c t i o n s i n expanded form, s i n c e we i n t e n d h e r e a f t e r t o confige ourselves o n l y t o a d i s c u s s i o n of t h e g e n e r a l p r i n c i p l e s of constructi n g t h e system of equations of t h e t h e o r y of s h e l l s by means of t h e method considered here. I n e x a c t l y t h e same manner, we can d e r i v e t h e components of t h e s u r f a c e f o r c e s on t h e boundary s u r f a c e s of t h e s h e l l and on i t s contour surf aces.
Before considering t n e conditions on t h e boundary s u r f a c e s , l e t u s assume t h a t t h e s u r f a c e f o r c e s a c t i n g on them and t h e stress t e n s o r components corresponding t o them have undergone p a r a l l e l displacement t o t h e b a s i c s u r f a c e Making use of (11, 8.13) we f i n d t h a t t h e followi n t h e sense of Levi-Civita. i n 2 conditions a r e s a t i s f i e d on t h e boundary s u r f a c e s of t h e shell:

where t h e s c a l a r $o i s determined from (11, 8.12); t h e s t r a s s t e n s o r compone n t s displaced t o t h e b a s i c s u r f a c e a r e dsnoted as i n Chapter 1 1 1 ; the sign (+) corresponds t o a bourldary s u r f a c e on. which t h e d i r e c t i o n of t h e u n i t v e c t o r of t h e e x t e r n a l normal coincides with t h e d i r e c t i o n of t h e v e c t o r & on t h e undeformed b a s i c s u r f a c e , while t h e s i g n (-) corresponds t o a boundary s u r f a c e on which t h e s e d i r e c t i o n s are opposite.

I n eqs.(l0.5)

t h e coordirlate z has a f i x e d value.

Let

T!iese equations determine t h e p o s i t i o n of t h e boundary s u r f a c e s f o r a e s h a l l hereafter p r e s c r i b e d p o s i t i o n g f t h e b a s i c s u r f a c e within t h e s h e l l . W assume t h a t h l and h2 a r e constants. On t h e contour s u r f a c e s , t h e follow-in.3 r e l a t i o n s are s a t i s f i e d :

The stress t e n s o r components are connected with t h e displacement v e c t o r components by t h e equations r e s u l t i n g from Yookes l a w (11, 4.3):

/256

25 7

L : 3 3 are expressed i n terms of t h e C h r i s t o f f e l The f u n c t i o n s G i , and 933 a r e nonlinear i n t h e displacement v e c t o r components and t h e i r derivatives w i t h r e s p e c t t o t h e coo r d i n a t e s sJ and z of terms e n t e r i n g i n t o Hooke's l a w (11, 4.3) whose composiThese f u n c t i o n s Lion may b e e s t a b l i s h e d , f o r i n s t a n c e , from e q . ( I I , 7.2). have a similar meaning i n t h e case of p h y s i c a l n o n l i n e a r i t y , i.e., of a Hookers l a w determined by (11, L . 7 ) .
where t h e f u n c t i o n s Lyi J

symbols

rp, j , rYi3,

, LYi3,

1 . Approximate Zxpressions of t h e Displacement Vector Components Section 1 and t h e Equations of Motion of t h e Shell
Let us assume t h a t t h e displacement v e c t o r components, displaced t o t h e b a s i c surface, can b e represented b y approximation formulas similar t o . t h o s e considered i n Chapter 1 1 1 .
N

(11.1)
m-:O

where &c

(2)

a r e prescribed f u n c t i o n s

(above, we mostly used cpm ( z ) =

t3

m!

while t h e c o e f f i c i e n t s uj")

are unknown f u n c t i o n s

be determined.

Tie right-hand s i d e of e q . ( l l . l ) c o n t a i n s a f i n i t e sum such that, h e r e as above, t h e number of degrees of freedom of t h e s h e l l i n t h e d i r e c t i o n of t h e coordinate x3 = z i s r e s t r i c t e d . S u b s t i t u t i n g e q s . ( l l . l ) i n t o eq.(10.3), we f i n d t h e f o l l o w b g approximate expressions f o r t h e components of t h e body forces :

The f u n c t i o n s Yi and Y 3 are t h e r e s u l t s of s u b s t i t u t i o n , i n t o t h e f u n c t i o n s ;i and +3,'of t h e appro,dmate eqs.(ll.l) f'or t'?c displacement vect.3r conponents.If we confine ourselves t o studying weakly nonlinear problems, then t h e funct i o n s Y t and Y3 W i l l b e polynomials of t h e functions cpPmand t h e i r d e r i v a t i v e s . L e t us now consider t h e approximate expressions f o r t h e s u r f a c e f o r c e s on t h e boundary s u r f a c e s of t h e s h e l l . Ikkirlg use of eqs.(l0.5) and (10.8a) (10.8c), we o b t a i n
N
1

=f
m =O

{ 'Pm (z('-1) [pdi l

L53

(si, z(

1) I&*']

+
(11. j a )

+ pq7;n (+))

up)+ 'Vi3 (si, z ' + ) , d k U W , It;");

P

The q u a n t i t i e s z (*) a r e determined from eqs. (10.6).

The fu_rlctions

c o n t a i n t h e s e t of nonlinear terms e n t e r i n g i n t o t h e s u r f a c e f o r c e s X+kli and X++(*)3. Let us set up t h e f u n c t i o n a l :

r=

fJ { f;(p,
0 (SI
--hi

- P F ; ) ~ (pF, - pF;y

+ (pF, - p F 3 (1 - k,z)x

259

I IIIII..

1111 I

111

II 1 1 1 1 1 1 1 1 1 1

I , ,

, , .,

. .,.... ..... -_-.

,, ,,

...~

+
t h e area of t h e b a s i c s u r f a c e .

I.&-)

--

x;-)
3)*

c l s dt.

(11.4)

where t, i s a n a r b i t r a r y i n s t a n t of t h e and t h e i n t e g r a l

1extends over

/258

(SI The o t h e r symbols are known from Chapter 1 1 1 .

W e shall determine t h e generalized coordinates up'.) of a system r e p l a c i n g t h e s h e l l from t h e c o n d i t i o n s of t h e minim?lm of t h e f u n c t i o n a l I. It i s w e l l known t h a t here we may u s e t h e d i r e c t methods, f o r example t h e R i t z method and t h e c l a s s i c a l method, by s e t t i n g up t h e Euler-Lagrange-Ostrogradskiy equations*. W e s h a l l not consider t h e IXtz method b u t d i s c u s s t h e Euler-LagrangeOstrogradskiy equations. L e t us i n t r o d u c e t h e n o t a t i o n :

The Euler-Lagrange-Ostrogradskiy equations can be represented i n t h e f o l lowin,? form:

(11.6a)

Cf., f o r example, V.I.Smirnov, Gostekhizdat, 1951.

Course i n Higher Mathematics, Vol.IV,

260

W e s h a l l consider t h e left-hand rivative formula

__

s i d e of e a . ( l l . 6 a )

a s t h e f u n c t i o n a l de-

aw' .

Then t h e system of eq.(ll.6a)

can be replaced by t h e s h o r t

Z U ; "

2 W -- 0.
6 I p

(11.6b)

Equations (11.6a) o r (11.6b) approximately determine t h e motion of an element of t h e s h e l l . Making use of eq.(11.5) for t h e f u n c t i o n IJ, l e t US put e q s . ( l l . 6 a ) i n t o t h e f o l l o d n g f o r m :

/259

or

( p = 1, 2, 3; m=0, 1, 2,

... , N ) .

(11.To)
separating t h e i r l i n e a r

Let us now consider i n more d e t a i l e q s . ( l l . % ) ,

261

(i, j,k = 1, 2; p = 1, 2, 3).

Substituting eqs.(ll.2a) (11.3b) and (11.8a) (11.9d) i n t o eqs.(ll.?b), we o b t a i n t h e equations of motion of an element of t h e s h e l l i n expanded form. m e n t h e l i n e a r p a r t s of t h e s e equations, however,will b e highly cumbersome. The equations of moti2n can be somewhat s i m p l i f i e d by changing t h e metric on t h e coordinate axis x , and by a n appropriate s e l e c t i o n of t h e functions c p . (2). Let us assume t h a t t h e b a s i c surface of t h e s h e l l coincides with i t s middle surface, i . e . , t h a t h, = h2 = h. Let us p u t

262

Hence, we f i n d

Selectin,? t h e constant C such t h a t , on v a r i a t i o n of z over t h e i n t e r v a l (-h, +h), t h e v a r i a b l e 5 v a r i e s over t h e symmetric i n t e r v a l ( 4 , &), we obtain

From eq.(a) r e s u l t s

/261
( 1 1 . loa)

where d s i s a n element of a r c of t h e t h i r d coordinate i n t e r s e c t i n g t h e undeformed b a s i c s u r f a c e a t a r i g h t angle, and

g3a=[I - ( ( k , + k , ) z + k , k , z ? ] - ' .

(11.1Ob)

I n t h e l a s t r e l a t i o n , z must b e regarded as a f u n c t i o n of 6 determined by e q . ( c > . O f course, f o r a s u f f i c i e n t l y t h i n s h e l l , a t small values of t h e product z k i , we can approximately put

Cz z 2.

(11.1oc)

This The i n t r o d u c t i o n of t h e v a r i 2 b l e 5 somewhat s i m p l i f i e s eq.(ll.'P). transformation of t h e coordinate x = z does not a f f e c t t h e form of t h e notat i o n of t h e o r i g i n a l equations (11.2a) - (11.3b), although t h e meaning of t h e v a r i a b l e c o e f f i c i e n t s of t h e functions up"' and t h e f i r s t d e r i v a t i v e s W i l l d i f f e r from t h e i r meanings i n t h e o r i g i n a l system of coordinates+$. I n p a r t i c ular, t h e q u a n t i t i e s c p . ( 5 " ) ) will b e f u n c t i o n s of t h e coordinates s3 , i . e . , t h e s h e l l of constant thickness, on i n t r o d u c t i o n of t h e v a r i a b l e 5, W i l l be, i n a manner of speaking, transformed i n t o a s h e l l of quasi-variable thickness. It i s obvious t h a t , a t t h e i n t e r n a l p o i n t s of t h e s h e l l , t h e coordinate 5 , on
-x- The i n t r o d u c t i o n of a new metric on t h e t h i r d coordinate ax5s would permit simplifying t h e n o t a t i o n for c e r t a i n of t h e equations considered i n Chapter 1 1 1 . I n this way, t h e t e r m s e x p l i c i t l y containing t h e curvature k , of t h e b a s i c s u r f a c e would be eliminated from t h e s e equations.

d i f f e r e n t i a t i o n , must b e regarded as a n independent variable. W e W i l l n o t ana l y z e t h e q u e s t i o n of t h e b e s t method of s e l e c t i n g t h e f u n c t i o n s y . ( ~ ) , which now must be r e p l a c e d i n t h e equations by t h e f u n c t i o n s cp, (5). I n v i e w of t h e r e l a t i v e l y small t h i c k n e s s of t h e s h e l l , t h e choice of cp, (L) i n t h e form of a power monomial 5 ' has considerable advantages. This case w a s e s s e n t i a l l y considered by us i n Chapter 1 1 1 . Here we s h a l l use a Let us p u t d i f f e r e n t p a r t i a l s e l e c t i o n of t h e f u n c t i o n s T.(<).
'Pzn = cos
ttxc -

'PZn+l = s i n -.

nnc I

( 1 1 . 1 1 )

Consequently,
'po= 1 ; (p, = 0;

.rrl ; cp3 = sin - ; cp2 = cos -

c.

(p4 = cos

2nr

etc.

W e r e c a l l t h a t t h e q u a n t i t y 4, i s a f u n c t i o n of t h e coordinates sJ The system of f u n c t i o n s cp. ( c ) introduced by eqs. ( 1 1 . 1 1 ) , i s orthogonal over t h e The same p r o p e r t v i s possessed by t h e d e r i v a t i v e s cpJ(5) i n t e r v a l ( 4 ,2 , ) . and 9; (cj.

W e a l s o note t h e following r e l a t i o n s which r e s u l t from e q s . ( l l . l l ) :

/262

(112 )
Consider now t h e equations of motion r e s u l t i n g from e q s . ( l l . n ) . A s will b e s e e n from e q s . ( l l . % ) , t h i s system of equations i s a n i n f i n i t e system c o n s i s t in,? of nonlinear d i f f e r e n t i a l equations of t h e f o u r t h order. It i s c l e a r t h a t a d i r e c t a p p l i c a t i o n of such a system i s unpromising. For t h i s reason, we must d e f i n e t h e c o n d i t i o n s m d e r which t h i s system i s resolved i n t o s e p a r a t e subsystems containing a f i n i t e number of equations. The i n f i n i t e system of equations r e s u l t i n g from e q s . ( l l . n ) can b e decomposed i f , i n s e t k i n g up t h e Euler-Lagrange-Ostrogradskiy equations, we e l i m i n a t e t h e n e c e s s i t y of v a r i a t i o n of t h e nonlinear terms and of t h e terms w i t h t h e c o e f f i c i e n t s Pt y , , N:q e n t e r i n g i n t o t h e composition of t h e body f o r c e s pFGq, and a l s o eliminate t h e v a r i a t i o n of t h e s u r f a c e f o r c e s X ? :( t ) P I n order t o eliminate t h e v a r i a t i o n of terms with t h e c o e f f i c i e n t s and NY, as w e l l as t h e nonlinear terms, we must r e p l a c e , i n t h e s e terms, t h e functions b y expansions of t h e form

up"'

26h.

(11.13)
where a,$) are constant c o e f f i c i e n t s t o be determined, and represent a system of f u n c t i o n s s a t i s f y i n g t h e kinematic boundary conditions and t h e cond i t i o n s of completeness, ensuring t h e p o s s i b i l i t y of approximation of solut i o n s of t h e equations of t h e theory of s h e l l s by expressions of t h e form of The second method of appro,dmate r e p r e s e n t a t i o n of t h e s e terms, eq. (ll.l3)*. based on t h e method of successive approximations, will b e given below.

cg)

To make i t unnecessary t o vary t h e s u r f a c e f o r c e s XTxl , l e t us employ t h e following method: L e t us r e s o l v e t h e components of t h e prescribed body i n t o t h e components k t i ) q X ( f ) q and(1 - k t + l q ) X c k ) q . The f i r s t forces X summand w i 1 b e regarded as a component of t h e body f o r c e S ( 5 + L)k(,),X(+),, A ) i s t h e d e l t a function. These body f o r c e s will b e a s s o c i a t e d where 6 ( L with t h e p r e s c r i b e d forces. Let u s a l s o a s s o c i a t e with t h e body f o r c e s t h e /263 expressed by terms of t h e body f o r c e s pF# containing t h e components eqs.(ll.l3). L e t u s d e n o t e l t h e new components of t h e body f o r c e s , f o r b r e v i t y , by t h e symbol pR,, b u t remember t h a t t h e y depend on t h e c o e f f i c i e n t s a$) Then, i n s t e a d of t h e f u r c t i o n a l I expressed by eq.(11.4), we o b t a i n

(=Y

1, =
0 ( S I

2W dS dt,

( 1 1 . 1 4 )

where, i n d i s t i n c t i o n t o eq. (11.3 1 ,

where p% are components of t h e body f o r c e s which can be determined from eqs.(ll.Za) (11.2b) i f t h e terms e n t e r i n g w i t h o p p o s i t e s i 5 n s i n t o t h e quant i t i e s p R , a r e cancelled out from t h e right-hand s i d e of t h e s e equations. Equating t o z e r o i n eq.(11.15) t h e terms containing t h e surface f o r c e s , w e find

?t

( 1 1 . 1 6 )

t i c i t y , where t h e R i t x method i s s e t f o r t h .

f a c e s of t h i s s h e l l , t h e n t h e q u a n t i t i e s k t 5 , q w i l l vanish. I n accordance wi-th t h e Kirchhoff-Love hypotheses, t h e components X*+,, will vanish and t h e of t h e nonvanishing X(*)q will b i - e q u a l t o unity. coefficients k (.) Tne conditinns of a minimum f o r t h e f u n c t i o n a l I, l e a d t o t h e N e r Lagrange-Ostroqradskiy system of equations, which i n t h i s case w i l l be of t h e following form :

( y - 1, 2, 3; m = O , 1,.!2,

- .-

IL).

I n a d d i t i o n t o e q s . ( l l . l 7 ) , we must a l s o b e a r i n mind t h e conditions deThese c o n d i t i o n s have t h e termining t h e c o e f f i c i e n t s a$$ i n eqs. (ll.13). following form, w e l l known from t h e R i t z method:

(11.18)
Thus, t h e complete sys: ~ 7 e f quations of t h e t h e o r y of s h e l l s now consists of e q s . ( l l . l 6 ) , ( 1 1 . 1 7 ) , and (1-1*18). This system combines t h e c l a s s i c a l Euler-Lagrange-Ostroyradskiy equations w i t h t h e equations r e s u l t i n g from the R i t z method. L e t u s now consider e q s . ( l l . l 7 ) . W e s h a l l f i r s t introduce a n abbreviated n o t a t i o n f o r s e v e r a l d i f f e r e n t i a l o p e r a t o r s :
'

266

(11.19a)

(11.19b)

The o p e r a t o r s P J ' ) ( q = 1, 2, 3 ) depend on t h e curvature k, of t h e b a s i c surface, s i n c e t h e s e c u r v a t u r e s e n t e r i n t o t h e parameter 1. However, t h i s dependence can be termed weak, e s p e c i a l l y f o r t h e case of t h i n s h e l l s with t h e 0.01. The operators Pp") are apparently c l o s e t o t h e operar a t i o 2h : Rmx t o r s describing t h e l i n e a r s t r e s s - s t r a i n s t a t e of a given p l a t e . Further, l e t u s denote t h e operators depending on t h e curvature of t h e b a s i c s u r f a c e of t h e s h e l l and on t h e coordinate 5 as follows:

(11.2Ob)
Then, b e a r i n g i n mind e q s . ( l l . l 2 ) , we f i n d

26 7

21

Let us continue our consideration of t h e q u a n t i t i e s entering i n t o Again bearing i n mind e q s . ( l l . l 2 ) , we f i n d eqs.(ll.l7).

(U. 22a )

(11.22b)

(11.232)

/266

(11 22d)
(11.22e)

( 11.22f)

( 11.22g)
( i , j , k = l , 2; p , q = 1 ,
2 , 3).

(11.22g) y i e l d t h e meaning of t h e abbreviated notaEquations (11.2La) t i o n [ e q . ( l l . l 7 ) ] for t h e system of equations of motion of a n element of t h e s h e l l . For this, i t i s s u f f i c i e n t t o give t o t h e index p t h e values 1, 2, 3 , e obtain t h e and t o take m successively equal t o 2n and 2h + 1. I n this way, w system of equations

268

P;?
(nt=O,

( 0 ) =0

-.

1, 2,

.. . ,

N).

(11.23)

where Pp'.) are t h e operators ?hose meaning i s given by e q s . ( l l . l q a ) (ll.19d). The components of t h e v e c t o r v a r e determined i n t h e following manner:

(ll.2.4)
(h=O,
1, 2, . ..; r = l , 2, 3).

Equations (11.23) c o n s t i t u t e , a t N a, an i n f i n i t e system of fourthorder equations. If no components of t h e multi-dimensional vector 5 e n t e r i n t h e f u n c t i o n R , , t h e n t h e system of equations (11.2%) i s resolved i n t o autonomous subsystems, each containing six equations. T h i s r e s o l u t i Q n w i l l take place d i r e c t l y only i n problems of t h e mechanics of p l a t e s o r s h e l l s w i t h zero curvatures k i of t h e b a s i c surface.
+

I n t h e remaining cases, no decomposition takes @ace. It i s t h e r e f o r e necessary t o determine t h e components of t h e v e c t o r u entering i n t o t h e quanI n t h i s case, as a l r e a d y noted, eqs.(11.23) Will t i t i e s PR, by e q s . ( l l . l 3 ) . contain indeterminate c o e f f i c i e n t s a,$) To determine t h e s e q u a n t i t i e s we /267 must use e q s . ( l l . l 8 ) , which a r e a l g e b r a i c equations t h a t a r e l i n e a r f o r t h e l i n e a r statement of t h e problem and nonlinear i n t h e general case. W e Will not e c a l l a t t e n t i o n only t o one of Tive t h e expanded form of t h e s e equations. W t h e f e a t u r e s of t h e method developed by us.

This method i s based on t h e combined use of eqs.(ll.23) - ( 1 1 . 2 4 ) determining t h e wanted functions as s o l u t i o n s of a c e r t a i n boundary problem, and on e q s . ( l l . l 8 ) which, when taken to,:ether With eqs. (11.13), y i e l d t h e approximate a n a l y t i c form of t h e s o l u t i o n . The method of combination of eqs.(ll.23) (11.21) and eq.(11.18) t o g e t h e r w i t h eq.(11.13) depends on t h e scope of t h e problem of mechanics involved. I n any p a r t i c u l a r approach t o i n v e s t i g a t i o n of t h e problem, however, we w i l l g b t a i n an approximate r e p r e s e n t a t i o n of t h e mot i o n of a s h e l l element, which d i f f e r s from s o l u t i o n s t h a t a r e analogous b u t a r e obtained from other equations i n t h a t , i n t h i s case, t h e necessary condit i o n s of t h e minimum of t h e quadratic d e v i a t i o n from t h e s o l u t i o n s of t h e three-dimensional theory of e l a s t i c i t y will be s a t i s f i e d .

'de note f i n a l l y t h a t t h e condition of minimum of t h e f u n c t i o n a l I, determined by e q . ( l l . l h ) , a c t u a l l y 'coincides with t h e Gauss p r i n c i p l e of least cons t r a i n t Z of t h e system, i f t h e c o n s t r a i n t Z i s averaged over t h e time i n t e r v a l (0, t,

>.

269

Section 1 2 . Boundary Conditions. Various Versions of theSolution of the General Problem of the D e c s of Shells. Initial Conditions
~~~~~ ~~

1 . Remarks on Boundary Conditions

~

Let us consider the boundary conditions that terminate the statement of the problem and briefly discuss the various versions of its solution.

To obtain a system of boundary conditions, let us find, on the contour surface of the shell, the components of the displacement vector and stress 1 1 . 1 ) . Then, considvector resulting from the approximate representations ( ering the shell as a three-dimensional body, let us set up the boundary conditions on the contour surface of the she61 in accordance with the statement of the three-dimensional problems of the theory of elasticity considered in Chap1 . Finally, using the method of least squares, let us require that the ter 1 quantities resulting from the approximate representations ( 1 1 . 1 ) shall satisfy, on the contour surface of the shell, the requirement of the least-square deviation from the corresponding functions prescribed on the contour surface in the formulation of the three-dimensional boundary problem of the theory of elasticity.
This program requires consideration of the following functionals:

/26r
(12.la)

1.

4-1

. ,
(12. lb)

, is that part of where C is the contour of the middle surface of the shell, C , are the comthe contour on which the displacements are prescribed, % and f ponents of the displacement and stress vectors prescribed on the respective parts of the contour surface, and u-zq and f z , are the approximate expressions of these components determined by eqs.(ll.l). For definiteness we may assume that the functions cp. ((;) are expressed by eqs. ( 1 1 . 1 1 ) . Let us denote

The conditions of a m i n i for the functionals Iz and I j lead to the following conditions of the middle surface of the shell:

( 12.3a )

or

( p = 1, 2. 3;

Nt : = 0, 1,

2,

. .. , N). aSJ
A "

(12.3~) must b expressed

W e re

1 1 t h a t , on t h e contour C y t h e derivatives

i n terms of t h e d e r i v a t i v e s along t h e tangent and along t h e p r i n c i p a l normal t o t h e contour C. W e s h a l l n o t consider t h e c o n d i t i o n s (12.3b) - ( 1 2 . 3 ~ )i n t h e expanded form and confine ourselves t o b r i e f remarks on t h e statement of the boundary problem of t h e dynamics of s h e l l s considered here.
a) W e cannot d i r e c t l y assert t h a t t h e boundary conditions (12.3b)-(12.3c) are natural f o r t h e v a r i a t i o n a l problem considered i n Sect.11, a t l e a s t not /269 e must, t h e r e f o r e , with r e s p e c t t o t h e method by which t h e y w e r e obtained. W d e f i n e t h e i r connection w i t h t h e natural boundary conditions. W e Will do t h i s later i n the text.

b ) Naturally t h e question arises as t o t h e e x i s t e n c e and uniqueness of a s o l u t i o n of t h e boundary problem under consideration.

2.

On t h e Existence and Uniqueness of S o l u t i o n s of t h e Boundary Problem Posed

It i s w e l l known t h a t theorems f o r t h e uniqueness of s o l u t i o n s of l i n e a r s t a t i c and dynamic three-dimensional problems of t h e theory of e l a s t i c i t y f o r f i n i t e regions have long s i n c e been proved+*. Theorems of t h e e x i s t e n c e of sol u t i o n s have been proved f o r three-dimensional l i n e a r problems of t h e s t a t i c s of an e l a s t i c problem and a l s o f o r a number of problems of dynamics%+.
Since we are h e r e i n v e s t i g a t i n g only t h e results of t h e approximation of t h e equations of t h e three-dimensional i n t e r n a l problem of t h e theory of elast i c i t y , l e a d i n g t o equations of t h e e l a s t o d p a m i c s of s h e l l s , we may i n advance assume w i t h considerable c e r t a i n t y t h a t t h e s e theorems on t h e e x i s t e n c e and uniqueness of s o l u t i o n s can a l s o be extended t o t h e boundary problems under consideration. O f course, t h i s i s merely a working hypothesis.

+$

C f . f o r i n s t a n c e , E.Trefftz,

Mathematical Theory of E l a s t i c i t y , ONTI, 1934

++%

See preceding f o o t n o t e and a l s o V.D.Kupdraze, Boundary Problems of t h e Theory of V i b r a t i o n s and I n t e g r a l Equations, Gostekhizdat, 1950

We will not f u r t h e r d i s c u s s t h e proofs of t h e s e theorems with r e s p e c t t o t h e theory of s h e l l s ; confining ourselves t o t h e following remark: If we do ( 1 2 . 3 ~ )f o r t h e not require s a t i s f a c t i o n of t h e necessary conditions (12.3b) j , then i t can be a s s e r t e d that t h e r e exextremum of t h e f u n c t i o n a l s I2 and I i s t s a s o l u t i o n of eqs.(ll.l8) and (11*23)f o r which t h e s u m I , + I3 will have a minimum. The q u e s t i o n of t h e uniqueness of such a s o l u t i o n remains open.

3. Natural Boundary Conditions

There i s no d i f f i c u l t y i n determining t h e n a t u r a l boundary conditions f o r t h e problem of t h e extremum of t h e f u n c t i o n a l I,, expressed by eq.(11.14). However, we will not i n v e s t i g a t e t h e s e conditions, r e c a l l i n g t h a t t h e boundary conditions f o r t h e three-dimensional problem of t h e t h e o r y of e l a s t i c i t y conThe conditions (12.3b) - ( 1 2 . 3 ~ )found by us s i d e r e d i n c h a p t - I I a r e natural*. a r e t h e r e s u l t s of t h e requirement of least-square d e v i a t i o n s of t h e s o l u t i o n s of t h e problem of t h e elastodynamics of s h e l l s from t h e natural boundary cond i t i o n s of t h e three-dimensional problem of t h e theory of e l a s t i c i t y on t h e /270 contour surface of t h e s h e l l . This requirement i s i n agreement w i t h t h e fundamental p r i n c i p l e of constructing t h e equations of t h e elastodynamics of s h e l l s I n t h i s connection, we may consider t h e conditions as applied i n Sect.11. (l2.3b) - ( 1 2 . 3 ~ ) as n a t u r a l conditions f o r t h e extremum of t h e f u n c t i o n a l I1 i n a n extended sense, even i f they do not coincide with t h e n a t u r a l boundary conditions of t h e v a r i a t i o n a l problem. I n conclusion, l e t us d i s c u s s p o s s i b l e versions of t h e s o l u t i o n of t h e 1 1 . 1 6 ) , (11.18), (11.23), ( 1 1 . 2 % ) w i t h t h e boundary consystem of equations ( d i t i o n s (12.3b) (12.3~).

F i r s t l e t us concentrate on t h e system of f u n c t i o n s ? l q . Although t h e system of equations constructed by us i s complete, i . e . , t e number of equat i o n s i s equal t o t h e number of f u n c t i o n s sought, i t i s not advisable t o a t tempt a n Itexacttt determination of t h e q u a n t i t i e s k ( i I q from these equations. The i t e r a t i o n method should be used. For t h e beginning, l e t us put t h e funct i o n s k ( f J p equal t o zero o r u n i t y , zero corresponding t o t h e exact s a t i s f a c t i o n of t h e boundary conditions on t h e boundary s u r f a c e s of t h e s h e l l , and u n i t y t o t h e a p p l i c a t i o n of t h e Kirchhoff-Love hypotheses. Then, adopting one or t h e other method, l e t u s proceed t o t h e s o l u t i o n of t n e boundary problem. I n Section 1 1 we mentioned one of t h e procedures f o r t h i s solution, based on e q s . ( l l . l J ) . A d i f f e r e n t version of t h e s o l u t i o n i s possible, permitting us t o exclude from consideration t h e a p p r o x h a t e expressions f o r t h e componThis v e r s i o n i s based on an i t e r a e n t s u$) and, consequently, e q s . ( l l . l 8 ) . t i o n process. Assigning t o t h e c o e f f i c i e n t s k ( k l q d e f i n i t e values and r e j e c t i n g i n t h e quasi-body f o r c e s PR, a l l terms depending on t h e q u a n t i t y ud, l e t us solve t h e bomdary problem of t h e s h e l l theory. Then, from eqs.(ll.l6),we f i n d t h e corrected values of k ( & l q and introduce i n t o P% t h e terms depending on t h e

Cf., f o r instance, V.I.Smirnov, Gostekhizdat, 1951

Course i n iIigher Mathematics, Vol.IV,

272

p.295,

q u a n t i t i e s $I found as a r e s u l t of t h e o r i g i n a l approximation. The process i s then repeated. Evidently, we can expect p o s i t i v e results when this method i s applied t o t h e c a l c u l a t i o n of plane s h e l l s , s i n c e most of t h e terms r e j e c t e d i n obtaining t h e o r i g i n a l approximation depend on t h e curvature of t h e middle surface of t h e s h e l l .

A s i n t h e preceding Chapter, we note t h e c o r r e l a t i o n of t h e s o l u t i o n of t h e e l a s t o d y n d c eq,uations of t h e theory of s h e l l s , found by t h e method under study, with t h e general methods of t h e mathematical theory of e l a s t i c i t y . &en determining t h e components of t h e displacement vector by t h e approximate f o r mulas ( l l . l ) , we a r e e v i d e n t l y a b l e t o f i n d t h e components of t h e s t r a i n tens o r , and, from Hookels l a w , t h e components of t h e s t r e s s t e n s o r .

The Saint-Venant compatibility conditions Will b e satisfied. The equat i o n s of motion on t h e boundary conditions W i l l be approximately s a t i s f i e d . Here tle s a t i s f a c t i o n of t h e boundary conditions on t h e boundary surfaces of p. The i n t r o d u c t i o n t h e s h e l l can be improved by s e l e c t i n g t h e functions k of t h e f u n c t i o n s k ( i t ) p i s one of t h e f e a t u r e s of t h e proposed method. 1271 W e s h a l l now take up a question not y e t discussed. The equations obtained i n Sect.11 d i f f e r from (111, 24.27 - 24.29) found from t h e general equation of dynamics, i n being of a higher order. T h i s r e s u l t s i n c e r t a i n complications i n s t a t i n g t h e i n i t i a l conditions. It i s n a t u r a l t o use here a method based on t h e approximate s a t i s f a c t i o n of t h e i n i t i a l . conditions, s t a r t i n g from t h e requirements of t h e least-square d e v i a t i o n of t h e approximation functions from t h e f u n c t i o n s describing t h e s e conditions. Let us assume i n accordance w i t h (11, 8.la 8.lb) t h a t , a t the i n i t i a l time to, t h e components of t h e displacement v e c t o r uio and of t h e v e l o c i t y vector & a r e assigned as functions of t h e coordinates of a p o i n t of t h e s h e l l . If t h e s e f u n c t i o n s a r e d i f f e r e n t i a b l e , then, by d i f f e r e n t i a t i n g them we s h a l l f i n d t h e f i e l d of t h e i n i t i a l r a t e of deformation.

I n c e r t a i n problems of dynamics, t h e functions ut0 and q.40a r e not d i f f e r entiable. A n example i s given by t h e i n i t i a l values of t h e l o n g i t u d i n a l vel o c i t i e s i n a rod under l o n g i t u d i n a l impact. If t h e f u n c t i o n s u10 and q o a r e not d i f f e r e n t i a b l e , t h e i n i t i a l f i e l d s of deformation and of t h e r a t e of deformation must b e independently prescribed.

To s e t up t h e i n i t i a l conditions of t h e problem, l e t u s consider t h e f o u r functionals :

(12.4a)

273

where E i s an a r b i t r a r y small t h e i n t e r v a l . The Lagrange f u n c t i o n L i s est a b l i s h e d from t h e components of t.he v e l o c i t y v e c t o r of a s h e l l element , j u s t as t h e f u n c t i o n L i s e s t a b l i s h e d from t h e components of t h e displacement vect o r . The o t h e r n o t a t i o n i s familiar from t h e preceding discussion.

W e w i l l d e f i n e t h e i n i t i a l conditions, s t a r t i n g from t h e conditions t h a t t h e f u n c t i o n a l s I,, 12, I3 and 4 s h a l l b e minimum. The p h y s i c a l meaning of this requirement i s obvious. A s f o r t h e f u n c t i o n a l 4 , i t contains t h e "energy of a c c e l e r a t i o n " which e n t e r s i n t o t h e Gauss p r i n c i p l e of l e a s t c o n s t r a i n t , and t h e r a t e - o f - s t r a i n energy. Thus, t h i s f u n c t i o n a l i s connected with t h e q u a n t i t i e s c h a r a c t e r i z i n g t h e minimum p r o p e r t i e s of t h e a c c e l e r a t i o n s of t h e a c t u a l motion of t h e system. It would a l s o be p o s s i b l e t o introduce d i r e c t - /272 l y t h e Gaussian c o n s t r a i n t Z of t h e system, b u t this would make i t impossible t o use t h e a n a l y t i c apparatus employed above.
The requirement f o r minimizing t h e f u n c t i o n a l s 11, 12, Is, and 4 i s obviously equivalent t o t h e optimum simulation of t h e i n i t i a l mechanical s t a t e of t h e s h e l l by t h e q u a n t i t i e s r e s u l t i n g from t h e approximate equations (11.1). If we introduce t h e n o t a t i o n

+I

Sl

then t h e necessary conditions f o r t h e f u n c t i o n a l s I,, 12, I3 and imum will take t h e following form:

t o be min-

(12.6a)

( p = 1, 2, 3; m = O

1, 2.

.. .).

E uations (12.6a) permit a d i r e c t determination o f t h e i n i t i a l values ~ $ 5 ) S u b s t i t u t i n g t h e s e values i n t o eqs.(12.6b), we o b t a i n t h e initial and upo Of values $8) and then, passing t o eqs.(12.6c), t h e i n i t i a l values '$('I

( 8 ...

274

course, t h e r e may b e o t h e r approaches t o t h e determination of t h e extended system of i n i t i a l conditions. I n t h i s Chapter, we have considered t h e method of l i n e a r approximation of t h e components of t h e finite-deformation t e n s o r . Such a n approximation permits e l i m i n a t i n g only a p a r t of t h e nonlinear terms from t h e Lam6 equations, b u t does not exhaust t h e p,roblem of l i n e a r i z a t i o n , s i n c e o t h e r sources of nonl i n e a r i t y s t i l l remain, i n c l u d i n g t h e C h r i s t o f f e l symbols {J: 3 of t h e Lagrangian coordinates of a medium w i t h extensive deformations and n o n l i n e a r i t y i n t h e boundary conditions. For this reason, we must apply t h e method of l i n e a r /273 approximation t o t h e quasi-body f o r c e s containing nonlinear terms e n t e r i n g i n t o t h e Lam6 equations. The complexity of t h e a n a l y t i c expressions f o r t h e s e quasi-body f o r c e s , however, prevents us f o r t h e t i m e being from developing t h e method of l i n e a r approximation i n a g e n e r a l form. S e c t i o n 13. Approximate Methods of I n .v _ estigating th -.e - Equilibrium ..... and O s c i l l a t i o n s of S h . e ~l l s asDiscrete-Continuum . ~~- - Systems*
~ ~

I n concluding our d i s c u s s i o n of t h e problem complex connected with t h e g e n e r a l problem of c o n s t r u c t i n g a mechanical system c l o s e l y r e l a t e d , according 50 c e r t a i n c r i t e r i a , t o some prescribed system, l e t us b r i e f l y d i s c u s s t h e method of reducing t h e problems of t h e t h e o r y of s h e l l s t o problems of t h e study of motion of systems w i t h a f i n i t e number of degrees of freedom. W e s h a l l c a l l such a system a discrete-continuous system. The concept of "discrete-continuous system" w a s introduced i n t o t h e t h e o r y I n h i s terminology such a system i s a t h i n of s h e l l s by V.Z.Vlasov (Bibl.3b). w a l l e d e l a s t i c two-dimensional system possessing a f i n i t e number of degrees of freedom along one of t h e coordinates and a n i n f i n i t e l y g r e a t number along t h e other.
W e will a l s o use t h i s t e r m i n what follows, b u t s h a l l g i v e i t a d i f f e r e n t meaning. By t h e term "discrete-continuous system!' we s h a l l understand a continuous medium whose dynamic s t a t e i s approximately determined by a system of f u n c t i o n s of t h e r e l a t e d t o a d i s c r e t e s e t of p o i n t s on t h e b a s i c s u r f a c e of t h e s h e l l . Whatever o t h e r a s p e c t s of t h e concept of discrete-continuous syet e m s a r e possible, will n o t b e considered here.

The i n v e s t i g a t i o n i n t h e f i n a l p a r t of t h i s Chapter proposes t o g i v e a method of reducing t h e problem of d e r i v i n g q u a n t i t i e s t h a t determine t h e s t r e s s - s t r a i n s t a t e of a s h e l l t o t h e s o l u t i o n of f i n i t e systems of a l g e b r a i c equations i f t h e s h e l l i s i n equilibrium, and t o t h e s o l u t i o n of a system of o r d i n a r y d i f f e r e n t i a l equations of second o r higher o r d e r i f v i b r a t i o n s of t h e s h e l l are t o be s t u d i e d .
The p r i n c i p l e a n a l y t i c approach t o t h e c o n s t r u c t i o n of equations describi n g t h e s t a t e of a s h e l l as a discrete-continuous system i s t h e use of i n t e r -:$

a s p r e s e n t e d t o t h e AllThe substance of t h i s and t h e f o l l o w h g S e c t i o n s w Union Conference on t h e Theory of Plates and S h e l l s held a t Kazan i n 1960. C f . , t h e a u t h o r ' s paper i n t h e Transactions of t h e Conference.

275

1 II

Ill

II

11111111111.1111111 11111111

p o l a t i o n formulas t h a t express t h e values of t h e functions sought i n terms of a d i s c r e t e s e t of t h e i r v a l u e s a t t h e nodes of a c e r t a i n net*. Consequently, here, too, as i n t h e e a r l i e r p a r t of t h i s Chapter, we i n t e n d t o use one of /274 t h e methods of approximation functions.

The choice of such a method i s a r b i t r a r y . W e have n o t i n v e s t i g a t e d t h e comparative e f f e c t i v e n e s s of t h e various methods of approximation functions i n a p p l i c a t i o n t o t h e b a s i c problem, t h e replacement of t h e s h e l l by a system with a f i n i t e number of degrees of freedom.
Section 1 4 . The Fundamental Discrete System of Unknowns

To construct t h e discrete-continuous system replacing t h e s h e l l , one of t h e above-considered reduction methods must be used.
W e s h a l l apply t h e methods studied a t t h e beginning of Chapter 1 1 1 . This method i s c l o s e s t t o t h e method based on t h e Kirchhoff-Love hypotheses. O u r results discussed below can, therefore, b e extended without fundamental comp l i c a t i o n s t o t h e c l a s s i c a l theory of s h e l l s . .

A s shown previously (111, Sect.5), t h e three-dimensional problem of t h e theory of e l a s t i c i t y can be reduced t o a determination of s i x functions of a p o i n t of t h e b a s i c s u r f a c e of t h e s h e l l . These functions a r e t h e displacement vector components u , of a p o i n t of t h e b a s i c s u r f a c e of t h e s h e l l , and t h e cov a r i a n t d e r i v a t i v e s V3u1

Assume t h a t t h e v a l u e s of t h e six functions, determining t h e s t a t e of t h e s h e l l , a r e known.at t h e nodes of some n e t on t h e b a s i c surface. Then t h e v a l u e of t h e s e functions a t t h e intermediate p o i n t s can be determined by one of t h e i n t e r p o l a t i o n formulas. This permits us t o express t h e s t r a i n s and, by means of Hookers law,the s t r e s s e s a t a n a r b i t r a r y p o i n t of t h e s h e l l i n terms of t h e values of unknowns a t t h e nodes of t h e n e t . The i n t e r p o l a t i o n method involves t h e r e l a t i v e order of t h e terms ret a i n e d on reduction of t h e three-dimensional problem of t h e theory of elast i c i t y t o a two-dimensional problem i n expansions of t h e form of (111, 4 . 2 ) , ( 1 1 1 , L , 5 a ) , (111, 4.5b) and subsequent r e l a t i o n s r e s u l t i n g from those enumerated. Making lise of eas.(II,2.11) and expansions of t h e form of (111, 4.2), /275 l e t us consider, f o r example, elementary and roughly approximate representa-s+

-::We r e c a l l t h a t t h e use of i n t e r p o l a t i o n formulas w a s given i n t i e theory of s h e l l s by 1.Ya.Shtayerman i n h i s work "On t h e Application of I n t e r p o l a t i o n Methods t o t h e Approximate I n t e g r a t i o n of t h e D i f f e r e n t i a l Equations of Equil i b r i u m of m a s t i c S h e l l s " , V i s t i KPI, Vo1.2, 1927 and i n t h e problems of s t r u c t u r a l mechanics, by N .V .Kornoukhov i n h i s paper "An Interpolation-Iterat i o n Method of S o l v i n g t h e D i f f e r e n t i a l Equations of Strength and S t a b i l i t y of Prismatic Rods", Sbornik trudov I n s t . s t r o i t . mekhan AN UkrSSR, Vol.11, 1949 3%- T h i s approximation corresponds t o t h e accuracy of determination of t h e s t r a i n tensor components adopted i n t h e c l a s s i c a l theory of s h e l l s .

276

t i o n s of t h e components of t h e s t r a i n tensor,

where
sj

=v3ujl z - 0 .

(14.2)

The meaning of t h e other symbols has been given i n Chapter 1 1 1 . I n s e t t i n g up eqs.(lf+.la) - (U+.lb), w e r e t a i n e d i n t h e nonlinear p a r t of t h e s t r a i n t e n s o r components only those terms w i t h t h e g r e a t e s t s i g n i f i c a n c e according t o t h e well-known p o s t u l a t e s t h a t can be t r a c e d back t o t h e i n v e s t i gations by T.Karman. The r i g h t -hand s i d e s of eqs ( 1 4 .l a ) (14. l b ) contain f i r s t - o r d e r derivatives with r e s p e c t t o t h e coordinates x i ( i = 1, 2). For this reason, remaini n g a t l e a s t w i t h i n t h e limits of accuracy adopted i n t h e net method, l e t us apply t h e following i n t e r p o l a t i o n methods: Let us cover t h e b a s i c surface with a t r i a n g u l a t i o n n e t , and within each t r i a n g l e l e t us i n t e r p o l a t e t h e unknown functions by l i n e a r functions of t h e coordinates of t h e b a s i c surface, taking values equal t o t h e values of t h e unknown functions a t t h e v e r t i c e s of t h e triangle. Consider f o r example t h e t r i a n g l e F l (p, q ) , && ( p + 1, q ) , b ! ' (p, q + 1 ) . tlere, p and q a r e t h e numbers of t h e nodes of t h e n e t on t h e coordinate l i n e s . The component of displacement uJ w i t h i n t h e t r i a n g l e MI, &, & w i l l be expressed by t h e equation

or

The c o e f f i c i e n t s of t h e l i n e a r t r i n o m i a l s which a r e t h e f a c t o r s of t h e components u J ( & ) i n eq.(l4.3b) can be found from a comparison of eqs.(l4.3a) W e w i l l not give t h e expressions f o r t h e s e components. The n i t h eq.(14.3b).

277

f u n c t i o n s 9~ w i t h i n t h e t r i a n g l e

MI&&

are similarly determined.

($) i n t h e problems of dynamics are f u n c t i o n s The q u a n t i t i e s uj (1% ) and of t h e t i m e t. I n problems of s t a t i c s , t h e y do n o t depend on t h e time. /;L76
Approximating t h e displacements uJ"1 by l i n e a r f u n c t i o n s of z, we complete t h e c o n s t r u c t i o n of t h e q u a n t i t i e s t h a t a p p r o d m a t e l y c h a r a c t e r i z e t h e s t a t e of a p r i s m a t i c element of t h e s h e l l r e s t i n g on t h e t r i a n g l e M1&&. Making use of e q s . ( l & . l a ) ( U . l b ) , we can approximately determine t h e s t r a i n - s t r e s s state of t h e s h e l l i n t h e p r i s m a t i c element r e s t i n g on t h e triI n r e l a t i v e accuracy, this determination corresponds ( f o r exangle MI&%. ample) t o t h e accuracy with which t h e s t r e s s - s t r a i n s t a t e i s determined f o r a one-dimensional rod i n l o n g i t u d i n a l v i b r a t i o n , i f t h e rod i s replaced b y a system of concentrated masses connected b y w e i g h t l e s s springs. The d i f f e r e n c e i s p r i m a r i l y t h a t , i n t h e rod, when this method of approximate s o l u t i o n of t h e problem of l o n g i t u d i n a l v i b r a t i o n s i s used, t h e s i n g l e component of t h e s t r a i n t e n s o r has d i s c o n t i n u i t i e s a t t h e p o i n t s a t which t h e mass i s concentrated, w h i l e i n t h e case under c o n s i d e r a t i o n t h e faces of t h e p r i s m a t i c elements will b e s u r f a c e s of s e p a r a t i o n of t h e s t r a i n t e n s o r components. The s i x p r i n c i p a l q u a n t i t i e s , however, w i l l r e t a i n t h e i r c o n t i n u i t y on t h e s e s u r f a c e s . Further refinements will l e a d t o a n i n c r e a s e i n t h e number of terms ret a i n e d i n expansions of t h e form of e q s . ( l 4 . l a ) - (14.lb) and t h e consequences, as we know from Chapter 1 1 1 , w i l l introduce i n t o t h e expansion derivatives of second and higher order with r e s p e c t t o t h e coordinates of t h e b a s i c surface, so t h a t t h e l i n e a r approximations of t h e form of eqs.(lL+.3a) ( l 4 . j b ) will become i n s u f f i c i e n t . V 'J e must t a k e recourse t o i n t e r p o l a t i o n formulas i n t h e form of polynomials of t h e coordinates d ( j = 1, 2 ) of t h e second, t-hird, and higher orders. This W i l l complicate t h e base r e g i o n of approximation.

I n a n approximation by l i n e a r trinomials, suzh a region, as a l r e a d y mentioned, i s a t r i a n g l e . For an approximation by polynomials of t h e second degree we may, f o r i n s t a n c e , use a tldoublettt r i a n g l e fill(p, q ) , &(p
+

l 2,

q),

mate t h e polynomials of t h e third-degree t l t r i p l e l tt r i a n g l e with an a d d i t i o n a l i n t e r n a l point, e t c . . R e flfractionalfl numbering of t h e i n t e r p o l a t i o n nodes i n d i c a t e s t3e p o s i t i o n of a n auxiliary node between t h e nodes of t h e main re<$on, which remains a t r i a n g l e w i t h i n t e g e r s used i n numbering i t s nodes. For
example, t h e node & ( p
+

1
2 '

q ) l i e s on t h e s t r a i g h t l i n e j o i n i n g t h e

nodes Ml(p, q ) and r/h ( p + 1, q ) . A l l these cases l e a d t o approximation formul a s f o r t h e six p r i n c i p a l f u n c t i o n s t h a t a r e l i n e a r with r e s p e c t t o t h e values of t h e s e f u n c t i o n s a t t h e nodes of t h e n e t .
The c o e f f i c i e n t s of t h e v a l u e s of t h e p r i n c i p a l unknowns a t t h e nodes of a n e t are polynomials of t h e coordinates x i ( i = 1, 2 ) , equal t o unity a t t h e

278

r e s p e c t i v e node and equal t o zero at a l 1 , o t h e r nodes.

We have

The sums a r e extended t o t h e base r e g i o n s of approximation i n d i c a t e d above. men L e t u s denote t h e coordinates of t h e nodes by and

&

4.

where, as mentioned above, cppp and t h e b a s i c s u r f a c e of t h e s h e l l .

q P q a r e polynomials of t h e coordinates x i of

The expressions (ll+.'.La) (l/+.Lb), as w e l l as t h e reduction formulas considered i n Chapter 1 1 1 , y i e l d approximate expressions f o r t h e p o t e n t i a l and k i n e t i c energy of t h e s h e l l and make i t p o s s i b l e t o c o n s t r u c t t h e Lagrange f u n c t i o n L-3. The generalized coordinates here W i l l b e t h e q u a n t i t i e s UJ (p, q) a n d a 3 ( p , q).. Among ttlese g e n e r a l i z e d coordinates, however, t h e r e may a l s o b e redundant coordinates, s i n c e t h e boundary conditions of t h e problem impose, on t h e q u a n t i t i e s uJ (p, q ) and a j(p, q ) , r e s t r i c t i o n s which a r e a n a l y t i c a l l y e-xpressed by t h e equations of SezJmetric connectivity, and i n t h e g e n e r a l case, of kinematic connectivity. Let us consider t h i s q u e s t i o n i n g r e a t e r d e t a i l . Section 15. Boundary Conditions ~- . and t h e Equations of Connectivity. I n i t i a l Conditions

I n considering t h e boundary conditions w e s h a l l s t a r t from t h e concepts of the three-dimensional s t r e s s - s t r a i n s t a t e of t h e s h e l l , as auopted i n Chapter 1 1 1 . The v a r i o u s boundary conditions i n t r o d u c e no a d d i t i o n a l complications i n t o t h e s o l u t i o n of t h e problem b y this method. K i n m a t i c , k i n e t i c and mixed boundary conditions may b e p r e s c r i b e d on t h e contour surface. These conditions l e a d t o equations of l i n e a r and nonlinear geometric and kinematic connectiv- /278 ity. Without going i n t o d e t a i l , l e t u s consider t h e cases of t h e p r i n c i p a l boundary conditions p r e s c r i b e d on t h e contour s u r f a c e of t h e s h e l l .

% r e L i s t h e Lagrange f u n c t i o n for t h e shell as a whole, r a t h e r t h a n t h e d e n s i t y of t h e Lagrange f u n c t i o n considered above, and i s termed f o r b r e v i t y t h e "Lagrange function".

279

1 . F i r s t Boundary Condition -On t h e contour s u r f a c e C, l e t t h e displacements * b e given:

(q')c = 'pi(d, 2, t )

(i = 1, 2, 3; j

= 1, 2).

Making use of t h e n o t a t i n n adopted i n Chapter 1 1 1 , and s e t t i n g

we f i n d , from (111, 6.ha) - (111, 6,!+b), surface, including i t s contour C.

equations v a l i d m e r tne e n t i r e b a s i c

( i , s = 1, 2).

Again making use of t h e expansions (111, 13.3) and confining ourselves t o t h r e e terms of the expansions on t h e right-hand s i d e s of eqs.(15.3a) - (15.3b) we obtain
(UJC

= (pi ( x i , 0,

t ) ; p i ) C = cpl') ( X j , 0, t ) ( i = 1, 2, 3;j = 1, 2).

Tne condition (15.L.a), a f t e r a p p l i c a t i o n of t h e i n t e r p o l a t i o n formulas,

Will lead t o equations of ,geometric c o n n e c t i v i t y r e l a t i v e t o t h e generalized

, ) and e i ( p , q). I n f a c t , making use of eqs.(U.L+a) coordinates ~ ( p q (U.L!.b), we obtain from eq.(l5./+a):

/279

*W e

r e c a l l t h a t on t h e contour s u r f a c e t h e coordinates xJ are connected by t h e equation of t h e contour of t h e b a s i c surface.

280

IFi

(i= 1, 2, 3; j = 1 , 2).

Equations (15.5) must be s e t up for segments of arc of t h e contour C belonging t o t h e b a s e r e g i o n s of approximation t o which t h e sums
CP 9)

are ex-

tended. It can b e s i m i l a r l y shown t h a t t h e c o n n e c t i v i t y determined b y eqs.(l5.!hb) - ( 1 5 . L ~ ) i n t h e g e n e r a l case i s n o t geometrical. W e s h a l l term i t kinematic, although, i n c o n t r a s t t o t h e kinematic c o n n e c t i v i t y of c l a s s i c a l d y n d c s , i t s equations c o n t a i n d e r i v a t i v e s of t h e second order with r e s p e c t t o t h e time t of t h e generalized coordinates, i f we confine ourselves t o t h e t h r e e f i r s t terms of t h e expansions on t h e right-hand s i d e s of eqs.(l5.3a) t o eqs (15.3b )

2. Second Boundary Problem

This problem has been considered previously (111, Sect.13) i n t n e formulat i o n t h a t b e s t corresponds t o t h e method under study. Making use of t h e boumiary conditions i n t h e form of t h e equations of Cnapter I11 (111, 13.9a), (111, 13.9b) we a g a i n f i n d equations of c o n n e c t i v i t y analogous t o those considered above. W e note here t h a t , r e t a i n i n g only t h e two f i r s t terms on t h e right-hand s i d e s of t h e expansions ( 1 1 1 , 13.6) we obt a i n , i n t h e g e n e r a l case, one equation of geometric connectivity, and one equation of kinematic connectivity, which r e s u l t s d i r e c t l y from a considerat i o n of t h e left-hand s i d e s of eqs.(III, 13.9a), (111, 13.9b). W e will not w r i t e out t h e equations of t h e s e c o n n e c t i v i t i e s . L e t us d i s cuss o n l y t h e c a s e s i n which t h e kinematic c o n n e c t i v i t y degenerates, as already discussed i n our c o n s i d e r a t i o n on t h e b a s i c boundary problems.
I f t h e c o n d i t i o n s of attachment of t h e contour of t h e b a s i c s u r f a c e i n clude i n themselves t h e conditions t h a t some displacement component on t h e contoilr C s h a l l vanish, then t h e wave operator M f o r t h i s component will be transformed, as will b e seen from ( 1 1 1 , 6 . 3 a ) , i n t o a Laplace operator on t h e contour C. I n t h i s case t h e equations of kinematic c o n n e c t i v i t y r e s u l t i n g from t h e conditions (15.4b) ( 1 5 . 4 ~ )are transformed i n t o equations of geometrical connectivity. Obviously, this does not a p l y t o t h e derivatives of t h e operator M with r e s p e c t t o t h e coordinates x' ( = 1 , 2). Tnese derivatives appear when introducing, i n t o t h e expansions of I $ ) i n powers of z, t e r m s contain- /280 i n g z i n degrees h i g h e r than t h e second.

281

I n conclusion, we will make a b r i e f statement on t h e correspondence between t h e boundary conditions and t h e equations of motion. L e t us t u r n a g a i n There, we showed t h a t t h e boundary conditions obtained as t o (111, Sect.13). a r e s u l t of t h e method of successive approximation cannot be satisfied w i t h t h e same r e l a t i v e accuracy as t h e system of e q u a t i o n s determining t h e displacements of t h e p o i n t s of t h e b a s i c surface.
These conclusions do n o t extend t o t h e method s t u d i e d here, s i n c e we do by means of successive approximations. The V a l not exclude t h e components f t ~ i d i t y of t h e above becomes obvious from a comparison of t h e number of equations if motion and t h e number of equations of connectivity.
j . I n i t i a l Conditions ..

The i n i t i a l c o n d i t i o n s were considered p r e v i o u s l y (111, S e c t . U ) , where we used t h e method of successive appro-ximation, which, as a l r e a d y mentioned, will not b e applied here. &king use of t h e conditions (111, 14.1) and t h e i r expansions i n s e r i e s i n ascending powers of z of t h e form (111, Ilc.2a), and a l s o making use of eqs.(15.3a) - (15.3b), we f i n d a system of i n i t i a l conditions similar t o t h e system considered above (111, Sect.14). dithout w r i t i n g out t h e s e conditions again, l e t u s f i r s t pose t h e question whether t h e i n i t i a l conditions correspond t o t h e o r d e r of t h e system of d i f f e r e n t i a l equations of motion mentioned b e f o r e A d e f i n i t e answer can b e g i v e n here only f o r t h e case of a (111, Sect.14). l i n e a r dependence between t h e displacements uJz)and t h e coordinates z. T h i s case corresponds t o t h e r e p r e s e n t a t i o n of t h e s t r a i n t e n s o r components by Then t h e numb$r of i n i t i a l conditions containing t h e eqs.(lb.la) (14.1b). i n i t i a l values uiO(p, q), e i o ( p , q), ulO(p, q), h i o ( p , q ) ( i = 1, 2, 3) will b e twelve for each p a i r of n m b e r s (p, q). The system of equations of motion cons i s t i n g of six d i f f e r e n t i a l equations of t h e second o r d e r f o r each p a i r of numb e r s (p, q) w i l l b e of a n order equal t o t h e number of i n i t i a l conditions.

If, i n t o t h e expansions of t h e displacement-vector and s t r a i n - t e n s o r components, we i n t r o d u c e terms with t h e f a c t o r z, then, considering t h e cases of degeneration of t h e kinematic boundary conditions, we f i n d t h a t t h e number of i n i t i a l conditions Will i n c r e a s e t o eighteen, w h i l e t h e system of d i f f e r e n t i a l equations of motion, as shown i n t h e following Section, w i l l c o n s i s t of t h r e e equations of t h e s i x t h order and t h r e e of t h e second order, f o r each p a i r of numbers (p, q).
here w i l l c o n t a i n q o ( p , - q), G 0 ( p , q), G 0 ( p , q), /281 (p, q). There a r e n o t enough of t h e s e condiLions, and we must consider t h e next terms of t h e expansions of t h e displaceIf, i n t h i s case, t h e r e l a t i v e acment v e c t o r cgmponents w i t h t h e f a c t o r z. curacy of t h e equations of motion and t h e equations of c o n n e c t i v i t y remains unchanged, t h e n t h e number of i n i t i a l conditions i n c r e a s e s t o twenty-four f o r each p a i r of numbers (p, q ) ; t h e s e w i l l n3w i n c l u d e t h e i n i t i a l v a l u e s of t n e d e r i v a t i v e s of t h e f o u r t h and f i f t h o r d e r w i t h r e s p e c t t o time, $6) (p, q ) and u{z) (p, q). Thus, t o s o l v e t h e problem, t h e i n i t i a l conditions must b e s a t i s f i e d with a higher degree of accuracy t h a n t h e r e l a t i v e accuracy of t h e

. .. The i n i t i a l conditions uio (p, q), e i 0(p, q ) and i i

equations of motion and t h e equations of c o n n e c t i v i t y . S e c t i o n 16. Equations of Motion of t h e S h e l l

I f t h e bourldary c o n d i t i o n s l e a d t o equations of geometric o r degenerate kinematic connectivity, we can make use of t h e Ostrogradskiy-Hamilton p r i n c i p l e t o s e t up t h e equations of motion of t h e nodes of t h e i n t e r p o l a t i o n n e t on t h e b a s i c s u r f a c e of t h e s h e l l .

The Ostrogradskiy-Hamilton following v a r i a t i o n a l equation:

p r i n c i p l e i s expressed, as w e know, by t h e

t.

s^

(6A

+ 6 L ) d t =0,

(16.1)

where bjA i s t h e elementary work of t h e nonconservative f o r c e s performed on passage of t h e p o i n t s of t h e system from t h e t r a j e c t o r i e s of a c t u a l motion onto t h e t r a j e c t o r i e s of comparison, and L i s t h e Lagrange f u n c t i o n of t h e s h e l l Itas a wholet1. I f we confine o u r s e l v e s t o t h e f i r s t two terms on t h e right-hand s i d e s of eqs. (15.3a) - (15.3b), and r e p r e s e n t t h e s t r a i n t e n s o r components by e q s . ( U . l a ) ( l / + . l b ) , t h e n t h e zquations of motion of t h e discrete-conti'nuous system rep l a c i n g t h e s h e l l Will b e of t h e form of Lagrange equations of t h e second kind:

(16.2a) (16.%b)
( i = 1, 2, 3 ) .

where Ql and cfl are g e n e r a l i z e d nonconservative f o r c e s . !-Iowever, as we know, we have t h e r i g h t t o i n c l u d e a l s o conservative f o r c e s i n t h e s e generalized f o r c e s , if t h i s can h e l p t o s i m p l i f y s o l u t i o n of t h e problem.
I f t h e equations of kinematic c o n n e c t i v i t y degenerate on consideration /282 of t h r e e terms i n t h e expansions of t h e displacement v e c t o r i n powers of t h e coordinate z, t h e n we can a g a i n make use of t h e Ostrogradskiy-Hamilton p r i n c i p l e (16.1) i n s e t t i n g up t h e equations of motion, b u t i n t h i s case t h e La.mange f u n c t i o n L Will c o n t a i n t h e second and t h i r d t i m e d e r i v a t i v e s of t h e i e n e r a l i z e d coordinates ui(p, q) agd, as before, t h e f i r s t t h e d e r i v a t i v e s of e can convince ourselves of t h i s by t h e generalized c o o r d i n a t e s a i (p, q). W Thus, under t h e considering t h e right-hand sides of eqs.(l5.3a) - (15.3b). adopted assumptions on t h e equations of c o n n e c t i v i t y , t h e equations of motion of .the discrete-continuous system r e p l a c i n g t h e s h e l l Will now b e of t h e fol-

lowing form:

(i'l,

2, 3).

N e c a l l t h e r e a d e r ' s a t t e n t i o n t o t h e d i f f e r e n c e i n t h e o r d e r s of t h e This d i f f e r e n c e of equations enterini: i n t o t h e subsystems (16.3a) - (16.3b). x d e r s , under t h e methad of r e d u c t i o n here adopted, w i l l occur when a n odd numb e r of terms i s r e t a i n e d on t h e right-hand s i d e s of t h e expansions i n t e n s o r s e r i e s of t h e displacement v e c t o r components q ( ' ) i n ascending powers of z . Nith a n even number of t e r m s r e t a i n e d i n t h e expansions, t h e o r d e r of t h e equat i o n s e n t e r i n 3 i n t o t h e subsystems (16.3a) - (16.3b) will b e t h e same*. This a s s e r t i o n i s i n p a r t i c u l a r i l l u s t r a t e d by eqs.(16.2a) (16.2b).

I f a s h e l l has a p a r t of t h e contour s u r f a c e f r e e of connectivity, then t h e equations of kinematic connectivity, r e s u l t i n g from t h e conditions (15.4.b) 50 ( 1 5 . 4 ~ ) w i l l not degenerate i n t o equations of geometric connectivity, and t h e Ostrogradskiy-Hamilton p r i n c i p l e W i l l not b e applicable, a t l e a s t not without a d d i t i o n a l i n v e s t i g a t i o n .

I n t h e s e cases, we may g i v e up t?ie method of reduction based on c m s i d e r a t i o n of six f u n c t i o n s of ut a n d a 1 of a p o i n t of t h e b a s i c s u r f a c e of t h e s h e l l , permikting u s t o c a r r y t h e reduction problem t o completion; i n s t e a d , we may use a p p r o e m a t e r e p r e s e n t a t i o w of t h e displacement v e c t o r by polynomials 1 1 i n c)ur study of aryanged i n powers of z, which were considered i n Chapter 1 reduction met5ods r e l y i n g on t h e general equation of dynamics. I n t h i s case, such d i f f i c u l t i e s w i l l not a r i s e , s i n c e t h e wave o p e r a t w M appears i n t h e /283 -) (l5.3b) as a r e s u l t of a p p l i c a t i o n of t h e Lame' equations i n relations ( 1 5 . 3 ~ ~ order t o eliminate, from t h e expansiorls i n Taylor t e n s o r s e r i e s , t h e covariant deri,vatives of t h e component ul of t h e second and higher o r d e r s with respect to x " = z. A t t h e same t i m e , i t can be s t a t e d t h a t a p p l i c a t i o n of t h e Lam6 equations improves t h e accuracy of t h e approximations. W e will not develop a v e r s i o n of t h e discrete-continuous method t h a t i s not connected with t h e u s e of Lam6 equations. Section 17. Concluding Remarks

Chapter I V covered a group of questions connected with t h e fundamental

$5 This statement supplements our paper read a t t h e Conference on S h e l l Theory a t Kazan i n 1960

problem of t h e s h e l l theory, which reduces t o t h e c o n s t r u c t i o n of a mechanical system approximately equivalent, according t o some c r i t e r i o n , t o t h e s h e l l as a three-dimensional e l a s t i c body.

As t h e a n a l y t i c ' c r i t e r i o n of approximate equivalence, we s e l e c t e d t h e magnitude of t h e q u a d r a t i c d e v i a t i o n of some f u n c t i o n c h a r a c t e r i z i n g t h e s t a t e of t h e mechanical system t o b e constructed from t h e corresponding f u n c t i o n c h a r a c t e r i z i n g t h e state of t h e three-dimensional body, namely, t h e s h e l l . As t h e f u n c t i o n we chose t h e d e n s i t y of t h e Lagrange f u n c t i o n s , t h e body f o r c e s , o r t h e s u r f a c e f o r c e s , depending on t h e s p e c i f i c problem involved. The solut i o n of v a r i o u s p h y s i c a l problems w a s u n i f i e d by t h e general requirement t h a t t h e s e q u a n t i t i e s f o r t h e approximately equivalent system show minimum d e v i a t i o n from t h e same q u a n t i t i e s f o r t h e three-dimensional body, t h e s h e l l .
Thus, Chapter IV contains t h e s o l u t i o n of a s e r i e s of imation f u n c t i o n s , which are r e f l e c t e d i n t h e mechanics of t h e r e f o r e a l s o included i n Chapter I V t h e f i r s t p r i n c i p l e s c o n s t r u c t i o n of a discrete-continuous system r e p l a c i n g t h e t e r p o l a t i o n formulas t o f i n d t h e required approximation. problems of approxshells. W e have of t h e theory of t h e s h e l l , and used i n -

I n meaning, t h e c o n s t r u c t i o n of a discrete-continuous system r e p l a c i n g a s h e l l i s c l o s e t o t h e f i n i t e - d i f f e r e n c e method.

The proposed method d i f f e r s from t h e f i n i t e - d i f f e r e n c e method, however, more exact, s i n c e c o n s t r u c t i o n of t h e equations of motion i s based on t h e operations of i n t e g r a t i o n r e q u i r e d f o r t h e c a l c u l a t i o n of both k i n e t i c and p o t e n t i a l energy. The o p e r a t i o n of i n t e g r a t i o n somewhat smoothes t h e e r r o r s introduced by t h e i n t e r p o l a t i o n formulas. One of t h e major e r r o r sources s t i l l p e r s i s t s , namely, t h e approximation formulas of c o n n e c t i v i t y t h a t r e s u l t from t h e bouvldary conditions.
i . 1being

iz shortcoming Qf t h e method i s t h e complexity of t h e equations of motion and the boundary conditions. The f i e l d of a p p l i c a b i l i t y of t h e method theref o r e encompasses a l l problems where t h e use of methods of t h e Bubnov-Galerkin type involves fundamental d i f f i c u l t i e s i n c o n s t r u c t i n 3 t h e systems of appro,dmaiion f u n c t i o n s .
A n elementary example of such problems i s t h e problem of t h e v i b r a t i o n s of a r e c t a n q u l a r p l a t e with mixed conditions on each s i d e of t h e r e c t a n g u l a r contour of i t s middle surface.

/284.

A n advantase of t h e method i s t h e s i m p l i c i t y of programming i n c a l c u l a t i o n on high-speed e l e c t r o n i c computers.

285

CHAPTER V
INTZRAL AND INTEGRO-DIFFERENTIAL FQUATIONS O F THE THEDRY OF SHELLS
Section 1. General C h a r a c t e r i s t i c s of t h e Contents of t h e Concluding Chapter The l a s t Chapter c o n t a i n s a discussion of p a r t of our results i n t h e methods of solving t h e boundary conditions of s h e l l theory, r e l y i n g on t h e integrod i f f e r e n t i a l and i n t e g r a l equations of t h e s t a t i c s and dynamics of s h e l l s , res u l t i n g from t h e theorem of work and reciprocity. These s t u d i e s were begun by us i n 1939-1940 and a r e s t i l l going on a t p r e s e n t (Bibl.23b-j). During t h e p a s t f i v e y e a r s t h e p o s s i b i l i t y o f applying t h e apparatus of i n t e g r o - d i f f e r e n t i a l and i n t e g r a l equations t o t h e s o l u t i o n of boundary probl e m s of t h e s h e l l theory has a t t r a c t e d t h e a t t e n t i o n of many workers. Besides t h e methods i n d i c a t e d above, t h e y have used o t h e r methods, based p a r t i c u l a r l y on t h e i n t e g r a l r e l a t i o n s generalized i n t h e Green formulas of t h e theory of t h e Newtonian p o t e n t i a l function. Limited space prevent us from giving a det a i l e d a n a l y s i s of t h e v a r i o u s methods of reducing t h e boundary problems of t h e s h e l l theory t o equivalent systems o f i n t e g r o - d i f f e r e n t i a l and i n t e g r a l equaMany q u e s t i o n s of t h e theory of this reduction, including t h e problem tion+. of equivalence, existence, and uniqueness of t h e s o l u t i o n s of t h e equations s e t up by u s w i l l n o t be exhaustively answered here. W e intend t o r e t u r n to them i n t h e second p a r t o f t h i s book.

The last S e c t i o n s of t h i s Chapter w i l l contain t h e i n t e g r o - d i f f e r e n t i a l /286 and i n t e g r a l equations of t h e dynamics of s h e l l s , t o g e t h e r w i t h s p e c i a l applica1 . t i o n s of t h e generalized r e c i p r o c a l theorem proved i n Chapter 1
Section 2 . Elementam S o l u t i o n s of Three-Dimensional Problems of E l a s t i c i t y Theory Containing S i n g u l a r P o i n t s and Lines I n t h e first. Section o f t h i s Chapter we applied t h e method of c o n s t r u c t i n g i n t e g r o - d i f f e r e n t i a l equations of t h e s h e l l theory based on t h e i n t r o d u c t i o n of s o l u t i o n s of t h e three-dimensional problem of e l a s t i c i t y theory containing sing u l a r i t i e s arranged along a c e r t a i n segment of a s t r a i g h t l i n e . W e used t h i s method e a r l i e r (Bib1.23b) i n 1939 1940, and w i l l t a k e t h e r e s u l t s , given below, from t h a t work. W e s h a l l consider s o l u t i o n s with s i n g u l a r i t i e s of t h e three-dimensional s t a t i c problem of t h e theory of e l a s t i c i t y found f o r a l i n e a r l y deformed medium.

During p r e p a r a t i o n of this work f o r t h e press, t h e book by D.V.Vaynberg and A .L.Sinyavskiy (Bibl.17) appeared which contains a b r i e f discussion of t h e method of c o n s t r u c t i o n of i n t e g r o - d i f f e r e n t i a l and i n t e g r a l equations of t h e s h e l l theory given by u s and o t h e r authors, t o g e t h e r w i t h s e v e r a l a p p l i c a t i o n s t o t h e theory of s p e c i f i c boundary problems, i n c l u d i n g t h e problems of t h e equil i b r i u m of c y l i n d r i c a l notched s h e l l s . It a l s o g i v e s (Bib1.17) a bibliography which i s incomplete b u t s t i l l deserves a t t e n t i o n .

286

It i s well known t h a t displacements i n an unbounded e l a s t i c medium, due t o t h e a c t i o n of a s i n g l e concentrated f o r c e d i r e c t e d along t h e axis OY, of a rect a n g u l a r Cartesian coordinate system y , and applied t o t h e p o i n t with coordin a t e s TIli are o f t h e following f o M :

q q i( Y j ,

qj)

2 m { 5 1--v -6v

-+

--

r3
~

a 2

2(1

ay;

(t)};

(i, j, k = 1 , 2, 3), ( i f 4).

where

I f , i n s t e a d of a single force, we apply a t t h e p o i n t M(T, ) t h e a r b i t r a r y f o r c e s Yk d i r e c t e d along t h e coordinate axes, then t h e displacements correspondi n g t o t h e s e f o r c e s w i l l be expressed by t h e e q u a l i t i e s :

The s t r e s s e s corresponding t o t h e displacements (2.la)

(2.lb) have the/287

form:

Cf,, f o r i n s t a n c e (Bibl.9b) o r E.Trefftz, ONTI, 1934, pp.39-40, 287

Mathematical Theory of m a s t i c i t y ,

The a r b i t r a r y s y s t e m of f o r c e s Yk applied a t t h e point M(Vj ) c r e a t e s a f i e l d of s t r e s s e s defined by t h e e q u a l i t i e s

L e t u s now consider t h e expressions f o r t h e displacement v e c t o r components 6, and t h e corresponding s t r e s s t e n s o r components i n a c u r v i l i n e a r system of coordinates by means of which t h e space i n s i d e t h e s h e l l i s arithmetized, under s p e c i f i c a t i o n , a s i n d i c a t e d below, of t h e d i r e c t i o n of t h e s i n g l e concentrated force.
Between t h e Cartesian coordinates y , and t h e i n t e r n a l c u r v i l i n e a r coordi- i of t h e p o i n t s of t h e s h e l l t h e r e exist r e l a t i o n s h i p s expressed by t h e nates x formulas of d i r e c t and i n v e r s e transformation:

x i =xi(yjj; y j = ~5 (xi)
If we put x ?
= 0, then t h e i r r e l a t i o n s

(i, j = 1, 2, 3 j .

(2.6

y i = y i ( X I , 9, 0)

li = 1, 2, 3)

(2.7)

w i l l be t h e equations of t h e b a s i c surface of t h e s h e l l , Below, we will primarily consider s h e l l s of constant thickness. I n t h i s case, t h e b a s i c surface w i l l coincide with t h e middle surface, and t h e boundary s u r f a c e s of t h e s h e l l w i l l be included i n t h e system of coordinate surfaces.
Let u s now assume t h a t a t some point M of t h e s h e l l a f o r c e i s applied having t h e components
Fk =6k
(i)

(2.8)

d i r e c t e d along t h e tangent t o t h e coordinate l i n e x ~ . L e t u s f i n d i t s com- /288 ponents i n t h e rectangular Cartesian coordinate system yj. Making use of t h e formulas f o r t h e transformation of t h e contravariant v e c t o r components ( I , 5.5) and t h e e q u a l i t i e s ( 2 . 6 ) , we o b t a i n

288

The index ( i ) of t h e force components in t h e rectangular system of Cartesian coordinates shows t h a t they are f o r c e components d i r e c t e d along a tangent t o t h e coordinate l i n e xi of t h e c u r v i l i n e a r coordinate s y s t e m . Equations (2.3), on choice of t h e f o r c e determined by eqs.(2.8) take t h e f o l l o w b g form:

- (2.91,
(2.10)

where v ( i ) j are t h e components of t h e vector of t h e displacements caused i n t h e e l a s t i c medium by t h e concentrated force defined by eqs.(2.8) - (2.9). These components express t h e displacement vector i n t h e rectangular Cartesian mord i n a t e system OY,. Returning again t o t h e coordinate system x i , l e t u s f i n d t h e covariant components u(*) of t h e vector of displacements caused by t h e a c t i o n of t h e & W e obtain force l

(2.11)

A comparison of eqs.(2.11) w i t h (I, 6.3) shows t h a t t h e q u a n t i t i e s u [ , ) ~ found by u s possess p e c u l i a r tensor properties. These q u a n t i t i e s may be considered as covariant components of t h e vector a t t h e point N ( x i ). But being functions of t h e p a i r of p o i n t s M and N, they a r e components of t h e covariant Further, from e q ~ ~ ( 2 . 5and ) tensor of rank two, connected with these point*. (2.9) w e find:

(2.12)

Again a p p l w g t h e transformation formulas (I, 6.3), w e obtain t h e contrav a r i a n t components of t h e stress t e n s o r in t h e c u r v i l i n e a r coordinate system
~~

* We

s h a l l not dwell on t h e analogy between these q u a n t i t i e s and t h e so-called **intermediatent e n s o r components. Cf.I.Schouten and D.Struik, Introduction t o N e w Methods of D i f f e r e n t i a l Geometry, ONTI, 1939, p.29.
289

connected with t h e s h e l l .

W e r e c a l l t h a t i n a r e c t a n g u l a r C a r t e s i a n coordinate system t h e noninvariant equality

i s satisfied.
Equations (2.13) show t h a t t h e q u a n t i t i e s aJJ, are components of a t h i r d rank t e n s o r of two points. A t p o i n t M, t h e s e q u a n t i t i e s a r e components of t h e v e c t o r w i t h t h e s u b s c r i p t s ( i ) and (p). A t point N, however, they a r e compone n t s of a second-rank tensor. Let u s pass now t o t h e construction o f new s o l u t i o n s of t h e t h r e e dimensional problem of t h e theory of e l a s t i c i t y w i t h s i n g u l a r i t i e s . The solut i o n s we have considered f o r t h e homogeneous s t a t i c equations of t h e e l a s t i c i t y theory s a t i s f y t h e s e equations f o r a l l values of t h e coordinates, except for t h e coordinates of t h e p o i n t of a p p l i c a t i o n of t h e concentrated force. T h i s point i s singular. I n it, t h e displacements become i n f i n i t e of t h e o r d e r r-' as r -, 0, and t h e s t r e s s e s become i n f i n i t e of t h e o r d e r f 2 . Such s i n g u l a r i t i e s a r e encountered i n t h e Newtonian p o t e n t i a l function. On t h e b a s i s of t h e derived s o l u t i o n s we can f i n d a s e r i e s of new s o l u t i o n s of t h e homogeneous skat i c equations of t h e theory of e l a s t i c i t y with a continuous d i s t r i b u t i o n of t h e s i n g u l a r p o i n t s along a c e r t a i n line. Let u s assume t h a t this l i n e i s a segment of a s t r a i g h t l i n e , of l e n g t h 2Mh, where M > 1 . Let t h e coordinates of t h e middle of t h i s segment in t h e rectangular C a r t e s i a n coordinate system be & ( i 1 , 2 , 3). Then, t h e coordin a t e s o f t h e p o i n t s o f t h e segment can be expressed by t h e e q u a l i t i e s :
qi-=ci+ani

( i = 1, 2, 31,

(2.14)

where

C a: = I .
Here, ai a r e t h e d i r e c t i o n cosines of t h e segment, and
cy

i s t h e d i s t a n c e from

t h e middle of t h e segment with fixed p o s i t i v e d i r e c t i o n .

290

L e t us d i s t r i b u t e on this segment t h e f o r c e load of l i n e a r d e n s i t y q ( a ) , d i r e c t e d along one of t h e axes of t h e l o c a l coordinate basis. Assume t h a t i n t h e r e c t a n g u l a r Cartesian coordinate system y i , t h e d i r e c t i o n of t h e f o r c e s /2qo Then, t h e displacements q(cr)d(cr) coincides with t h e d i r e c t i o n of t h e axis OY, caused by t h e s e f o r c e s w i l l b e expressed by t h e e q u a l i t i e s

(2.15)

(2.lb) i f where t h e f u n c t i o n s Q t J ) ((Y)a r e determined by t h e r e l a t i o n s (2.la) t h e coordinates of t h e p o i n t of a p p l i c a t i o n of t h e f o r c e s a r e expressed by t h e e q u a l i t i e s (2.14).

The f u n c t i o n s w ( sa ~t i s f ~y t h ~e homogeneous s t a t i c equations of t h e e l a s t i c i t y theory f o r all v a l u e s of t h e coordinates, except t h e coordinates of t h e p o i n t s on t h e segment b e a r i n g t h e f o r c e load. These coordinates are expressed This segment i s t h u s a s i n g u l a r l i n e f o r t h e f u n c t i o n s w t J ) by eqsm(2.14). The l o a d d e n s i t y can b e s e l e c t e d under very wide assumptions r e l a t i v e t o t h e a n a l y t i c p r o p e r t i e s of t h e f u n c t i o n s q((Y)m

Consider t h e f u n c t i o n s q ( a ) causing t h e l o c a l perturbations. I n other words, l e t u s s e l e c t a function q(cy) such t h a t t h e displacements and s t r e s s e s due t o t h e r e s p e c t i v e load will r a p i d l y a t t e n u a t e w i t h i n c r e a s i n g d i s t a n c e of e shall t h e p o i n t N(yi ) from t h e segment on which t h e load i s d i s t r i b u t e d . W c a l l such a load a focusing load, s i n c e it w i l l subsequently permit us t o sepa r a t e p a r t of t h e wanted f i e l d of displacement in t h e neighborhood of t h e sing u l a r l i n e , and t o "liquidate" t h e r e s i d u a l For t h e c o n s t r u c t i o n of t h e focusing load we employ t h e method given elsewhere (Bibl.23b). O f course t h i s method cannot be considered optimum, b u t we w i l l n o t f u r t h e r d i s c u s s t h e methods of optimum choice of t h e focusing load.

F i r s t l e t u s analyze t h e conditions whose s a t i s f a c t i o n enables u s t o repres e n t t h e f u n c t i o n s Q(,), ( c y ) in t h e form of s e r i e s i n ascending p o s i t i v e powers of CY.

It w i l l be c l e a r from e q s 0 ( 2 . l a ) ( 2 , l b ) t h a t t h e s i n g u l a r i t y contained i n t h e functions k (j, k = 1 , 2, 3 ) depends on a f a c t o r of t h e form r-n The q u e s t i o n o f t h e p o s s i b i l i t y of expanding t h e s e f u n c t i o n s i n series i n ascending powers of CY t n e r e f o r e reduces t o an a n a l y s i s of t h e p o s s i b i l i t y of such expanFrom eqm(2.2) we f i n d sion f o r the function f n

* This

term i s borrowed from t h e book by C.Lanczos " P r a c t i c a l Methods of Applied Analysis", Fizmatgiz, 1961, p.220. The meaning of this term i s extended by US.

o r , in o t h e r words,

where

and 2~-C , OS cp = 2 r 0 - a
-9

i s t h e doubled scalar product of t h e v e c t o r s ro and

.-.
CY.

From eq.(d),

we f i n d

The expansion of F n ( a ) in a s e r i e s i n ascending powers of when t h e i n e q u a l i t y

CY

i s possible

(2.16b )

i s s a t i s f i e d , or, strengthening t h e i n e q u a l i t y ,

e find* Hence, w

Here, of course, w e have taken a p o s i t i v e value f o r J 2 . The u m value of is Mh. Consequently, a t t h e p o i n t s s a t i s f y i n g t h e condition

ICY!

See also M.A.Lavrentfyev The esthhate [ e q . ( 2 . 1 7 ) ] i s t o o high; cf. Sect.3. and B.V.Shabat, Methods o f t h e Theory of Functions of a Complex Variable, Gostekhizdat, 1951, pp.501 502.

292

t h e functions 8 ( powers of CY:

J)

(CY) can be represented by t h e convergent series i n ascending

v, k =.

1, 2, 3).

These series w i l l a b s o l u t e l y converge f o r a l l values of CY l y i n g on t h e segment (-MH, + M h p . On t h e b a s i s of eqs.(2.19), t h e f u n c t i o n s wc1) defined by /292 eqs.(2.15) can be represented by expansions of t h e form

(2.20)

where t h e c o e f f i c i e n t s m,, are defined by t h e formulas

(2.21)

L e t u s assume a t first t h a t t h e functions q ( o ) a r e everywhere bounded on t h e i n t e r v a l (a, +Mh) and are continuous, except f o r a f i n i t e number of p o i n t s of d i s c o n t i n u i t y of t h e f i r s t kind. Assume f u r t h e r t h a t on t h e c o n t i n u i t y int e r v a l s t h e function q ( a ) i s represented by polynomials o f degree N, where N i s f o r t h e time being an a r b i t r a r y number. Then, c o e f f i c i e n t s of t h e s e polynomi a l s can always be s e l e c t e d such t h a t t h e e q u a l i t i e s
I n n = 0 (I2 = 0,

1, 2 ,..., N - 1)

shall be s a t i s f i e d .
Equations (2.22) form a system of l i n e a r a l g e b r a i c equation from which t h e

N c o e f f i c i e n t s of t h e polynomial representing t h e function q ( a ) can be d e t e r c o e f f i c i e n t i f t h e function q ( a ) i s continuous mined i n terms of t h e (N + l)th W e shall consider t h e case of t h e disconover t h e e n t i r e i n t e r v a l (-Mh, +a).
tinuous f u n c t i o n q ( a ) somewhat l a t e r .

It follows from eq.(2.21) t h a t t h e c o e f f i c i e n t mN i s of a r e l a t i v e o r d e r Consequently, t h e displacen o t lower than t h e o r d e r of t h e q u a n t i t y (Mh)N+l. ments wt1) k defined by eqs. (2.20) w i l l be of an o r d e r not lower than t h e o r d e r of t h e r a t i o (Mh)N+l : r : + l over t h e e n t i r e region in which ro s a t i s f i e s t h e i n -

* On a b s o l u t e

and uniform convergence of t h e s e expansions, cf., f o r instance, E.T.Whittaker and G.N.Watson, Course i n Modern Analysis, Vol.11, Gostekhisdat, 1934, pp. 91-92,

293

e q u a l i t y (2.17). Since, i n this region, t h e i n e q u a l i t y

Mfi -<r,,

1 2,5 '

i s s a t i s f i e d , it follows from t h e above t h a t i n t h i s region we can construct displacements w ( i ) k which are n e g l i g i b l y s m a l l in a b s o l u t e magnitude. L e t u s i n v e s t i g a t e this question i n more d e t a i l , considering t h e concrete construction of a function q ( a ) with t h e above-noted focusing properties.

(*,

Let u s consider t h e d e n s i t y nf t h e load q ( a ) determined over t h e i n t e r v a l +~ha ) s follows: a ) t h e function q ( a ) i s piecewise-continuous over t h e interval

(-a, +Mh);

+ eh), (-h

b ) t h e function q(m) i s zero over t h e i n t e r v a l ( e h , +eh); (h e ,. -h + s h ) , where L << 1; c ) t h e function q ( a ) i s normed by t h e condition

- sh,

h +

/zq3

(2.23)

Let u s denote t h e value of t h e piecewise-continuous function q ( a ) over t n e b t e r v a l (h + sh, Mh) by % ( C Y over ); t h e i n t e r v a l (eh, h c h ) by % ( C Y over ); ,the i n t e r v a l (-h + ch, -sh) by %(CY);over t h e i n t e r v a l (-h - ch, -Mh), by 9+ (4.

L e t u s now impose on t h e functions Q (CY), Q ( C Y ) t h e condition t h a t 0) be self-balanced: t h e load on t h e i n t e r v a l s (0, Mh) and (-a,

...,

a"q (a)da = 0;
0 0

(n=O, 1 , 2,

... , N ) .
From t h e s e

Consider, f o r example, t h e first group of conditions (2.24). conditions, it follows t h a t

294

( 2 . 2 5 )

where pn ( t ) and pn

(CY)

a r e c e r t a i n polynomials.

Assume t h a t t h e f u n c t i o n %(a) is assigned. For d e f i n i t e n e s s and c e r t a i n s i m p l i f i c a t i o n s in t h e subsequent c a l c u l a t i o n s , l e t us put

Let u s perform t h e s u b s t i t u t i o n of t h e v a r i a b l e s b r i n g i n g t h e i n t e g r a t i o n intervals in e q . ( 2 . 2 5 ) t o t h e standard i n t e r v a l (-1, + l ) . L e t u s pt

t=- I
a=-

It (1

- 28) (.+
2

1-22E

1.
'
1- 2 E

It (1 - 2s)

(.+

1- 2 E

Let us s e l e c t t h e polynomial pn ( t ) such t h a t , on this s u b s t i t u t i o n of t h e vari a b l e t, it s h a l l be transformed i n t o t h e Legendre polynomial Pn(z):

Then t h e polynomial &(a) i s transformed as follows:

pn (a) = P,(z') =P, z-

1-2E

)-

Then e q . ( 2 . 2 5 )

takes t h e following form:

.r 1
+I

P,,(:)Q1 ( z ) ~ B = -1 -1

295

Equation (2.27) determines t h e c o e f f i c i e n t s of t h e expansion o f t h e funct i o n Ql ( 2 ) i n Legendre polynomials, i f , a s already specified, we p r e s c r i b e t h e function (a)

Assume t h a t t h e function Q l ( z ) can be approximately represented by t h e Then this polynomial will be represented by an polynomial GN)(z)of degree N. expansion in Legendre polynomials

QiN)( zj where
I 1

BnPn (z),
n-0

By i n c r e a s i n g N i n formula (2,28), we o b t a i n an i n f i n i t e sequence o f functions

QdN) ( 2 ) .

Speaking generally, this sequence can be divergent, b u t it always r e t a i n s Let us a c e r t a i n mechanical meaning. N o w l e t u s consider t h e load %("(t). c o n s t r u c t t h e function

This function may be regarded a s t h e mean value o f t h e r e s u l t a n t of f o r c e s with a d i s t r i b u t i o n d e n s i t y q l ( N ) ( t ) applied t o 1/N p a r t of t h e i n t e r v a l over which these forces a r e distributed.

If t h e r e exists l i m Pl(N) ( t ) = P ( t ) n o t equal t o zero, then we may a s s e r t Nt h a t t h e s i n g u l a r i t i e s corresponding t o t h e l i m i t i n g values o f t h e d e n s i t y (N) ( t ) are t h e result of t h e continuous d i s t r i b u t i o n of concentrated f o r c e s of f i n i t e magnitude over t h e i n t e r v a l (h + ch, 2h ch). If l i m P , ( N ) ( t ) does

N-

a .

n o t e x i s t , on this i n t e r v a l t h e r e a r e d i s t r i b u t e d s i n g u l a r i t i e s of t h e f o r c e f i e l d of t h e type of f o r c e dipoles, etc.. Thus t h e limit of t h e sequence /295 QlcN) ( 2 ) determines t h e s i n g u l a r i t i e s of t h e f o r c e f i e l d constructed by us beyond t h e limits of t h e interval (-h, +h). The f u n c t i o n Q(N) ( t )i s constructed s i m i l a r l y t o t h e function %(N) ( t ) i f t h e function %(cy) i s prescribed. This exhausts t h e question of construction of t h e function q ( a ) over t h e i n t e r v a l ( a h , +Mh), O f course, t h i s extension of t h e c l a s s of f u n c t i o n s q(cy) demands a corresponding extension of the' i n t e g r a b i l i t y conditions of t h e s e functions, W e w i l l n o t f u r t h e r d i s c u s s this ques-

296

t i o n and w i l l assume t h a t t h e necessarg .&ension of t h e i n t e g r a b i l i t y condit i o n s can be found, Thus, by s e l e c t i n g t h e f u n c t i o n s 'qa(cy )-and q 3 ( )~such %hat t h e condition (2.23) i s s a t i s f i e d , we can c o n s t r u c t t h e functkons Q (cy) and (CY),s t a r t i n g But then, according t o e q ~ ~ ( 2 . 2 0 ) (2,ZL), o u t from t h e conditions (2.24). t h e displacements w C i l k will vanish beyond t h e sphere S of r a d i u s ro > 2.5 Mh, I n t h i s region e x t e r n a l t o t h e sphere S , t h e s t r e s s e s corresponding to t h e displacements w ( ~ and ) ~ t h e d e r i v a t i v e s of those q u a n t i t i e s w i l l a l s o vanish. Withi n t h e sphere S t h e displacements found by us w i l l s a t i s f y t h e c o n t i n u i t y cond i t i o n s and t h e equations of e l a s t i c i t y theory everywhere, except on t h e l i n e on which t h e r e are s i n g u l a r i t i e s ,

The d e n s i t y q(cY) with t h e considered p r o p e r t i e s focuses t h e f i e l d of d i s placements and s t r e s s n e a r t h e segment on which it i s d i s t r i b u t e d ,

If n e i t h e r 'limQ1(N) (z) nor l i m Qq CN) (z) exist, then e q ~ ~ ( 2 . 2 2 w ) i l l be

N-, OJ N+OJ s a t i s f i e d o n l y f o r l i m i t e d values of N, In this case, t h e f i e l d of displacements and s t r e s s e s w i l l not disappear beyond t h e limits of t h e sphere s, i,e,, t h e focusing p r o p e r t i e s of t h e load q(cy) w i l l be weakened. T h i s can apparently i s bounded on t h e i n t e r v a l (-Mh, t a k e place everywhere on c o n d i t i o n t h a t

Iq(c~)l

+Mh),
I n conclusion, l e t u s make an approximate evaluation of t h e v a r i a b i l i t y of t h e displacements wCI,k and of the. corresponding s t r e s s e s in t h e neighborhood of a segment of t h e s t r a i g h t line over'which a load of d e n s i t y q(a) i s d i s t r i These formulas can be repre(2.lb). buted, L e t u s r e t u r n t o eqsm(2.1a) sented in t h e following form:

B(i)i

=- [C

A r

+ B cos2(ryl)];
B=

o(i) k

=-cos (ryJ cos (ryk), ..

AB r

(2.30)

where

A=-. 1

24~G '

3 ; 2 (1 - v)

c= 9 - 12v
2(1 - v ) .

It follows t h a t t h e displacement y e c t o r components determined by eqs.(2,30)

can b e represented a s

/a6

(2.31)

where

@(CY) i s a c e r t a i n bounded f u n c t i o n of t h e parameter

CY,

O f course, this f u n c t i o n a l s o depends on t h e coordinates of t h e p o i n t N ( y i ) a t which t h e displacements a r e determined and on t h e coordinates & of t h e ten-

297

t e r of t h e segment on which t h e load i s placed.

Further, applying t h e theorem of t h e i n t e g r a l mean, we f i n d
h
i ,

where E l i e s i n t h e i n t e r v a l (a, b).

If t h e point N(yi ) a t which t h e displacements are determined l i e s o u t s i d e b t h e l i n e s bearing t h e load, then t h e i n t e g r a l w i l l b e nonsingular. a r Integrating, we f i n d f r o m eq. (p),

(2.32)

where p = -ro cos cp and r ( a ) and r ( b ) are t h e d i s t a n c e s from t h e respective p o i n t s a and b t o t h e p o i n t N(yi ).
If cp i s zero o r n , then t h e point N, a t which t h e displacements w ( l ) k a r e determined, l i e s on t h e s t r a i g h t l i n e bearing t h e load. Two cases must be d i s tinguished here. If t h e point N does not l i e on t h e p a r t of t h e s t r a i g h t l i n e over which t h e load i s d i s t r i b u t e d , then t h e i n t e r v a l under consideration, as above, will not be singular.

In t h e case where t h e p o i n t N does l i e on t h e i n t e r v a l (a, b), this i n t e If t h e g r a l will be improper, b u t w5th an e x i s t i n g Cauchy p r i n c i p a l value. p o i n t N coincides with one end of t h e i n t e r v a l (a, b ) , then t h e i n t e g r a l
be divergent. A s w i l l be seen from eqs.(2.32) a have ~ logarithmic f e a t u r e s . p o i n t s t h e functions w ( ~ )
q r

da f will

and (2.15),

a t these

O f course a l l these conclusions a r e v a l i d only f o r narrow c l a s s e s of funct i o n s q(a), which i n p a r t i c u l a r , admit of t h e a p p l i c a t i o n of t h e theorem of t h e (a,) such e can always s e l e c t f u n c t i o n s 92 ( a , ) and i n t e g r a l mean. However w t h a t t h e conditions of a p p l i c a b i l i t y of t h e theorems of c l a s s i c a l a n a l y s i s will and &(CY), they l i e on p a r t s of the/- be s a t i s f i e d . A s for t h e f u n c t i o n s %(CY) s t r a i g h t line running o u t s i d e t h e region f i l l e d by t h e material of t h e s h e l l , which permits us t o arrange them more a r b i t r a r i l y than t h e functions %(CY)and

93 (a).

298

Consider now t h e v a r i a b i l i t y of t h e s t r e s s tensor components corresponding t o t h e displacements wcl)k i n t h e neighborhood of a singular line. Equations (2.4a) (2.4d) can be put i n t o t h e following form:

q i ) ik

cos ( r y k )[l + B cos2 (ry,)], r2

(2.33~ )

where
1- 2 v _ -

A=-

8x(1 - v )

It will be seen from eqs.(22.3a) - (2.33d) t h a t t h e components of the stress t e n s o r can be described by the following formula:

where $ c l ) r s i s a function of t h e parameter cy, and, of course, of the coordir , and of t h e parameters den a t e s of t h e point N(yl ) of t h e s t r e s s f i e l d termining t h e position of t h e load-carrying segments. Applying again t o t h e theorem of t h e i n t e g r a l mean, l e t u s consider the integral

The meaning of t h e notation here used will be c l e a r from t h e above exposition. The i n t e g r a l

s- i s nonsingular a 2
d c y

i f t h e point N ( y , ) l i e s outside t h e

29 9

s t r a i g h t l i n e b e a r i n g t h e load q(a) o r l i e s on a p a r t of t h e s t r a i g h t l i n e f r e e from load. If t h e p o i n t N(yi ) l i e s on a segment of t h e s t r a i g h t l i n e over which t h e load i s d i s t r i b u t e d with a d e n s i t y q(a), then this i n t e g r a l w i l l be improper and divergent. Assuming a t first t h a t t h e p o i n t N(yi ) l i e s o u t s i d e t h e segment o f t h e s t r a i g h t line on which t h e function q(a) d i f f e r s from zero, we f i n d from eq. that
h

t a n - 1--

0 - I-, cos r0 sin 'p

'p

-tan

- 1 n --- ro cos cp
--

Yo sin

cp

(2.35)

This e q u a l i t y confirms t h e above a s s e r t i o n s on t h e p r o p e r t i e s of t h e int e g r a l under consideration, i f w e i n v e s t i g a t e t h e l i m i t passage of t h e p o i n t N(yi ) on t h e s t r a i g h t l i n e bearing t h e load. I n p a r t i c u l a r , when t h e p o i n t N(yi ) approaches t h e segment w i t h t h e nonzero load d e n s i t y q ( c y ) , t h e i n t e g r a l ( 9 ) i n c r e a s e s without l i m i t , b u t not more r a p i d l y than t h e function r-' a s r 0.

All t h e above e s t a b l i s h e s t h e p r o p e r t i e s of v a r i a b i l i t y of t h e s t r e s s i n5 t) he f i e l d corresponding t o t h e displacements w ( i ) k (htermined by ~ ~ ~ ( 2 . 1 neighborhood of t h e s i n g u l a r l i n e bearing t h e load of d e n s i t y q(cy). W e emphas i z e again t h a t t h e a n a l y s i s given here by no means exhausts all p r o p e r t i e s of t h e displacement f i e l d s w ( i ) k and of t h e corresponding s t r e s s f i e l d s , s i n c e it has been performed under simplified i d e a s a s t o t h e a n a l y t i c p r o p e r t i e s of t h e function q(cr ).
Section 3. I n t e g r o d i f f e r e n t i a l and I n t e g r a l Equations of t h e S t a t i c s of S h e l l s . w i t h Focusing Kernels

W e give below a method of solving t h e boundary problems of s t a t i c s of W e will s h e l l s , r e l y i n g on t h e theorem of work and r e c i p r o c i t y (11, Sect.12). confine t h e d i s c u s s i o n f o r t h e time being t o t h e formulation of t h i s theorem, known from t h e l i n e a r theory of e l a s t i c i t y . The e n t i r e method can be considered as a development of t h e well-known method of Somiglianox-. It i s well known t h a t t h e theorem of work and r e c i p r o c i t y , o r t h e Reciproc a l Theorem, makes it p o s s i b l e t o e s t a b l i s h an i n t e r r e l a t i o n between two f i e l d s o f displacements and S t r e s s e s induced i n an e l a s t i c body by two systems o f e s h a l l d i s t i n g u i s h t h e main and a u x i l i a r y f i e l d s of f o r c e s applied t o it. W displacements, s t r e s s e s , and forces.

* A.Love,

Mathematical Theory of E l a s t i c i t y , ONTI, 1935.

W e w i l l apply t h e term *%basic" t o t h e f i e l d s of displacements and stresses due t o a load a c t i n g on t h e s h e l l i n accordance with t h e conditions of t h e boundary problem t o b e considered. These f i e l d s are u s u a l l y unknown and must b e determined. The n o t a t i o n of t h e components of t h e v e c t o r of b a s i c displacements and t h e components of t h e t e n s o r of b a s i c stresses are known from t h e earlier Chapters of this book.

L e t u s p a s s t o t h e consideration of t h e a w d l i a r y d i s p l a c e m e n t and stress f i e l d s and t o t h e corresponding s u r f a c e and body forces. Let us again con- /299 s i d e r an unbounded e l a s t i c medium and imagine t h a t part of t h i s medium i s t h e s h e l l we a r e studying. L e t t h e unbounded mediumbe under t h e a c t i o n of f o r c e s applied t o a c e r t a i n segment (-Mh, +Mh) of a s t r a i g h t l i n e i n t h e manner i n d i cated i n t h e preceding Section.
Let u s d i r e c t t h e segment (-Mh, +Mh) of t h e load-carrying s t r a i g h t l i n e along t h e normal t o t h e middle s u r f a c e of t h e s h e l l , L e t u s assume t h a t a p o i n t l y i n g on t h e middle s u r f a c e corresponds t o t h e zero value of t h e paI amet e r cy on t h e load-carrying segment. Let u s f u r t h e r assme t h a t t h e d i s t r i b u t i o n d e n s i t y of t h e l o a d s q(w) i s determined by conditions ( a ) and ( b ) of t h e preceding Section, and a l s o t h a t it s a t i s f i e s eqs.(2.23) (2.24). Under t h e s e conditions, two segments of t h e s i n g u l a r l i n e on which t h e function q ( u ) does n o t vanish will l i e i n s i d e t h e e note t h a t under conditions ( a ) and ( b ) of t h e preceding Section t h e shell. W load-carrying segments do not i n t e r s e c t t h e middle surface nor t h e boundary s u r f a c e s of t h e s h e l l . These s u r f a c e s are free from s i n g u l a r p o i n t s both of t h e displacement f i e l d and of t h e stress f i e l d caused by t h e load d i s t r i b u t e d on t h e s i n g u l a r l i n e .

Let u s assum2 t h a t t h e f o r c e s applied t o t h e singular l i n e are d i r e c t e d along t h e v e c t o r ei of t h e l o c a l coordinate bases, where t h e indzx i i s fixed. W e r e c a l l t h a t , under t h e assumptions adopted by us, t h e v e c t o r e d i r e c t e d along t h e norm& t o t h e undefomned middle-surface has t h e same d i r e c t i o n as t h e segment of t h e s t r a i g h t 1b.e bearing t h e load.
W e s h a l l apply t h e term "auxiliary" t o those stresses created b o t h i n t h e s h e l l , and i n a p a r t d i m , by t h e load of t h e above-mentioned s i n g u l a r e x i s t i n a s h e l l c u t o u t of t h e unbounded e l a s t i c forces, determined from t h e known s t r e s s e s by t h e must b e applied t o t h e s h e l l .

f i e l d s of displacements and of t h e unbounded e l a s t i c mel i n e . For t h e s e f i e l d s t o medium, a system of s u r f a c e formulas (11, 8.2a 8.2b),

The surface f o r c e s s o found, t o g e t h e r with t h e f o r c e s d i s t r i b u t e d on t h a t p a r t of t h e s i n g u l a r l i n e l y i n g within t h e s h e l l , form t h e system of a u x i l i a r y loads.

L e t u s consider t h e a n a l y t i c expressions f o r t h e auxiliarg displacements i n t h e c u r v i l i n e a r system of coordinates x1 connected with t h e s h e l l .

To avoid d i f f i c u l t i e s i n determining t h e f i e l d of auxiliary displacements in t h e c u r v i l i n e a r coordinate s y s t e m , l e t u s f i r s t use a rectangular Cartesian

coordinate s y s t e m , and then apply t h e formulas of transformation of t h e compone n t s of t e n s o r q u a n t i t i e s .

W e s h a l l r e t a i n t h e e a r l i e r n o t a t i o n f o r t h e load d e n s i t y q(CY). The vect o r os t h e corresponding f o r c e i s d i r e c t e d along t h e t a n g e n t s t o t h e coordinate L e t u s f i n d t h e components of/300 l i n e e, of t h e c u r v i l i n e a r coordinate s y s t e m . t h e v e c t o r d e n s i t y o f t h e load q(J) (CY) i n t h i s system:
q j ( a ) == q
(2) 2 ;

(i, j = 1, 2, 3).

(3.1 1

gous t o eq.(2.9)

Passing t o r e c t a n g u l a r Cartesian coordinates, we o b t a i n a r e l a t i o n analof o r t h e v e c t o r d e n s i t y components:

The f i e l d of displacements due t o t h e load w i t h t h e v e c t o r d e n s i t y p t i ) k x t h e r e c t a n g u l a r Cartesian coordinates will be determined by f o n m l a s q s . ( 2 . 1 0 ) and (2.15): analogous t o e
X(CY) i n

(3.3)
(i, j = 1, 2, 3).

Let u s r e t u r n now t o t h e c u r v i l i n e a r coordinates. The covariant compone n t s of t h e v e c t o r of a u x i l i a r y displacements a r e expressed by an e q u a l i t y analogous t o t h e r e l a t i o n (2.11):

(3.4)
(i, j ,
12,

= ;

1, 2, 3).

The s i g n s of swmnation over k and p a r e omitted. Equation (3.4) can be s i m p l i f i e d by using a moving Cartesian system of coY , coincides with t h e s t r a i g h t l i n e bearo r d i n a t e s and assuming t h a t t h e a x i s O i n g t h e load, d i r e c t e d , as already s t a t e d , along t h e normal t o t h e undeformed middle surface, Le., along t h e coordinate l i n e 3 of t h e c u r v i l i n e a r coordinate system. I n this case,

302

and eq.(3.4)

t a k e s t h e following form:

( 3 . 5 1

T h i s expression of a u x i l i a r y displacements i s n o t i n v a r i a n t ; t h e diaplacements w ( J ) k a r e defined by e q ~ ~ ( 2 . 1 5 ) . The expressions of t h e stress t e n s o r components and of t h e components of t h e surface f o r c e s defined according t o /3Ol t h e components v c i ) from Hookefs l a w w i l l n o t be given hare.

Applying t h e Reciprocal Theorem t o t h e b a s i c and a u x i l i a r y systems of d i s placements and loads, we f i n d

+h

(3.6)

where t h e SJ(,) denote t h e components of t h e a u x i l i a r g s u r f a c e forces, XJ and FJ a r e t h e components of t h e s u r f a c e and body f o r c e s of t h e b a s i c system, w h i l e u , a r e t h e covariant components of t h e v e c t o r of t h e p r i n c i p a l displacement. The integrals

JJ
( S I

extend over t h e boundary and contour s u r f a c e s of t h e s h e l l and

t h e integrals

lJ[
I V)

extend over i t s volume.

The expression

-h

s"

q (a) ui (3) da = ui (.d)

(j = 1, 2)

(3.7)

can be regarded t h e average value, with t h e weight q(cr), of t h e displacement v e c t o r components q. This q u a n t i t y i s a f u n c t i o n of t h e coordinates of t h e point M ( x J ) i n t e r s e c t e d by t h e s t r a i g h t line b e a r i n g t h e auxiliary load, and t h e middle s u r f a c e of t h e s h e l l .

303

I n connection with eq. (2.23) t h e q u a n t i t i e s can b e conventionally regarded as components o f t h e displacements o f t h e two-dimensional continuum stud i e d i n t h e s h e l l theory.
The function q ( m ) can always be taken such t h a t t h e condition

<

will be s a t i s f i e d , where q( (x*) are t h e displacements o f t h e p o i n t & (x*) lyi n g on t h e middle surface, and A i s a constant.
The condition ( 3 . 8 ) assures t h e p o s s i b i l i t y of an approximate s u b s t i t u t i o n of t h e i n t e g r a l (3.7), Le., o f t h e averaged displacements &, by t h e displacements of t h e middle s u r f a c e of t h e s h e l l . This solves p a r t of t h e general problem of reduction of t h e three-dimensional problem of t h e theory of elastic i t y t o t h e two-dimensional problem of t h e t h e o r y of s h e l l s .

up)

W e s h a l l now i n d i c a t e an elementary method of constructing t h e function Assume t h a t t h e displacement q(cy) permitting us t o s a t i s f y condition (3.8). u, (cy) can be approximated by t h e polyonomial:

(3.9)
Then, t h e averaged displacements

( a ) can be represented i n t h e form:

(302

(3.10)
where

(3.11)

cyJ

are t h e p o s i t i v e powers contravariant vector1 )

f t h e parameter a (but n t o f t h e component

of t h e

O n t h e b a s i s of t h e p r o p e r t i e s o f t h e function q(cy), considered i n Sect.2, l e t us s e l e c t t h e functions a , ( c y ) and q , (cy) i n t h e form of polynomials such t h a t t h e conditions
11 J. = z0

( j = 1, 2 ,..., N ) .

(3.12)

s h a l l be s a t i s f i e d .
If we confine o u r s e l v e s t o t h e r e l a t i v e accuracy adopted i n t h e c l a s s i c a l
301:

t h e o r y and i n Chapter I 1 1 of this book in s e t t i n g up t h e system of d i f f e r e n t i a l equations of motion of an element of t h e s h e l l , t h e n it i s s u f f i c i e n t to p u t

L e t u s c o n s t r u c t t h e polynomials 9 , (a)and g, ( w ) such t h a t t h e f u n c t i o n s q(w) s h a l l be even over t h e i n t e r v a l (-h, +h). I n this case, even under condiq , ( 3 . 1 0 ) , we s h a l l have nonvanishing cot i o n (3.13) i n t h e right-hand s i d e of e e f f i c i e n t s f o r t h e terms containing h5 and higher o r d e r s of h, and t h e c o e f f i c i e n t s of u y w i l l a l s o be nonvanishing. It i s n o t hard t o see t h a t f o r this it i s s u f f i c i e n t t o put

q2 (a) = a,

a +2 a; h

q3 (a)= n o -

n 2 a.

(3.14)

For d e f i n i t e n e s s we s h a l l assume that t h e parameter e i n t h e c o n d i t i o n s ( a ) Then, w e obtain and ( b ) imposed on t h e function q(w) i n Sect.2 i s 0.25.

All t h e c o e f f i c i e n t s of t h e nJ w i t h odd i n d i c e s j v a n i s h i d e n t i c a l l y . Equati n g t h e c o e f f i c i e n t Q , according t o t h e condition (2.23), t o unity, and t h e /303 c o e f f i c i e n t m, t o zero, we f i n d t h e c o e f f i c i e n t s a , and a,, Equation (3.14) t a k e s t h e following form:

W e f i n d t h e f u n c t i o n s %(cy) and %(cy) from s e l v e s t o t h e i r approximate r e p r e s e n t a t i o n

W e also n o t e t h a t this elementary method of c o n s t r u c t i n g t h e f u n c t i o n s q(m) involves a n a l y t i c r e s t r i c t i o n s imposed on t h e components u , of t h e displacement vector. These r e s t r i c t i o n s are expressed by t h e assumption t h a t an approximate 305

r e p r e s e n t a t i o n of t h e s e components by polynomials according t o eq.(3.9) i s poss i b l e . O f course, by employing more general methods of c o n s t r u c t i n g t h e funct i o n s q(a,) than t h o s e j u s t described, we should be a b l e t o eliminate t h e redund a n t a n a l y t i c r e s t r i c t i o n s imposed on t h e displacement v e c t o r components q. W e will n o t d i r e c t o u r i n v e s t i g a t i o n s d o n g t h i s l i n e and r a t h e r confine o u r s e l v e s t o a result o f t h e same r e l a t i v e accuracy as that given i n Chapter 111. The approximate r e l a t i o n (3.9) will enable u s then t o o b t a i n a number of r e l e v a n t conclusions by r a t h e r elementary means. Let u s r e t u r n t o

e q . ( 3 . 6 ) .

Since we can now approximately put

we f i n d from eq.(3.6)

(i, j = 1, 2, 3j.

where t h e s u p e r s c r i p t s (0) show t h a t t h e a u x i l i a r y system of displacements corresponds t o t h e l o a d on t h e s i n g u l a r l i n e , necessary t o determine t h e c o e f f i cients q ( ) .

W e s h a l l now show t h a t by varying t h e function q(a) w e can, t o within required accuracy, f i n d t h e c o e f f i c i e n t s I$) q @ ) e t c . , without having t o d i f f e r e n t i a t e eqs. (3.9).

i s s u f f i c i e n t t o take t h e functions

Indeed, f o r determining t h e c o e f f i c i e n t q)with t h e necessary accuracy,it Q (a,) and 93 (a,) as follows:

q2(a) = q3(a) = ala

a3a3

(3.18)

, from t h e conditions and t o determine t h e c o e f f i c i e n t s al and a

I n this case, a l l c o e f f i c i e n t s o f nJ with even s u b s c r i p t s j vanish.

/304

Putting, as before, e = 0.25, we f i n d t h e c o e f f i c i e n t s a, and a , from conThe f u n c t i o n s Q (a,) and g, ( c y ) w i l l be of t h e form d i t i o n s (3.19). (3.20)

The c o e f f i c i e n t from eq. (3.6):

4) will be

determined by t h e following equation r e s u l t i n g

( i , j = l , 2, 3).

where t h e index (1)has a meaning s i m i l a r t o t h a t of t h e s u p e r s c r i p t ( 0 ) in eq. (3.17). Obviously, f o r a n a r b i t r a r y c o e f f i c i e n t ulk) of t h e function q(cy), we can s e t up t h e equation

,after

a suitable selection

(i, j = =1, 2, 3 ; /2=0 1, 2 , . . . , N ) .

It i s n o t d i f f i c u l t t o prove t h a t e q ~ ~ ( 3 . 2 2y ) i e l d a new s o l u t i o n of t h e problem o f reduction of t h e three-dimensional problems of t h e s t a t i c s of an e l a s t i c body t o t h e two-dimensional problems of s h e l l theory. This s o l u t i o n does n o t r e q u i r e s a t i s f a c t i o n o f t h e condition t h a t t h e components of t h e vece w i l l explain t h e t o r s of t h e f o r c e s a c t i n g on t h e s h e l l be d i f f e r e n t i a b l e . W d e t a i l s of t h e new reduction method below,
L e t u s study t h e i n t e g r a l s on t h e right-hand s i d e s of eqs.(3.17),

and (3.22).

The i n t e g r a l s of t h e form

sss
f

(3,2l),

F ~ V dV !~.
wi

should b e considered as

VI

prescribed f u n c t i o n s of t h e coordinates of t h e Roint M ( x * L e . , as t h e p o i n t s of i n t e r s e c t i o n between t h e s t r a i g h t l i n e bearing t h e a d d i t i o n a l load and t h e middle surface. The i n t e g r a l s

>,

Jl

over t h e surface S of t h e s h e l l can be

(S)

represented i n t h e form of sums of i n t e g r a l s over t h e boundary s u r f a c e s S of Since t h e load on t h e boundary sur$%es t h e s h e l l and i t s contour s u r f a c e S,. of t h e s h e l l i s u s u a l l y known, t h e i n t e g r a l s of t h e form
JJXv;$dS,,
121

where t h e

s i g n (+) has been introduced i n s t e a d of t h e symbol S(*) t o shorten t h e fonmilas, must be considered known functions. The i n t e g r a l s

JJ S;i,jri,dS
(!

contain the/305

covariant components o f t h e required displacements on t h e boundary s u r f a c e s of t h e s h e l l , Finally, t h e i n t e g r a l s over t h e contour s u r f a c e S , of t h e s h e l l determine known and unknown functions.

3c7

The i n t e g r a l

SC1

1 s

XjgW

Wi

dS

i s a known function, if t h e f o r c e s a c t i n g on t h e

contour surface a r e prescribed. Usually t h e f o r c e s a c t i n g on t h e part of t h e contour surface t h a t is free from c o n n e c t i v i t y are known. O n the p a r t s of the contour surface with c o n n e c t i v i t y t h e s e f o r c e s are unknown, and t h e correspondi n g p a r t of t h e i n t e g r a l

Js
S ,1

Xjv$j dS

will contain d e r i v a t i v e s of t h e componSimi-

e n t s of t h e required displacements of t h e p o i n t s of t h e contour surface.

larly, t h e i n t e g r a l

JJS{,k,)jujdS i s decomposed i n t o two parts.

SC )

That p a r t of

t h e i n t e g r a l taken over t h e region of t h e contour s u r f a c e w i t h prescribed d i s placement components i s a h o v m function. The o t h e r p a r t of t h i s i n t e g r a l cont a i n s components of t h e displacement sought. O n t h e b a s i s of t h e p r o p e r t i e s of t h e a u x i l i a r y displacements considered i n Sect.2, it can be a s s e r t e d t h a t t h e s u r f a c e i n t e g r a l s w i l l n o t be s i n g u l a r if t h e point M(xj ) does n o t l i e on t h e contour surface. If t h e p o i n t M ( x j ) does l i e on t h e contour surface, then t h e s e i n t e g r a l s w i l l be improper, b u t convergent. The volume i n t e g r a l w i l l l i k e w i s e be improper b u t convergent. Bearing a l l t h e above i n mind, l e t u s now introduce t h e n o t a t i o n

-JJS;?,)u;dS
(11)

(i, r = l , 2,3 ; j = l , 2; k = 0 ,

I , 2, . . . , N ) ;
(3.23

where

dj
(1)

i s t h e i n t e g r a l over t h a t p a r t of t h e contour s u r f a c e with pre-

scribed components of t h e f o r c e s of t h e b a s i c system, and

.lJ (11)

rf

i s the inte-

g r a l over t h a t p a r t of t h e contour s u r f a c e with t h e prescribed components of displacement of t h e b a s i c system. Then, eqs.(3.22) t a k e t h e form:

(i,

I=

1,

2, 3; j = 1, 2; k = O , 1, 2,.. . , N ) ,

(3.24)

and enable u s t o f i n d t h e approximate expression f o r t h e displacement v e c t o r components a t an a r b i t r a r y p o i n t of t h e s h e l l .

Making use of eq.(3.9), n a t e z, we o b t a i n

a f t e r s u b s t i t u t i n g t h e parameter

cy

by t h e coordi-

(i, 1=1,

2, 3; j = 1 , 2).

(3.25) The l a t t e r stateThe e q u a l i t i e s (3 25 ) are approximate and noninvariant. ment i s connected with t h e f a c t t h a t t h e s c a l a r products i n these e q u a l i t i e s cannot be considered as absolute s c a l a r s . The p r o p e r t i e s o f t h e i n t e g r a l s ent e r i n g i n t o eqs.(3.25) have already been discussed. Thus, f o r an approximate determination of t h e f i e l d of displacements i n t h e s h e l l we must f i n d t h e components of t h e displacements sought on t h e boundary s u r f a c e s of t h e s h e l l and on p a r t ( I ) of t h e contour surface, as w e l l as t h e components of t h e b a s i c s y s t e m of f o r c e s on p a r t (11) of t h e contour surface.

To solve t h e problem, we m u s t set up equations i n t h e above unknowns. Seve r a l versions f o r constructing t h e required systems of equations may b e given. W e confine o u r s e l v e s i n t h i s Section t o two versions.
1 . Let u s put z = fi h i n eq.(3.25).

Introduce t h e n o t a t i o n

Then, from eq.(3.25), tions:

we f i n d t h e following system of i n t e g r o d i f f e r e n t i a l equa-

309

(i, r = l , 2, 3; j - 1, 2).
(3.2%

Let u s now assume t h a t t h e s y s t e m of auxiliary displacements i s due t o t h e a c t i o n of t h e focusing load considered i n Sect.2. Let t h e point M(xJ, 0) on h bordering i t s contour. t h e middle s u r f a c e l i e o u t s i d e t h e zone of width 2.5 M Then, i n eqs.(2,27a) (3.2%) we may omit t h e i n t e g r a l s over t h e contour surf a c e of t h e s h e l l , and confine t h e i n t e g r a t i o n region i n t h e i n t e g r a l s over t h e boundary s u r f a c e s t o t h e region l y i n g w i t h i n t h e c i r c u l a r cylinder of r a d i u s 2 Mh, with i t s axis coinciding w i t h t h e s t r a i g h t l i n e bearing t h e load of dens i t y q(cu).

The system of equations (3.2%)

(3.2%)

is now s i m p l i f i e d and t a k e s t h e

form:

(i,

Y=

1, 2, 3; j = 1, 2 ) .

where t h e regions of i n t e g r a t i o n ( + ) and (-) a r e bounded as noted above.

L e t u s assume t h a t t h e point H ( x ' ) l i e s in t h e zone of width 2 M h borderi n g t h e contour of t h e middle surface. Then, i n t h e i n t e g r a l s over t h e contour s u r f a c e s we must r e t a i n only t h e parts t h a t correspond t o t h e i n t e g r a t i o n over t h e region enclosed within t h e c i r c u l a r c y l i n d e r having a r a d i u s 2 and an a x i s coinciding with t h e s t r a i g h t l i n e bearing a load o f d e n s i t y q(cr)o

Thus t h e a p p l i c a t i o n of t h e focusing load permits u s t o r e s t r i c t t h e i n t e g r a t i o n region t o r e l a t i v e l y small regions with movable boundaries, varying t h e i r p o s i t i o n on displacement of t h e point M ( x 1 ) over t h e middle s u r f a c e of t h e s h e l l . I n t h e following Section, w-e make use of equations with t h e integ r a t i o n regions r e s t r i c t e d in this manner. 2. Let us now consider t h e second method of s e t t i n g up t h e system of i n t e g r o d i f f e r e n t i a l equations based on t h e a p p l i c a t i o n of e q s . (3.9) and (3.24). Assume t h a t t h e right-hand s i d e of e ~ ~ ( 3 . 9contains ) an expression ensuring t h e approximate displacement of t h e v e c t o r u, t o t h e middle s u r f a c e of t h e s h e l l . W e note t h a t t h e elements of a r e a on t h e boundary s u r f a c e s and on t h e middle s u r f a c e are connected by t h e r e l a t i o n s

where dS{*) a r e elements of a r e a of t h e boundary surfaces, dS i s an element of a r e a of t h e middle surface, and k, a r e t h e p r i n c i p a l curvatures of t h e middle s u r f ace. Making use of t h e approximation equation (3.9), l e t u s r e p r e s e n t t h e conponents of t h e s u r f a c e f o r c e s in t h e region (11)of t h e contour surface by t h e expansions

Based on t h e approximate e q ~ ~ ( 3 . 9 and ) (3.30) and on eq.(3.29), eqs. (3.24) t h e following form:

we give

where t h e i n t e g r a l s
(1)

and
(11)

a r e taken over those parts of t h e contour of

t h e middle s u r f a c e s of t h e s h e l l l y i n g i n regions ( I ) and (11) of t h e contour surface; t h e components am equal t o t h e v e c t o r components S o ) on t h e

F L g

311

boundary surface z = +h, which v e c t o r has undergone p a r a l l e l displacement t o t h e middle surface; t h e components T I ! { ' are equal t o - t h e v e c t o r components S$]'r on t h e boundary s u r f a c e z = -h, which v e c t o r has undergone p a r a l l e l displacement t o t h e middle surface+!-. /309

Let us introduce t h e notation:

-h

-h

(3.32a)

Then, we o b t a i n from eq.(3.31):

(3.33

T h e s i g n s of summation over (...p)

are omitted.

Equations (3.33) must be regarded as a s y s t e m of i n t e g r o d i f f e r e n t i a l equat i o n s with unknown functions q c k ) of a p o i n t of t h e middle s u r f a c e s of t h e s h e l l . These equations are t h e i n t e g r a l analogs of t h e d i f f e r e n t i a l equations considered i n Chapter 111. There is, however, a s u b s t a n t i a l d i f f e r e n c e between eqso(3.33) and t h e equations of Chapter 111. This d i f f e r e n c e l i e s i n t h e f a c t t h a t eqs.(3.33) do n o t contain t h e d i f f e r e n t i a t i o n operation f o r t h e components of t h e v e c t o r f o r c e s a c t i n g on t h e s h e l l . Consequently, t h e s e equations remain v a l i d even i n cases where concentrated f o r c e s are applied t o t h e boundarg s u r f a c e s of t h e s h e l l . Obviously, this remark a l s o a p p l i e s t o eqsO(3.27a) -* (3.2%).

Now l e t u s assume, a s above, t h a t t h e load on t h e s i n g u l a r l i n e i s focush bordering t h e contour of t h e ing. Then, o u t s i d e of t h e s t r i p of wid.th 2.5 M w) i l l be of t h e following form: middle surface of t h e s h e l l , e q ~ ~ ( 3 . 3 3

*W e

have taken advantage of t h e f a c t t h a t t h e s c a l a r product does not vary und e r p a r a l l e l displacement, and t h i s operation can be performed by s e p a r a t e l y d i s p l a c i n g t h e c o f a c t o r s (I, Sect.lO.1).

312

(i, r = l , 2, 3; j = 1 , 2;

k, p = o ,

1,

2 , - - - 1

N)

where t h e region ( C ) l i e s w i t h i n a c i r c u l a r c y l i n d e r of r a d i u s 2.5 M h and a x i s coinciding with t h e s t r a i g h t l i n e bearing t h e load of d e n s i t y q(a).

I f t h e p o i n t M ( x l ) of i n t e r s e c t i o n between t h e middle s u r f a c e and t h e / 3 l O load-carrying s t r a i g h t l i n e l i e s i n t h e s t r i p of width 2 . 5 M h bordering t h e cont o u r of t h e middle surface, t h e n t h e i n t e g r a t i o n region (C) and t h e segments of a r c of t h e contour t o which t h e c u r v i l i n e a r integrals e n t e r i n g i n t o eqs.(3.33) extend must be r e s t r i c t e d t o t h e i n s i d e of t h e c i r c u l a r c y l i n d e r having a radiu s 2.5 M h and an axis coinciding with t h e s t r a i g h t l i n e bearing t h e load q(cu).

The equations of t h e form (3.34) a r e close in t h e i r mechanical meaning t o t h e d i f f e r e n t i a l equations of Chapter 1 1 1 , since they d e s c r i b e t h e mechanical s t a t e of a small b u t f i n i t e p a r t of t h e s h e l l , while t h e d i f f e r e n t i a l equations d e s c r i b e t h e state of an element of t h e s h e l l . It can be predicted t h a t , by modifying t h e s t r u c t u r e of t h e focusing load and passing t o t h e l i m i t , we w i l l be a b l e t o f i n d t h e d i f f e r e n t i a l equations of equilibrium from equations analogous t o eqs.(3.34). Several concluding statements a r e made below:
a ) The preceding conclusions were based on t h e assumption of existence of a focusing load constructed by t h e method given i n Sect.2. W e s h a l l n o t invest i g a t e t h i s question f u r t h e r , s i n c e t h i s involves s e v e r a l problems of mathemati c a l a n a l y s i s of t h e same n a t u r e a s t h e c l a s s i c a l problem of moments-. These problems cannot be discussed here and w i l l be taken up i n t h e second volume of t h i s book. However, i n Sect.9 we w i l l give one of t h e o t h e r p o s s i b l e methods of constructing a focusing load, which involves no fundamental d i f f i c u l t i e s , e s p e c i a l l y questions of existence.

b ) I n s e t t i n g up t h e i n t e g r a l equations (3.28a) (3.2823) and (3.34), we r e s t r i c t e d t h e region of i n t e g r a t i o n t o i n t e g r a l s containing t h e wanted funct i o n s . The same r e s t r i c t i o n of t h e i n t e g r a t i o n regions can be c a r r i e d o u t i n t h e i n t e g r a l s containing t h e prescribed functions. These i n t e g r a l s e n t e r i n t o t h e right-hand s i d e of eq. (3.23).

It is, however, necessary t o determine f i r s t whether t h e r e a r e i n t e g r a l s over a region e x t e r n a l t o ( C ) and dominant r e l a t i v e t o t h e i n t e g r a l s over t h e This case may occur, f o r example, i f prescribed e x t e r i o r f o r c e s region (Z). a r e absent from t h e regions (E). Then, of course, t h e r e s t r i c t i o n of t h e integ r a t i o n region of i n t e g r a l s containing t h e prescribed f u n c t i o n s t o t h e regior! (C) cannot be applied.
c ) m u a t i o n s (3.28a) (3.28b) and (3.34) are i n t e g r a l equations with vari a b l e i n t e g r a t i o n limits, which are f u n c t i o n s of t h e p o i n t M ( d ) of i n t e r s e c t i o n of t h e s t r a i g h t l i n e b e a r i n g t h e load q(a) with t h e middle s u r f a c e of t h e shell.
2

N.1 .Akhiyeaer,

The C l a s s i c a l Problems of Moments.

Fizmatgiz, 1961.

313

O n displacement of t h e p o i n t

M ( d ) over t h e middle surface, t h e region@)

I n t h e boundary s t r i p of width 2.5 Mh bordering t h e contour o f t h e middle surface, t h e contour i n t e g r a l s e n t e r i n t o t h e equations, as a l r e a d y noted.

will cover t h e e n t i r e middle surface.

d ) If t h e focusing load i s constructed approximately, eqs.(3.28a) (3.28b) and (3.34) should b e considered as approximate. If t h e focusing load i s cons t r u c t e d with s u f f i c i e n t accuracy, it will be p o s s i b l e t o r e f i n e t h e value of t h e radius r , of t h e sphere which d i s s e c t s t h e region (T) on t h e middle s u r f a c e of t h e s h e l l . For t h i s purpose, and on t h e b a s i s of elementary geometrical cons i d e r a t i o n s , eq. (2.16b) must b e changed t o read

& $ I i s t h e minimum r a d i u s of curvature of t h e middle surface i n t h e v i where , , > 2.51.11 we f i n d c i n i t y of t h e p o i n t pi(xl). Then, i n s t e a d of t h e i n e q u a l i t y r

(3.35)

and s i n c e

]@Imax

= M h , we o b t a i n
TQ

zhllz

1 Mh 1f 2 Rmin

----+ . . .

(3.36 )

e ) Appiication of focusing kernels, which i s t h e foundation f o r constructi n g eqs0(3.28a) (3.28b) and (3.34), l e a d s t o t h e conclusion t h a t t h e boundary conditions have only a weak influence on t h e s t r e s s - s t r a i n s t a t e of t h e s h e l l a t p o i n t s s u f f i c i e n t l y d i s t a n t from t h e contour of t h e middle surface.

O f course, t h e influence of t h e boundary conditions i s not confined t o t h e s t r i p of width 2.5 M h o r t o a narrower s t r i p i n accordance with t h e i n e q u a l i t y (3.36), s i n c e when t h e p o i n t M(xi) goes beyond t h e boundary of t h i s s t r i p , t h e region ( C ) can include part of t h e bordering s t r i p . For t h i s reason, t h e influence of t h e boundary conditions i s r e f l e c t e d on t h e s t r e s s - s t r a i n s t a t e of t h e e n t i r e s h e l l . However, t h e s t r u c t u r e of t h e above-derived equations permits t h e conclusion t h a t t h e influence of t h e boundary conditions i s weakened with i n c r e a s i n g d i s t a n c e of t h e p o i n t M ( x ) from t h e contour of t h e middle surface.

Section

4. Methods of Approximate Solution of a System of I n t e g r a l

Eauations of S h e l l Theorv

W e r e c a l l t h a t t h e approximate methods of s o l u t i o n of a system of i n t e g r a l equations i s most o f t e n based on approximate s u b s t i t u t i o n of t h e s e systems by

314

systems of a l g e b r a i c equations, Use of focusing new k e r n e l s permits r e s t r i c t i n g t h e number of unknowns e n t e r i n g i n t o t h e a l g e b r a i c equations approximately equivalent t o t h e i n t e g r a l equations. W e will give one of t h e p o s s i b l e methods of s e t t i n g up t h e s e equations.

L e t u s first consider t h e approximate r e p r e s e n t a t i o n of a c e r t a i n func- /312 tion ~ ( 5 1 , ) within t h e r e c t a n g l e MI (x + a, y + b ) , & (x a, y + b), M3 (x a, y b ) , h ( x + a, y b ) , assuming t h a t c e r t a i n of i t s values ~ ( x , y ), u)(x a, y), cp(x + a, y), co(x, y S), ( ~ ( xy , + b ) are known. Then, i n t e r p o l a t i n g t h e function cp(5, ? ) by a paraboloid, we o b t a i n

To shorten t h e formulas, l e t u s introduce t h e following notation:

The absolute values of t h e coordinate increments will be denoted by h j ( j =

=

1, 2).

L e t us r e t u r n t o t h e system of equations (3.214)~ The r e l a t i v e smallness of t h e i n t e g r a t i o n region (X) permits u s t o r e p r e s e n t t h e integrand functions assuming t h e region ( C ) by i n t e r p o l a t i o n polynomials of t h e form of eq.(4.1), t o be bounded by t h e r e c t a n g l e M,&F&& o r t o l i e i n s i d e it. Using t h e notae obtain t i o n (4,Z), w
I/(:)

(0, 0) -i- ujf" (0, 0)

JJ I(\:;;(0,ti)(is+
iL]

I .

3 15

(i, r = 1 , 2, 3; j = =1, 2; k ,

/I

- - = 0 ,1, 2,. . . , IV).

(4.3)

W e r e c a l l t h a t N ( 5 J ) i s an a r b i t r a r y p o i n t of t h e region ( E ) . The point/313 M ( x J ), as above, coincides w i t h t h e i n t e r s e c t i o n between t h e s i n g u l a r l i n e beari n g t h e a u x i l i a q y load and t h e middle surface of t h e s h e l l . Equations (L.3) c o n s t i t u t e a system of l i n e a r a l g e b r a i c equations. T h i s system cannot be solved autonomously, since t h e number of unknowns i s f i v e times as g r e a t a s t h e number of equations. However, by giving t h e coordinates XJ t h e increments I h j , we o b t a i n a s y s t e m of equations general f o r t h e e n t i r e s h e l l and containing t h e same number o f wanted equations and unknowns.

W e emphasize t h a t t h e r e s u l t a n t system i s more exact than t h a t obtained by t h e u s u a l methods, which a r e based on t h e s u b s t i t u t i o n of p a r t i a l d e r i v a t i v e s by f i n i t e - d i f f e r e r c e r a t i o s from t h e system of p a r t i a l d i f f e r e n t i a l equations of s h e l l theory, :-.:-:,-: i n t e g r a t i o n "smoothes out" t h e e r r o r s introduced by t h e use of t h e interpL; i i . ' , l ? i L formulas.
If w e replace t h e d e r i v a t i v e s by f i n i t e d i f f e r e n c e s , neglecting small q u a n t i t i e s of t h e f i r s t and higher orders, then t h e system of d i f f e r e n c e q u a t i o n s (4.3) i s converted i n t o a system of d i f f e r e n t i a l equations.
n n

f 2)

(i, r = = 1, 2, 3; j ,

S = 1,

2; k , p " 0 ,

1, 2 , . . . , .\"'.

(L.4.)

m u a t i o n s (4.3) and (4.4) approximately d e s c r i b e t h e s t r a i n e d s t a t e of t h e , bordering t h e contour of t h e middle s h e l l , o u t s i d e of t h e s t r i p of width r surface of t h e s h e l l . Constructing equations s u i t a b l e w i t h i ? t h i s s t r i p inW e s h a l l not consider t h e s e equations here. The methvolves no d i f f i c u l t i e s . od of e s t a b l i s h i n g them will be given l a t e r i n t h e t e x t .

It i s not hard t o convince ourselves t h a t t h e Let u s return t o e q s . ( 4 . 3 ) . f i r s t and second terms on t h e left-hand s i d e s of t h e s e equations a r e dominant,

'jlh

a s results from t h e following considerations: W u a t i o n s (3.34) can be represented in t h e form of

(i,

r = 1, 2, 3; j = 1, 2; k ,

p=@, 1, 2,. .. , 1).

(4.5)

A comparison o f eqs.(4.3) and (4.5) shows t h a t t h e f i r s t two terms on t h e i r left-hand s i d e s coincide. The t h i r d term on t h e left-hand s i d e of equat i o n (4.5), a f t e r a p p l i c a t i o n o f t h e i n t e r p o l a t i o n formulas, i s reduced t o /314 For t h i s reason, an t h e remaining terms on t h e left-hand s i d e of eq.(4.3). evaluation of t h e t h i r d term on t h e left-hand s i d e of eq.(L.S) i s equivalent t o an evaluation of t h e terms corresponding t o it i n eq.(4.3).

Assuming, according t o t h e e v a l u a t i o n s given i n t h e monograph (Bibl.lO), that

where L i s a c h a r a c t e r i s t i c dimension of t h e s h e l l , we f i n d f o r a t h i n s h e l l a t 2h : L = 0 . 0 1 ;M = 1 . 5 and L = &m:

where t h e p o i n t N(5: ) l i e s between t h e p o i n t M(xl ) and t h e boundary of t h e region (C). The e v a l u a t i o n ( b ) shows t h a t t h e G a u s s S e i d e l i t e r a t i o n process i s Retaining, f o r example, t h e first a p p l i c a b l e t o t h e system of equations (4.3). e o b t a i n t h e system of two terms on t h e left-hand s i d e of eqs.(4.3) o r (4.5), w equations of t h e f i r s t ( i n i t i a l ) approximation:

9
~

See, f o r instance, M.J.SaLvadori,

Numerical Methods i n Ehgineering, IL,lq55.

317

Equations (4.6) can e v i d e n t l y a l s o b e obtained from e q ~ ~ ( 3 . 3 4 by ) applicat i o n of p r o b a b i l i t y methods, f o r example of t h e Monte Carlo method The /315 p o s s i b i l i t y o f applying t h i s method i s based on t h e r e l a t i v e smallness of t h e I n t h i s case, t h e most c o r r e c t choice of t h e value of t h e required region (G). function $(*) (SJ ) under t h e i n t e g r a t i o n s i g n i s t h e choice of i t s value a t t h e p o i n t M(xJ ) a t t h e c e n t e r of t h e region (C); we again o b t a i n eqs.(4.6). The general a n a l y s i s o f t h e s o l v a b i l i t y o f t h e s e equations i s d i f f i c u l t , and we s h a l l n o t give it here, W e r e c a l l only that t h e Gauss method assumes t h e arbit r a r y choice of t h e i n i t i a l v a l u e s of t h e nondominating unknowns, permitting us, i n t h e case of complications, t o introduce i n t o t h e right-hand s i d e of eqs.(Ic.6) a d d i t i o n a l small terms, a t t r i b u t i n g a r b i t r a r y i n i t i a l values t o t h e r e j e c t e d e unknowns. The f u r t h e r cause of t h e G a u s s S e i d e l process i s well known, and w s h a l l n o t d i s c u s s it f u r t h e r .

*.

The s o l u t i o n s of t h e system of a l g e b r a i c equations (4.6) permit f i n d i n g t h e i n i t i a l approximate v a l u e s of t h e required f u n c t i o n s qk) ( x J ) . Their dependence on t h e boundary conditions i s r e f l e c t e d i n t h e composition of t h e f u n c t i o n s cPlk)(xJ), which contain a l l t h e assigned elements of t h e boundary conditions. The s o l u t i o n o f t h e system (4.6), c o n s i s t i n g o f 3(N + 1) equations, may be w r i t t e n out, using t h e well-known a l g e b r a i c formulas

T h i s n o t a t i o n permits us, in a more e a s i l y v i s u a l i z e d form than t h e t e n s o r equation (&,6), t o demonstrate t h e s t r u c t u r e of t h e s y s t e m of equations t o be solved. The determinant of t h e system of equations (4.6) i s of t h e form:

..............
-(E) ...........................

...

..............

...............
(2) ..............

...[J Klily,dS (8) ..............

(E) ..............

(4.8)

(E) ...........................
(8)

For t h e Monte Carlo method see, f o r instance, Modern Mathematics f o r Ehgineers, E.F.Bekkenbakh, Editor, S t a t e Publishing House f o r I n t e r n a t i o n a l L i t e r a t u r e , 1958, pp.275-287.

318

Each column of t h e determinant A corresponds t o an unknown function u$P)./316 This i s denoted by t h e n o t a t i o n above t h e h o r i z o n t a l l i n e a t t h e t o p of t h e determinant. To each value p, from 0 t o N t h e r e correspond t h r e e columns with values of r equal t o 1 , 2 , 3. The rows of t h e determinant correspond t o t h e i n d i c e s ( k ) and (i). To each value of k from 0 t o N t h e r e correspond t h r e e , 2 , 3. This e s t a b l i s h e s t h e r u l e f o r s e t t i n g rows with values of i equal t o 1 up t h e determinants

W e have

......................
q 1= ;

(Jy ...

II . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

......................

ay) ...

......... .........
.........

(4.9 1

O f course, a t high v a l u e s of N, even t h i s s i m p l i f i e d s o l u t i o n involves a degree of computing work which, i n l a b o r involved, does n o t correspond t o t h e requirements of r e l a t i v e accuracy adopted in t h e s h e l l theory. As will b e c l e a r from Chapter 1 1 1 , we have N s 3. I f N e q u a l s three, then t h e number of equations i n t h e s y s t e m (4.6) w i l l be twelve. I n t h i s case, we must n o t use t h e exact methods of c a l c u l a t i n g t h e determinants A$P) and A, b u t must have recourse eo approximate methods. Further, as a l r e a d y noted, we must f i n d t h e second approximation by s u b s t i t u t i n g t h e required f u n c t i o n s found from t h e f i r s t approximation i n t o t h e left-hand s i d e s of e q ~ ( 4 . 3 ) .

Equations (4.6) are a p p l i c a b l e i n t h e same region i n which eqs.(3.34) are If applicable, i.e., o u t s i d e t h e s t r i p determined by t h e i n e q u a l i t y (3.36). t h e point M(x1 ) i s i n s i d e t h e s t r i p bordering t h e contour of t h e middle surface of t h e s h e l l , then t h e right-hand s i d e of t h e equations will have t h e c u r v i l i n ear integrals
(1)

and
rii)

, which w i l l

extend over t h o s e parts of t h e contour

l y i n g i n s i d e a c i r c u l a r c y l i n d e r of r a d i u s ro with i t s a x i s coinciding with t h e s t r a i g h t l i n e b e a r i n g t h e a u x i l i a r y load. The t a s k of i n v e s t i g a t i n g t h e s t r e s s - s t r a i n s t a t e of t h e s h e l l i n s i d e t h i s s t r i p i s highly complex. Here we have a s u b s t a n t i a l l y three-dimensional d i s t r i b u t i o n of s t r a i n s and stresses, so t h a t a reduction of t h e three-dimensional problem of e l a s t i c i t y theory t o t h e two-dimensional problem of s h e l l theory /31Z w i l l undoubtedly d i s t o r t r e a l i t y , even i f we c o n s t r u c t an trexacttg s o l u t i o n of t h e boundary problem of t h e s h e l l theory. Evidently, t h e known s o l u t i o n s of t h e boundary problems o f t h e s h e l l theory, s a t i s f y i n g t h e c l a s s i c a l boundary condi-

319

t i o n s , describe only approximately t h e a c t u a l state of t h e s h e l l c l o s e t o i t s edges, although this approximation, as shown by observation and experiment, i s s u f f i c i e n t f o r p r a c t i c a l purposes. For this reason, t h e r e i s n o t much use in t r y i n g t o f i n d t h e exact s o l u t i o n s of t h e boundary problems of s h e l l theory if such s o l u t i o n s r e q u i r e a l a r g e expenditure of man-hours and money. W e should i n s t e a d attempt t o c o n s t r u c t an approximate solution, with an accuracy s a t i s f y i n g p r a c t i c a l means. W e w i l l not seek an exact s o l u t i o n in t h e border s t r i p , since t h e e n t i r e method considered here i s l a r g e l y approximate. W e w i l l confine ourselves t o t h e approximate solution, r e l y i n g on t h e f i n i t e - d i f f e r e n c e method.

L e t t h e a r c of t h e contour have no corner points. Let us consider t h e family of normals t o this arc. If t h e a r c of t h e contour belongs t o t h e segment (11), then t h e displacements a r e prescribed on i t and on t h e corresponding p a r t o f t h e contour Surface. A t a l l p o i n t s of t h e noma1 t o t h e a r c (11) of t h e contour on t h e boundary and beyond t h e border s t r i p o f width ro, t h e d i s placement vector components are approximately known. By using linear interpol a t i o n we will then be a b l e t o construct, i n first approximation, t h e displacement f i e l d within t h e zone and then t o f i n d t h e d e r i v a t i v e s of t h e displacement components from t h e v e c t o r s o f t h e l o c a l coordinate b a s i s , belonging t o t h e cont o u r surface; by Hookefs l a w we will f u r t h e r be a b l e t o determine t h e compone n t s of t h e s t r e s s v e c t o r on p a r t (IT) of t h e contour surface. Then, by applyextending over t h e region (F), w e ini n g equations o f t h e form of eq.(3.33), troduce c o r r e c t i o n s i n t o t h e displacement f i e l d determined i n f i r s t approximation.
Continuation of this process i s t h e o r e t i c a l l y p o s s i b l e ; i t s convergence i s e v i d e n t l y ensured by t h e smoothing influence of i n t e g r a t i o n . The case of a region (C) l y i n g i n s i d e t h e s t r i p bordering t h e contour of t h e middle surface, and enclosing p a r t of t h e contour (I), with t h e stress vect o r components prescribed on t h e contour surface, i s somewhat more complicated than t h e l a s t case. Here, t h e i n i t i a l approximation by t h e displacement v e c t o r components on t h e contour susface must be determined by e x t r a p o l a t i o n i n terms of t h e required values of t h e displacement vector components i n s i d e t h e zone and t h e values of t h e s e components o u t s i d e t h e zone; we must set up modified equations ( 3 0 3 3 ) t o determine t h e f i e l d of displacements i n s i d e t h e zone, s i n c e extrapolation introduces t h e required displacement v e c t o r components of an int e r n a l point of t h e zone i n t o t h e i n t e g r a l

s.
(1)

W e are unable t o d i s c u s s

/318

t h e s e questions in d e t a i l here, o r t o i n v e s t i g a t e t h e displacement f i e l d s in t h e neighborhood of a corner p o i n t o f t h e contour. Section 5. The I n t e g r o d i f f e r e n t i a l and I n t e g r a l Equations of t h e Dynamics of S h e l l s I n order t o o b t a i n t h e equations of motion from t h e equations of equilibrium, it i s s u f f i c i e n t t o include in t h e body f o r c e s t h e i n e r t i a f o r c e s

3 20

Q u a t i o n (3.24) then t a k e s t h e following form:

(i, r

.1

1, 3, 3; j .:= 1, 2 ; k -=O,

1, 2 ,

. . . , N).

W e r e c a l l t h a t an element o f volume reads

where dS i s an element of a r e a of t h e middle surface of t h e s h e l l .

I n t h e expressions of t h e i n e r t i a forces, l e t u s pass t o t h e covariant components, changing t h e arrangement of t h e i n d i c e s i n t h e corresponding s c a l a r Finally, l e t u s again make use of t h e approxproduct e n t e r i n g i n t o eq.(5.1). imate equation ( 3 0 9 )

P -0

Introduce now t h e notation

(5.3)

Then, as a result of transformations analogous t o those considered i n d e r i v i n g t h e system (3.33), we find

(i, r = 1, 2, 3; j - 1, 2; k , p = O ,

1, 2,. .. , N ) .

Equations (5.4) a r e t o be regarded as a s y s t e m of i n t e g r o d i f f e r e n t i a l /319 equations of t h e dynamics o f s h e l l s i n t h e unknown functions uIk)(t, XJ ) where t h e p o i n t M(xJ) belongs t o t h e middle surface of t h e s h e l l . The statements made in d i s c u s s i n g eqs.(3.33) eqs.

are l i k e w i s e a p p l i c a b l e t o

( 5 . 4 ) .

L e t u s now consider two conditions of v i b r a t o r y motion of t h e s h e l l which are encountered i n t h e s o l u t i o n of concrete problems.
1 . S t a t i o n a r y O s c i l l a t o r y Process

If t h e process i s s t a t i o n a r y , w e may p u t

( i = 1, 2 , 3; k - 0 ,
Similarly

1, . . . , A,).

(5.5

( r = l , 2, s ; p - o ,

1, 2) . . . , N ) .

will assume t h a t t h i s sequence i s d i s c r e t e .

The sequence of frequencies w may be e i t h e r f i n i t e o r i n f i n i t e , b u t we S u b s t i t u t i n g eqs.(5.5) and (5.6) i n t o eqs.(5.4), and equating t o zero t h e c o e f f i c i e n t s of cos (wt + e,), we f i n d

(S)

(i, r = 1 , 2, 3; j = 1, 2; k , p = o ,

1, 2,

. .. , IV).

where w runs through a d i s c r e t e s e r i e s of values according t o e q ~ ( 5 . 5 )

(5.6).

The system of equations (5.7) does not e s s e n t i a l l y d i f f e r from eqs.(3.33), and we s h a l l t h e r e f o r e n o t d i s c u s s it. The system (5.8) c o n t a i n s t h e parameter w . W e w i l l n o t f u r t h e r analyze t h e conditions of s o l v a b i l i t y of eqs.(5.8). Rather, we r e c a l l t h a t , s i n c e t h e l i n e a r system of i n t e g r o d i f f e r e n t i a l equat i o n s can be approximately replaced by a system of l i n e a r a l g e b r a i c equations, it follows t h a t ( i f , f o r some values of t h e parameter w, t h e r e e x i s t s a non- /320 t r i v i a l s o l u t i o n o f t h e homogeneous equations, i.e., equations with t h e f r e e terms iPJk) equal t o z e r o ) t h e system of inhomogeneous equations (5.8) has no sol u t i o n f o r t h e s e values of t h e parameter w. These cases which a r e cases of resonance w i l l r e q u i r e s e p a r a t e i n v e s t i g a t i o n .
2 . Nonstationary - .O s c i l l a t o r y Process
~~

I n t r a n s i e n t regimes of loading o f a s h e l l , it may be impossible t o repres e n t t h e f u n c t i o n s iPi(k) ( t , xJ ) by equations o f t h e form of e q . ( 5 . 5 ) . In t h i s case w e must t u r n t o t h e Laplace-Carson transformation o f t h e system of i n t e g r o d i f f e r e n t i a l equations (5.4).

It i s well known t h a t t h e r e p r e s e n t a t i o n of t h e f u n c t i o n s f ( t ) according t o Laplace and Carson i s expressed by t h e following formulas:

co

f ( y ) .=p j e - P f ( i ) d t .
0

Applying t h e Laplace-Carson transform t o t h e second time d e r i v a t i v e o f t h e function f ( t ) , we f i n d

ca

pSs~~f(/)dt=.pj(p)--p?f(O)-ppf(O).
0

Thus t h e r e p r e s e n t a t i o n of t h e second d e r i v a t i v e f ( t ) i s expressed i n terms of t h e r e p r e s e n t a t i o n o f t h e function f ( t ) and i t s value, and a l s o i t s f i r s t time d e r i v a t i v e a t t h e i n i t i a l time t = 0. Applying t h e Laplace-Carson transform t o t h e system of equations (5.4) and
~

Cf.,

f o r instance, A.I.LurTye,

Operational Calculus, Gostekhizdat, 1950.

3 23

u s i n g eqs. ( b ) and (c), we o b t a i n t h e following s y s t e m of o p e r a t i o n a l integrod i f f e r e n t i a l equations:

(5.9)
(i,
f =

1, 2, 3; j -

1, 2; k , q = o ,

1, 2,. . . , N ) .

The first t h r e e terms on t h e right-hand s i d e s of eqs.(5,9) are prescribed functions. The i n i t i a l conditions which must be s a t i s f i e d by t h e wanted funct i o n s e n t e r i n t o these terms.

/3n

After determining t h e r e p r e s e n t a t i o n s o f t h e q u a n t i t i e s ulk)from eqs. (5.9) we must f i n d t h e i r o r i g i n a l s , applying i n t h e general case t h e Riemann-Mellin formula:
e t i m

-+

I n s p e c i a l cases, Tables can be r e f e r r e d t o t h a t c o r r e l a t e t h e elementary and higher transcendental functions w i t h t h e i r transform#.

The described method w a s applied by G,Ye.Kazantseva i n solving problems of O n t h e Vibrations of Thin C i r c u l a r v i b r a t i o n of round p l a t e s (G.Ye.Kazantseva, P l a t e s , Thesis, Kiev Polytechnic I n s t i t u t e , 1956). Section 6. & a 1 Systems of I n t e g r o d i f f e r e n t i a l Q u a t i o n s o f t h e Dynamics o f S h e l l s w i t h Focusing Kernels and t h e i r . A z r r x i m a t e Solution

Heye, we will b r i e f l y d i s c u s s t h e approximate methods of s o l u t i o n o f t h e equations derived i n Sect.5, based on t h e focusing p r o p e r t i e s of systems of aGi1iax-y loads and t h e r e l a t e d systems of auxiliary displacements of stresses,
.3t

f o r instance, V.A,Ditkin and A,P.Prudnikov, Cf. Calculus of Operations. Fizmatgiz, 1961,

I n t e g r a l Transformations and

324

A s i n t h e l a s t Section, t h e equations of s t a t i o n a r y and nonstationary o s c i l l a t o r y processes will be considered separately.

1 . S t a t i o n a r y O s c i l l a t o r y Processes. The Frequency Spectrum

O n t h e b a s i s of t h e p r o p e r t i e s of equations with focusing kernels, t h e (5.8), o u t s i d e t h e s t r i p of width ro bordering t h e contour of system (5.7) t h e middle surface, can b e s i m p l i f i e d and represented i n t h e form:

The approximate s o l u t i o n of eqs,(6.1_3.) - (6.lb) can be performed a s in-/322 d i c a t e d i n o u r discussion on t h e s t a t i c equations, provided t h a t t h e frequency of t h e f r e e o s c i l l a t i o n s i s not w, i.e,,9 t h a t resonance i s absent. I n connection with this question, l e t u s consider t h e approximate d e t e r mination of t h e frequency spectrum by use o f t h e approximate s o l u t i o n (6.Yo). F i r s t , we note t h a t t h e approximate equation (6.1b) and t h e more exact equations (5.8) o f t h e s h e l l theory p e r n i t an approximate determination only o f a portion of t h e frequency spectrum o f the three-dimensional problem o f e l a s t i c i t y theory. The dimensions of t h e frequency regions a c c e s s i b l e for d e t e r minations based on t h e equations of s h e l l theory depend primarily on t h e nume approxon t h e number o f terms i n t h e polynomials of z by which w b e r N, i.e., imated t h e displacement v e c t o r components. Further r e s t r i c t i o n of t h e frequency region depends on t h e approximate methods used i n solving t h e i n t e g r o d i f f e r e n t i a l equations of t h e s h e l l theory. To o b t a i n an i d e a as t o t h e e f f e c t exerted on t h e frequency equation when replacing t h e approximate equation (6.lb) with focusing k e r n e l s by equations with k e r n e l s t h a t smoothly vary a t any v a r i a t i o n of t h e mutual p o s i t i o n of t h e p o i n t s M(xJ ) and N ( S J ) on t h e middle surface, l e t u s have recourse t o t h e class i c a l argument based on t h e approximate replacement of eqs.(6.lb) by a system of a l g e b r a i c equations

L e t u s imagine, on t h e middle surface of t h e s h e l l , an orthogonal coordin a t e n e t where t h e s i d e s of t h e meshes are s h o r t e r than r,, with ro equal t o M h (1 + -

1 % -+ according t o t h e condition ( 3 . 3 6 ) . Then, t h e region 2 r If n, i s t h e (C) will include a f i n i t e number of nodes of t h e coordinate net.

...)

3 25

number of nodes of t h e coordinate n e t i n t h e region (C), and q, i s t h e number of nodes on t h e e n t i r e middle surface (S), then we o b t a i n t h e approximate equation
/I,

:/I,

GS

(S) : (S).

W e w i l l consider t h e values of uJP) a t t h e nodes of t h e coordinate n e t as unknowns. Then, p l a c i n g t h e p o i n t M(xJ ) a t t h e nodes of t h e n e t and approximately representing t h e i n t e g r a l s i n eq.(6.lb) by f i n i t e sums under a p p l i c a t i o n of t h e formulas of mechanical quadrature, we r e p l a c e t h e system of equations (6.lb) by systems of a l g e b r a i c equations corresponding t o fixed p o s i t i o n s of This procedure i s applicable b o t h t o equations with focusing kert h e p o i n t M. n e l s and t o equations with k e r n e l s t h a t do n o t possess focusing properties. If we compare t h e determinant of t h e s y s t e m of a l g e b r a i c equations obtained from t h e i n t e g r o d i f f e r e n t i a l equations having k e r n e l s without focusing p r o p e r t i e s with t h e determinants of t h e system r e s u l t i n g from eqs. (6.lb),we /323 s h a l l see t h a t t h e l a t t e r has most of i t s elements equal t o zero, s i n c e most of t h e nodes of t h e net, a s can be seen from eqs.(6.2), a r e excluded from t h e region (z).
The determinant of t h e system w i t h focusing k e r n e l s includes elements t h a t depend on t h e boundaqy conditions. This f a c t i s due t o t h e influence of t h e boundary condition on t h e frequency spectrum. However, e v i d e n t l y t h e r e exists some p a r t of t h e frequency spectrum t h a t depends b u t weakly on t h e boundary conditions. W e will o f f e r suggestions t h a t seem t o confirm t h i s conclusion. It must be emphasized t h a t we consider t h e arguments developed below as being merely h e u r i s t i c

L e t u s perform, on t h e s y s t e m (6.lb), s i m p l i f i c a t i o n s analogous t o those e f i n d t h e system of a l g e b r a i c used i n studying t h e s t a t i c system ( 3 . 3 4 ) . W equations analogous t o t h e system (4.6):

(i, r = l , 2, 3; k , p = o ,

1, 2, . . . , N ) .

(6.3

The determinant of t h e system of a l g e b r a i c equations (6.3) can be obtained from eq.(b.8) on replacing t h e kernel K(l;)J 'by t h e d i f f e r e n c e

When t h e p o i n t M(xJ ) i s superposed on t h e fixed nodes of t h e coordinate n e t , and when t h e i n t e g r a l s e n t e r i n g i n t o t h e determinant A of t h e system of a l g e b r a i c equations (6.3) a r e replaced by f i n i t e sums r e s u l t i n g from t h e formul a s of mechanical quadrature, t h e determinant A w i l l have numerical elements i n s t e a d of f u n c t i o n a l ones. E q u a t h g t h e determinant A t o zero, we f i n d t h e

3 26

c h a r a c t e r i s t i c values of t h e parameter w, t i o n s of t h e s h e l l .

t h e frequencies ,of t h e free vibra-

The determinant A, a s shown i n Sect.l, i s of t h e o r d e r 3 ( N + 1). I f t h e number of nodes of t h e coordinate n e t on t h e middle surface S i s n,, then t h e t o t a l number of unknowns w i l l b e 3 ( N + l)b ; t h i s number i s equal t o t h e degree The r o o t s of t h e of t h e complete equation of frequencies with r e s p e c t t o w2. complete equation of frequencies,i.e., of t h e equation including t h e c o e f f i c i e n t s of t h e unknowns a t a l l nodes of t h e coordinate net, must include r o o t s f o r which t h e determinant A, derived f o r t h e s y s t e m o f equation (6.3), approximatel y vanishes. I n t h e opposite case, t h e homogeneous system of equations corresponding t o eq.(6.3) would b e only t r i v i a l s o l u t i o n s f o r t h e s e roots, which i n turn would mean t h a t t h e s o l u t i o n s of t h e system (6.3) had d e v i a t i o n s a r b i t r a r i l y g r e a t i n a b s o l u t e value from t h e s o l u t i o n s of t h e o r i g i n a l system cons t r u c t e d without t h e above j u s t i f i e d s i m p l i f i c a t i o n s , a case which would cont r a d i c t a l l of o u r e a r l i e r conclusions. But then, t o o b t a i n approximately a portion of t h e frequency spectrum, /324

it i s enough t o equate t h e determinant d of t h e system of a l g e b r a i c equation (6.3) t o zero, placing t h e p o i n t M ( x J ) a t one of t h e nodes of t h e coordinate
net. Thus, w e f i n d t h e approximate values of t h e 3 ( N + 1) r o o t s of t h e frewuency equation. Placing t h e g o i n t M(xJ ) a t a l l t h e nodes of t h e net, we f i n d T h i s shows t h a t t h e use of focusing k e r n e l s a l l t h e 3 ( N + l)% values of w enables u s approximately t o r e p r e s e n t t h e frequency equation a s a product of n , factors.

This again l e a d s t o t h e concept of t h e existence of two groups of frequencies. The frequencies of one group a r e weakly connected with t h e boundary conThese are frequencies approximately determined by t h e equations t h a t ditions. are obtained by equating t h e determinants o f t h e system (6.3) t o zero. The second group o f frequencies depends s u b s t a n t i a l l y on t h e boundary conditions. These frequencies a r e obtained from t h e determinants of systems analogous t o system (6.3) b u t set up under t h e assumption t h a t t h e point M(xJ ) belongs t o t h e s t r i p of width r , bordering t h e contour of t h e middle surface. This subd i v i s i o n of frequencies i n t o two groups should become more d i s t i n c t f o r t h i n n e r s h e l l s . Such phenomena are obviously connected w i t h t h e dynamic boundary eff e c t s mentioned i n A.LOve*s book+. A n i n d i r e c t confirmation of t h e correctness of t h e s e conclusions i s t h e weak dependence of t h e c r i t i c a l values o f t h e load on t h e boundary conditions, which i s mentioned i n works on t h e theory of stab i l i t y of s h e l l s , i n t h e presence of l a r g e regions on t h e middle surface s u f f i 0 ) . The l a t t e r conclusions, however, c i e n t l y remote from i t s contour (Bibl.4, 1 r e q u i r e a d d i t i o n a l research.

2. Nonstationaw Processes
I n t h e case of equations with focusing kernels, eqs.(5.9) can be simplif i e d i f t h e p o i n t M(xJ ) l i e s o u t s i d e t h e s t r i p bordering t h e contour of t h e e find middle surface, as mentioned above. W
-3

A.Love,

Mathematical Theory of m a s t i c i t y , ONTI, 1935, p.575.

3 27

P)

(i,

: = 1,

2, 3; j - 1, 2; It, q = o , 1, 2,

... , N ) .

The s i m p l i f i c a t i o n s made i n s e t t i n g up t h e equations of t h e s t a t i c s of s h e l l s y i e l d a system of a l g e b r a i c equations analogous t o system (4.3) and approximately replacing t h e system (6.L):

/325

The meaning of t h e n o t a t i o n introduced here i s given i n constructing t h e system of equations (4.3). Again basing our c a l c u l a t i o n on t h e scheme o f t h e

Gauss-Seidel i t e r a t i o n process, we o b t a i n t h e following system of equations of t h e f i r s t ( i n i t i a l ) approximation:

(E)

(i, r = l , 2, 3; j = 1, 2; k , q - 0 ,

1, 2,. . . , A).

/326
Solving t h i s system according t o eqs.(6.7) on replacing t h e k e r n e l s K f i \ ; we f i n d t h e approximate expressions of t h e reby t h e i r sums Kc($hi + p V tck iq) quired r e p r e s e n t a t i o n s L + ( ~ ) i n t h e form of i n t e g r a l r a t i o n a l functions o f t h e parameter p, and then continue t h e process of i t e r a t i o n . The elementary consid e r a t i o n s based on eqs.(4.7) and an evaluation of t h e degree of p i n t h e i r numerators and denominators show t h a t , from t h e r e s u l t a n t transforms, t h e o r i g i n a l s can be found, i.e., t h e s e transforms permit inversion.

In t h e s t r i p bordering t h e contour of t h e middle surface, we must u s e t h e These methods again reduce t h e problem t o approximate methods given i n Sect.4. t h e s o l u t i o n of a system of a l g e b r a i c equations containing, a s unknowns, t h e transforms of t h e functions sought. W e w i l l not f u r t h e r develop t h i s method b u t r e t u r n t o t h e discrete-continuous method considered i n Chapter IV.
Section 7. Application of t h e Discrete-Continuum - Method ~
-

I n Chapter I V we considered an approximate method f o r studying t h e dynami c s of s h e l l s , based on replacing t h e s h e l l by a discrete-continuous system. The meaning of t h e concept of a discrete-continuous system has been given i n Sect.13, Chapter IV.
W e r e c a l l t h a t , t o s e t up t h e equations of motion and t h e equations of conn e c t i v i t y t h a t had t o be s a t i s f i e d by t h e generalized coordinates of t h e d i s crete-continuous system, w e used a method of reducing t h e three-dimensional problems of t h e theory of e l a s t i c i t y t o t h e two-di.mensj.ona1 problems of t h e theory of s h e l l s , based on expansion of t h e wanted. .tc,r!;:tions i n Maclaurin tens o r s e r i e s i n escending powers of t h e coordinate zs

This method has t h e shortcoming noted i n Chapter 111, The complications encountered i n t h e a n a l y t i c composition of t h e c o n n e c t i v i t y equations r e s u l t i n g e r e c a l l t h a t t h e s e equafrom t h e boundary conditions are very s u b s t a n t i a l , W tions, i n t h e general case, contained t h e second time d e r i v a t i v e s ( a s w e l l as t h e d e r i v a t i v e s of o d e r s higher than t h e second) of t h e generalized coordinates; this precludes t h e use of t h e apparatus of c l a s s i c a l dynamics i n con-

329

I
I

II

I 1111 II1111

II

s t r u c t i n g t h e equations o f motion, although i n s p e c i a l cases t h i s a p p a r a t m i s a p p l icab1e I n t h e preceding S e c t i o n s o f this Chapter we considered a new method of reduction, based on t h e theorem of work and r e c i p r o c i t y and t h e use of focusing a u x i l i a r y loads. This reduction method permits e l i m i n a t i o n of t h e above comp l i c a t i o n s and t o map o u t a more e f f e c t i v e v e r s i o n of .the discrete-continuous method.
w i t h focusing kernels,

W e s h a l l s t a r t o u t from t h e system of i n t e g r o d i f f e r e n t i a l equations (5.4) s t i l l r e s t r i c t i n g t h e i n t e g r a t i o n region t o t h e region (C). Then, t h e system of i n t e g r o d i f f e r e n t i a l equations (5.4), o u t s i d e t h e /327 s t r i p o f width r , bordering t h e contour of t h e middle surface, can be given t h e following form:

(7.1)
( i , r = l , 2, 3; k , p = O ,

I , 2 , - . . ,N).

Let u s make f u r t h e r use o f t h e i n t e r p o l a t i o n formula (4.1). t h e notation:

W e introduce

Then, eqs.(7,1) tem (4.3):

w i l l y i e l d t h e system of equations analogous t o t h e sys-

('Zp r = I, 2, 3;j = l , 2; k , p=O, 1, 2,

... , N).
/328

(7.3a) t h e

O n t h e b a s i s of t h e estimates given in Sect.4, we find from the system sgstem of equations of t h e first ( i n i t i a l ) approximation.

The System o f equations ( 7 . 3 ) i s ana.Iogous t o t h e s y s t e m of algebraic equations (4.6) of t h e s t a t i c s ,of shells and contains d e r i v a t t v e a o n l y with respect t o t h e t i m e t. L e t us replace tire triangulation- net introduced in Sect.14 of Chapter I V by an orthogonal coordinate net corresponding t o the i n L e r p o b ~ n formula (4.1). W e shall superpose t h e point M(xf ) with t h e nodes o f t h e n e t and consid e r t h e q u a n t i t i e s u $ P ) a t t h e nodes as generalized coordinates, W e assume t h a t T h iilr t h e region t h e meshes of t h e net considerably exceed t h e region ( E ) t h e r e can be cmly one node of the net. Since t h e coordinaLes of the nodes are fixed, t h e integrals over t h e region (C) e n t e r i n g lnto e q s . ( 7 . 3 b ) w i l l be n mfnd eqs0(5d), we b i functions of t h e " h e r of nodes of t h e net. Bearing i troduce t h e notation:

e).

y!;;'

( i t )=

JJ v:z{
(3)

(x;,

E") dS;

m;;y ( / I ) =JJ V(\$)r

(E)

(27, E9) (ti

a;

331

where n is t h e number of t h e node of t h e n e t and t h e t h e node n The system of equations (7.3a) t a k e s t h e form

are t h e coordinates of

/329

[i, r = 1 ,2 , 3; j = 1 , 2; k, p = 0, 1 ,2 , formed according t o t h e conventional r u l e ) .

...,

N; t h e summation over j i s per-

The s y s t e m of equations of t h e i n i t i a l approximation i s of t h e form:

(i,

r = l , 2, 3; k, p = o , 1, 2,...., N).

Equations (7.5a) and (7.5b) are s e t up f o r each node of t h e n e t t h a t does n o t belong t o t h e s t r i p of width ro bordering t h e contour of t h e middle surface. The n a t u r a l question arises as t o t h e degree of approximation t o t h e descript i o n of real motion of s h e l l elements by t h e equations of t h e i n i t i a l approxiThis question i s complex but i s of p r a c t i c a l importance, s i n c e mation (7.9). a c l e a r d e f i n i t i o n of t h e r o l e of eqs. (7.9) will decrease t h e amount of comp-

332

c1
t a t i o n a l work necessary t o solve t h e problem.

W e s h a l l confine o u r s e l v e s t o t h e following remark: Although e ~ s . ( 7 . % ) , are s e t up autonomously f o r each node o f t h e net, which apparently i s inconsisf+ e n t with t h e interdependence between t h e motions of t h e nodes, this dependence is i m p l i c i t l y r e f l e c t e d in t h e composition of t h e c o e f f i c i e n t s of t h e s e equat i o n s and of t h e i r right-hand sides, which helped t o reduce t h e error. Actually, t h e c o e f f i c i e n t s of eqsm(7.5a) (7.5b) are expressed by i n t e g r a l s extending over a m a l l b u t f i n i t e region (C) of t h e middle s u r f a c e of t h e s h e l l , Including, as they do, t h e e l a s t i c constants, t h e d e n s i t y of t h e s h e l l , i t s thiclmess, and t h e p r i n c i p a l curvatures of t h e middle surface, t h e s e c o e f f i c i e n t s . r e f l e c t t h e b a s i c p r o p e r t i e s of t h e s h e l l as a continuous medium. They are links that connect t h e state of t h e s h e l l a t a c e r t a i n point with i t s stat4 i n t h e region surrounding t h a t point, This connectivity i s t h e result of app l i c a t i o n of t h e Reciprocal Theorem, one of t h e fundamental theorems of t h e mechanics of e l a s t i c bodies,

I n determining t h e first approximation f r o m eqsm(7,5b), we f i n d t h e following approximations from eqs. (7.5a), by s u b s t i t u t i n g i n them t h e second-degree A t t h e s e stages, t h e i n t e r r e l a t i o n beterms of t h e s o l u t i o n of eqs.(7,5b). tween t h e motions of a d j a c e n t nodes of t h e n e t i s e x p l i c i t l y revealed.
I n conclusion we n o t e that, on a p p l i c a t i o n of equations t h a t have ker- /330 n e l s without focusing p r o p e r t i e s , t h e i n t e g r a l s i n t h e i n t e g r o d i f f e r e n t i a l equations (5.4) would extend over t h e e n t i r e middle surface, and a l l t h e generalized coordinates would e n t e r into them?+.
L e t u s d i s c u s s t h e equations of motion of elements of a s h e l l l y i n g i n a s t r i p of width r , where ro i s determined by e q . ( 3 . 3 6 ) , bordering t h e contour o f t h e middle s u r f a c e of a s h e l l . The g r e a t e s t d i f f e r e n c e between t h e s e q u a t i o n s and t h e equations considered i n Chapter I V (Sects.15 16) l i e s i n t h e f a c t t h a t components of t h e displacement and stress vectors, prescribed over t h e contour surface, a r e included in t h e prescribed f u n c t i o n s Q : , which are here analogs of t h e generalized forces. This r e c a l l s t h e "automatic'' i n c l u s i o n o f t h e equations of geometric c o n n e c t i v i t y i n t h e system of Lagrange equations o f t h e second kind on s e l e c t i o n of t h e generalized coordinates corresponding t o t h e conditions of a concrete problem o f mechanics without i n t r o d u c t i o n of redundant coordinates. Evidently, it i s n o t a question here of a simple e x t e r n a l s i m i l a r i t y b u t of a complex i n t e r n a l r e l a t i o n , s i n c e t h e Reciprocal Theorem i s connected with t h e p r i n c i p l e of p o s s i b l e displacement*.

The above p r o p e r t i e s f o r equations with focusing k e r n e l s and t h e reMtant s i m p l i f i c a t i o n s of t h e system of equations of motion of s h e l l s were given by us i n t h e Note "Approximate Methods o f I n v e s t i g a t i n g t h e Equilibrium and Vibrao . 2 t i o n s of S h e l l s as Discrete-Continuous Systems", in: Information B u l l e t i n N of t h e S c i e n t i f i c Council on " S c i e n t i f i c P r i n c i p l e s of S t r e n g t h and P l a s t i c -

itsw, 1961,

+w T h i s i n t e r r e l a t i o n was c a l l e d t o o u r a t t e n t i o n by'Yu.N.Shevchenko, S c i e n t i s t , h s t i t u t e of Mechanics, Academy of Sciences UkrSSR.

Senior

333

Thus, t h e complications mentioned in Sect.15 of Chapter IV do n o t a c t u a l l y It i s t r u e , t h e s e complications w e r e due t o t h e method of reduction arise. used in t h a t Section and d i d n o t r e s u l t from t h e essence of t h e mechanical 'problem. W e r e c a l l t h a t t h e appearance o f time d e r i v a t i v e s i n t h e equations of c o n n e c t i v i t y w a s due, in this case, t o t h e successive elimination of derivat i v e s of t h e form 0 , .V,q on t h e b a s i s of t h e equations of motion in t h e Lam6 form.

..

To set up t h e d i f f e r e n t i a l equations of m t i o n of t h e nodes of t h e coordin a t e n e t in t h e s t r i p bordering t h e contour o f t h e middle surface, it i s necessary, as pointed o u t in Sect.4, t o use u n i l a t e r a l f i n i t e d i f f e r e n c e s f o r exp r e s s i n g t h e unknown v a l u e s of t h e components of displacement and s t r e s s on t h e contour s u r f a c e in terms of t h e i r v a l u e s i n s i d e t h e s h e l l . The forms of t h e s e equations depend s u b s t a n t i a l l y on t h e form of t h e contour and t h e geometrical p r o p e r t i e s of t h e middle surface. Here, t h e s e equations are n o t considered. The discrete-continuous method always allows u s t o o b t a i n a s y s t e m of ord i n a r y d i f f e r e n t i a l equations, determinate in t h e sense t h a t t h e number o f equations equals t h e number of f u n c t i o n s sought. This confirms t h e existence and uniqueness of t h e s o l u t i o n of t h e system of d i f f e r e n t i a l equations of mot i o n of t h e s h e l l , constructed by us. W e r e c a l l that t h e appraximate replacement o f t h e i n t e g r a l s by f i n i t e

/331

sums, has been, ever s i n c e Mer's day, t h e c l a s s i c a l method of i n v e s t i g a t i n g

v a r i o u s q u e s t i o n s of mathematical physic+. The requirements of m a t h e m t i c a l r i g o r compel u s t o pass t o t h e limit, from t h e wanted systems of d i f f e r e n t i a l equations with a f i n i t e number of f u n c t i o n s t o systems of i n t e g r o d i f f e r e n t i a l equations. This passage t o t h e limit must show t h a t t h e s o l u t i o n s o f t h e systems of d i f f e r e n t i a l equations obtained from t h e i n t e g r o d i f f e r e n t i a l equations (5.4) by i n t e r p o l a t i o n formulas, L e . , by t h e discrete-continuous method, coinc i d e a t t h e l i m i t with t h e s o l u t i o n s of t h e initial system (5.4).

O f course, t h e s o l u t i o n s io t h e s i m p l i f i e d systems of t h e form of equat i o n s (7.5a) (7.5b) a r e o n l y r a t h e r rough approximations t o t h e exact solut i o n s , and cannot be used t o c o n s t r u c t a sequence of f h c t i o n s converging t o t h e exact s o l u t i o n of t h e s y s t e m (5.4).

W e have n o t i n v e s t i g a t e d t h e above-mentioned process of convergence in det a i l s i n c e t h e technique of such i n v e s t i g a t i o n has been thoroughly studied so t h a t i t s r e s u l t i s known i n advance, although one would have t o overcome cons i d e r a b l e d i f f i c u l t i e s when i n v e s t i g a t i n g a method of mathematically d e s c r i b i n g t h e process adapted t o t h e s p e c i a l features of t h e problem. Such i n v e s t i g a t i o n s would b e o u t s i d e t h e scope o f our study.
Section 8. Nonlinear I n t e g r o d i f f e r e n t i a l Equations of t h e Dynamics of S h e l l s I n Sect.12 of Chapter 1 1 , we proved a theorem which w e c a l l e d t h e theorem T h i s theorem is of work and r e c i p r o c i t y in t h e nonlinear theory of e l a s t i c i t y .

* See,
Vol,l

- 2,

f o r instance, R.Courant and DAilbert, Methods of Mathematical Physics, Gostekhizdat, 1933 19%

334

a p p l i c a b l e t o a n a n i s o t r o p i c medium with geometrical and physical nonlinearity. The Reciprocal Theorem makes it p o s s i b l e t o construct nonlinear integrod i f f e r e n t i a l equations of t h e dynamics of s h e l l s .

W e s h a l l confine o u r s e l v e s t o consideration of a s h e l l made of an i s o t r o It p i c material, and s h a l l consider only t h e case of geometrical nonlinearity. i s n a t u r a l t o assume t h a t f i n i t e deformations are developed i n t h e s h e l l under t h e a c t i o n of t h e applied load and under c e r t a i n conditions on i t s surface contour, a s a r e s u l t of i t s f l e x i b i l i t y (Bibl.4).. For this reason f i n i t e deformat i o n s c h a r a c t e r i z e t h e b a s i c state of t h e s h e l l b u t not i t s auxiliary state.

As before, t h e a u x i l i a r y s t a t e w i l l be constructed from t h e displacementsB32 and s t r e s s e s i n a l i n e a r l y deformed unbounded medium under t h e a c t i o n of t h e I n this case, t h e s h e l l , as already pointed out, load i n d i c a t e d i n Sect.2. must be regarded a s part of an unbounded medium.
This argument e l i m i n a t e s t h e fundamental d i f f i c u l t y connected with t h e above procedure f o r a p p l i c a t i o n of t h e Reciprocal Theorem, s i n c e it i s then no longer necessary t o determine t h e p a r t i c u l a r s o l u t i o n s of t h e equations of t h e theory of e l a s t i c i t y , corresponding t o t h e a c t i o n o f a concentrated f o r c e i n a nonlinear deformable e l a s t i c medium.

L e t u s again consider (11, 12.10) a s w e l l as (11, 12.2a), (11, 12.2b), (11, 12.8) and (11, 1 2 . 9 ) . W e s h a l l attempt t o reduce t h e s e r e l a t i o n s t o exp r e s s i o n s containing t h e b a s i c system of required f u n c t i o n s ur(p) introduced i n t h i s Chapter, To o b t a i n results s u i t a b l e f o r concrete c a l c u l a t i o n s , we s h a l l introduce, i n t o t h e n o n l i n e a r p a r t of t h e components of t h e f i n i t e - d e f o r m a t i o n , , o n l y t h o s e terms which have been considered dominant ever since t h e tensor D f i r s t s t u d i e s made by T.Karman, These a r e t h e terms depending on t h e displace0 ) . L e t u s use t h e notation: ment v e c t o r components u, (Bibl.4, 1

W e make use here of t h e metric of t h e unstrained s h e l l , i n which, i n accordance w i t h Sect.2 of Chapter 1 1 , t h e covariant d e r i v a t i v e s 0, and v, a r e dee note t h a t , i n c o n t r a s t t o t h e modern theory of f l e f i b l e p l a t e s termined. W and s h e l l s , we s h a l l n o t put ( V 3 u , ) = 0 .
The covariant d e r i v a t i v e s V , u , ( i = 1 , 2 ) do n o t e n t e r i n t o t h e system o f functions g*). Therefore, t o accomplish t h e program of s e t t i n g up equations i n t h e unknown f u n c t i o n s we s h a l l u s e t h e approximate r e l a t i o n r e s u l t i n g from t h e e q u a l i t i e s (111, 7.4a), supplementing it by t h e nonlinear f a c t o r

( i s 1, 2).

335

where X(+li and X ( - l i are t h e s t r e s s vector components determined with respect to the forces a c t i n g on t h e deformed boundarg surfaces. This results from t h e
conditions ( 1 1 , 8.2b), From eqs.(8.2), which a.lso have a meaning f o r f i n i t e deformations. t h e following approximate r e l a t i o n i s obtained:

W e introduced t h e notation Y , = &+)

i

L222
- Xi-) I

( i = 1, 2 ) .

2P

Under t h e above assumptions, the r e l a t i o n ( 1 1 ,1 2 . 2 b ) can be represented

as follows:

W e took account here of t h e orthogonality of t h e system of coordinates t h e d e f o r m e d middle surface.

on

Making use of e q s . ( I I , 12.2a) and (8.5a) (8.R), we find, f o r t h e b a s i c system of forces a c t i n g on t h e s h e l l and t h e r e s u l t a n t strains and s t r e s s e s :

(8.6a)

336

a*

= ui3 - pup [ Yi- r c j q (1

@)-t;

a * ,

=Q33 - 1 (u3 (1))2

( j = 1, 2).

(8.6b)

Further,

it follows from ( 1 1 ,1 2 . 8 ) that

The meaning of t h e second t e r m o n t h e right-hand side of eQ.(a) will be c l e a r The first term i s connected with t h e s t r e s s vector on t h e from t h e foregoing. To write eq.(a) i n the surface of t h e s h e l l , a s will be seen f r o m ( 1 1 , 8.13). expanded form, one must separately consider t h e s t r e s s vector on the boundary n t h e boundary surfaces, surfaces of t h e s h e l l and on i t s contour surface. O eqs.(II, 8.3) can be represented in t h e form

%=!E
and, on t h e contour surface,

(i=

1, 2); x 3 = + h ;

Making use of ( 1 1 , 8.yb)j we f i n d on t h e boundarg surfaces:

B33

= 1.

A l l t h e remaining q u a n t i t i e s B,, contour surface

on t h e boundary surfaces

( r = 1, 2, 3).

Making use of (11, 6.5) and of t h e simplifying assumptions by T ~ K a m n ,we obt a i n on t h e boundary surfaces of t h e s h e l l :

and, applying (11, 8.13),

we f i n d on t h e boundarg surfaces of t h e s h e l l :

337

111111.

..,..

. ..-. - .

a. .tli
iJ

--

t 1xktc)i +X[,) j(l +l l i l )

-11;)

+ . ..)

(i= 1, 2).

(8.7)

Noting t h a t on t h e boundary s u r f a c e s of t h e s h e l l , f o r z = &h,

q . ( a ) , we find from e

(i = 1, 2);

This W e note t h a t t h e q u a n t i t i e s a r e small by comparison with permits a s b p l i f i c a t i o n i n deriving, from t h e equations of motion, t h e equaE s s e n t i a l l y eqs.($.8a) (8.8b) show t h a t , under t i o n s of first approximation. Kamanfs simplifying assumptions, t h e s t r e s s v e c t o r components on t h e boundary If, according t o Karmanfs s u r f a c e s can be determined b t h e l i n e a r theory. evaluations, t h e o r d e r of ui5 i s equal to t h e o r d e r o f (i, k = 1, 2), and t h e o r d e r of 4) i s higher than t h e o r d e r of e l k , then a l l introduced nonlinear

dl .

mk

terms w i l l be of t h e o r d e r

peik and w i k .

3/2

If terms with t h e f a c t o r s epq (p, q = 1 , 2 ) a r e r e t a i n e d i n eqs.(8.8a) t o (8,8b), then t h e right-;and s i d e s of eqs.(8.8a) (8.8b) w i l l have a d d i t i o n a l m teImS of t h e o r d e r peik O f course, i f considerable accuracy i s desired, one

- .

must r e t a i n i n eqs0(8.8a)

(8.8b) first of a l l t h e terms of o r d e r

. 3/2

O n t h e contour surface,

Here, we dropped t h e f a c t o r (1+ t . + $ )

7.

Making use of eqs.(8.2),

we f i n d

338

Here again t h e term YPuD() i s of t h e order

elk lected. The ~$1 up() i s of t h e order construction of t h e second approximation.

- .
3/a

e l k , and this term should be negSuch terms can be retained in t h 6

O n t h e contour surface,
ni#O

( i = 1, 2);

n,3=0.

L e t u s assume, t o simplify t h e calculations, t h a t there are no loads on the contours of t h e boundary surfaces. In t h a t case, t h e components Yp vanish. . l 3 ) , and r e l a t i o n ( a ) of this Section, we obtain on Again making use of (11, 8 t h e contour surface

( 8 . 1 0 )
(i= 1, 2, 3).
( 8 0 % ) . It will be The components T,, are determined from eqs0(8.5a) c l e a r from eqs.(8,5a) t h a t t h e right-hand side of eq.(8.10) contains t h e nonThus, i n contrast $ ) (i, k = 1 , 2 ) , of t h e order of VBlk l i n e a r terms u t o the stress vector components on t h e boundary surfaces, t h e components of t h e vector on t h e contour surface a r e s u b s t a n t i a l l y nonlinear, even i f we use t h e simplified K a m n theory.

- .

The calculations performed here show t h a t t h e only sources of substantia l l y nonlinear terms, which a r e of the r e l a t i v e o r d e r - p q k , i n t h e s t r e s s tensor components, a r e represented by t h e components of t h e tensor Tlk ( i , k = This could have been foreseen. Indeed, the s u b s t a n t i a l l y nonlinear = 1, 2). terms depend on t h e components of the antisymmetric tensor Q s r a t h e r than on t h e tensor of s m a l l deformations e r S . This i s confirmed, f o r instance, by (11,
9.2).

Hereafter, t o simplify the calculations we s h a l l omit terms of t h e r e l a - a t i v e order

and of higher orders.

Let u s now consider t h e body forces @ , d etermined from (11, 12.9). First, l e t u s study a l l terms containing t h e q u a n t i t i e s P t g ? . I n accordance with t h e must be replaced by t h e componr e l a t i v e accuracy adopted, t h e q u a n t i t i e s Pi; O n the basis e n t s of the tensor Nil;? of rank two, determined by (11, 6 . 1 2 b ) . of t h e above study we conclude t h a t t h e sum

does not contain terms of t h e r e l a t i v e o r d e r -

weik

and should be omitted. Con-

sequently,

339

(i, k = 1, 2, 3).

( 8 . 1 1 )

After completing t h e preliminary a n a l y s i s , we make use o f ( 1 1 ,1 2 . 1 0 ) , apressing t h e generalized Reciprocal Theorem. The above expression ( 1 1 ,1 2 . 1 0 ) , under t h e previous assumptions on t h e a w d l i a r g system of f o r c e s , displacements, q . ( r . l ) , with s e v e r a l and s t r e s s e s , again l e a d s t o an equation of t h e form of e
a d d i t i o n a l terms, i n c l u d i n g t h e i n t e g r a l

( 8 . 1 2 )

Making use of t h e Ostrogradskiy-Gauss formula, l e t us transform t h e integ r a l I as follows:

From eqs.(a) and ( k ) of t h i s Section it w i l l b e seen t h a t t h e i n t e g r a l s J J

i s ,

over t h e surface of t h e s h e l l , containing t h e components of t h e t e n s o r T i k , a r e cancelled.

W e neglect t h e n o n l i n e a r terms i n t h e composition o f t h e surface f o r c e s /337 depending on according t o eqs.(8.8a) (8.8b) and (8.10), s i n c e t h e s e terms

are of an order of smallness higher than peik

3/ 2

Consequently, a t t h e degree of r e l a t i v e accuracy adopted by us, t h e only source of nonlinear terms i n t h e i n t e g r c d i f f e r e n t i a l equations of t h e theory of s h e l l s will be t h e i n t e g r a l

340

Making use of we o b t a i n

(8,5a),

and r e t a i n i n g i n it t h e terms of t h e order

welk

where

The i n t e g r a l I, e n t e r s i n t o t h e right-hand s i d e of eq.(5.4) t i v e sign, y i e l d i n g

with a posi-

(8.16)
( S I

(i, t = l , 2, 3; j, q = 1, 2; k , p = o , 1, 2,

... , N).

Equations (8.16 ) form t h e required system of nonlinear i n t e g r o d i f f e r e n t i a l equations of t h e theory of s h e l l s . This system occupies a place intermediate between t h e system of equations (5.4) of t h e l i n e a r theory and t h e general system of equations of **strong flexure''. However, so f a r as we h o w , t h e l a t t e r 1 3 3 8 system has never been s e t up o r a t l e a s t has not yet been s e r i o u s l y studied, Let u s assume, a s above, t h a t t h e Let u s r e t u r n t o t h e system (8.16). a u x i l i a r y system of l o a d s l e a d s t o t h e construction of focusing k e r n e l s of int e g r o d i f f e r e n t i a l equations. Then, o u t s i d e a s t r i p of width M h
1 Itill

3 u

bordering t h e contour of t h e middle surface, t h e system (8.16) t a k e s t h e f o l lowing form:

1%)

( i , r = l , 2, 3 ; j , q==.1, 2; k , p = O , 1, 2, * . * , N).

Applying i n t e r p o l a t i o n formulas of t h e type of eq.(b.l), we replace t h e 8 . 1 7 ) by a system of nonlinear d i f f e r system of i n t e g r o d i f f e r e n t i a l equations ( e n t i a l equations.

W e s h a l l only give t h e system of equations of t h e i n i t i a l approximation, since t h e construction o f a system analogous t o eq.(7.5a) involves no fundamental difficulties, W e find

(l, r= I ,

2, 3; j , q = 1, 2; k , p = o , 1, 2,

... , N).

(8.18)

If t h e s h e l l i s i n equilibrium, then t h e system of d i f f e r e n t i a l equations (8.18) is transformed i n t o a system of nonlinear a l g e b r a i c equations. The sol u t i o n of t h e system (8.18), and of t h e more general and analogous system (7.5a), is, i n t h e g e n e r a l case, a non-single-valued function. I n o r d e r t o sel e c t t h e required branch of t h e s o l u t i o n , one must i n v e s t i g a t e t h e gradual development of deformations of t h e s h e l l from i t s i n i t i a l undeformed s t a t e . The branch p o i n t s o r l i m i t p o i n t s will correspond t o t h e c r i t i c a l values of t h e load.
One of t h e p r a c t i c a l methods o f solving t h e s y s t e m of equations (8.17) i s t o replace it by a system of ordinarg d i f f e r e n t i a l equations, using t h e d i s - & Crete-continuous method. Making use of a coordinate n e t on t h e middle surface, superposing t h e p o i n t M(xJ ) on t h e nodes of t h e net, and considering t h e quantities $ I a t t h e nodes as generalized coordinates, we s h a l l o b t a i n a system of W e (7.B). ordinary d i f f e r e n t i a l equations analogous t o t h e systems (7.5a)

342

r e c a l l t h a t t h e s e systems could b e constructed f o r s u f f i c i e n t l y strong focusing p r o p e r t i e s of t h e k e r n e l s of t h e i n t e g r o d i f f e r e n t i a l equations and f o r dimens i o n s of t h e n e t such that, on superposition of t h e p o i n t M(xJ)on one of t h e nodes i n t h e region (C), t h e r e would b e no neighboring nodes. I n t h i s case, as noted i n Sect.?, t h e integrals over t h e region (C) e n t e r i n g i n t o e q ~ ~ ( 8 . 1 8 ) w i l l be f u n c t i o n s of t h e nmibers of t h e nodes of t h e net. L e t u s make use of t h e n o t a t i o n (7.4a) ( 7 . 4 ~ ) and introduce t h e a d d i t i o n a l n o t a t i o n

(8.19)

Here, a s in Sect.?,

n i s t h e number of a node of t h e net.

The system of equations (8.18) now t a k e s t h e following form:

(i, I-=. 1, 2, 3; j , s = 1, 2; k , p = 0 , 1, 2,

... , N ; n =

1, 2,

...

(8.20)

The above statements on consideration of t h e system of equations ( 7 . B ) can be applied, with c e r t a i n a d d i t i o n s , t o t h e system of equations (8.20). Since, under t h e simplifying assumptions adopted by us, t h e nonlinear terms disappear from t h e boundary conditions, t h e equations of motion i n t h e zone bordering t h e contour of t h e middle s u r f a c e w i l l n o t d i f f e r s u b s t a n t i a l l y from eqs.(8.20). The method of approximate c o n s t r u c t i o n of t h e equations in t h e bordering zone i n d i c a t e d i n Sect.7 remains v a l i d also in t h i s case. F i n a l l y we note t h a t t h e a p p l i c a t i o n o f t h e method of equivalent l i n e a r i zation, analyzed i n Chapter IV, makes it p o s s i b l e t o r e p l a c e t h e s y s t e m (8.20) by t h e l i n e a r system of equations (?.%), provided t h e e l a s t i c constants a r e properly s u b s t i t u t e d . I n t h e concluding Section of t h i s Chapter we w i l l draw g e n e r a l i z i n g conc l u s i o n s on t h e proposed methods and give a general c h a r a c t e r i z a t i o n o f t h e i r s i g n i f i c a n c e f o r t h e mechanics of shells. Section 9. O n t h e Construction of Kernels o f _ - r n t e g r o d i f f e r e n t i a l Equations with Focusing Prope%ies The preceding p o r t i o n of t h i s Chapter was based on t h e p o s s i b i l i t y of cons t r u c t i n g a s y s t e m of a u x i l i a r y displacements, l e a d i n g t o k e r n e l s of i n t e g r o d i f f e r e n t i a l equations with p r o p e r t i e s which, i n accordance with C.LanczoSn-, w e
x- See t h e passage c i t e d i n Section 2 of t h e book by C.Lanczos.

343

It w i l l be c l e a r from t h e content' of Sects.2 and 3 t h a t an w i l l c a l l focusing. , i d e a l focus is e v i d e n t l y impossible, j u s t as it i s impossible i n o p t i c a l instruments. However, as we have shown i n Sects.2 and 3,.a focusing property s u f f i c i e n t t o guarantee t h e required accuracy o f t h e results can indeed be obtained.
However, t h e u s e of t h e theory o f generalized f u n c t i o n s opens new possib i l i t i e s f o r t h e construction o f k e r n e l s with i d e a l focusing p r o p e r t i e s , i.e., p r o p e r t i e s corresponding t o t h e p e r t u r b a t i o n s of an e l a s t i c medium embracing a s t r i c t l y bounded region of space and vanishing beyond t h o s e bounds. These app l i c a t i o n s of t h e t h e o r y of generalized f u n c t i o n s would go beyond t h e scope of t h e present study.

In Sects.2 and 3, t h e focusing a c t i o n of t h e l o a d was obtained by conand % (CY) a ccording t o eqs. (2.27) (2.29). These s t r u c t i n g t h e f u n c t i o n s % (CY) functions determined t h e d i s t r i b u t i o n of t h e s i n g u l a r i t i e s of t h e f i e l d of ami l i a r y displacements and s t r e s s e s on t h e segments of t h e s i n g u l a r l i n e going beyond t h e boundaries of t h e region bounded by t h e s u r f a c e of t h e s h e l l . As already noted, t h e exact s o l u t i o n of t h e problem of t h e required d i s t r i b u t i o n of s i n g u l a r i t i e s can go beyond t h e limits of c l a s s i c a l a n a l y s i s . If we abandon t h e requirement of "exact focusing", then t h e methods o f constructing t h e focusing load given i n Sects.2 and 3 can be simplified.

Let u s give up t h e determination of t h e f u n c t i o n s %(CY) and &(CY). The f i e l d of displacements and s t r e s s e s caused in t h e s h e l l by a c t i o n of t h e l o a d s & (a) and 99 (CY) will l i k e w i s e possess weak focusing p r o p e r t i e s . To strengthen t h e s e properties, l e t us m u l t i p l y t h e displacements caused by t h e l o a d s %(CY) and %(CY) by a f u n c t i o n o f t h e p o i n t N , which i s damped with s u f f i c i e n t rapidi t y with i n c r e a s i n g d i s t a n c e of t h e point N from t h e s i n g u l a r l i n e , and which, on t h e s i n g u l a r l i n e , becomes unity.

An example of a f a c t o r strengthening t h e focusing a c t i o n of t h e load i s t h e function

where

k >> 0, while f ( r ) i s a monotonically increasing, everywhere d i f f e r e n t i a b l e , and continuous function s a t i s f y i n g t h e condition

f ( 0 )=0

(9.3)

as well as t h e condition t h a t t h e d e r i v a t i v e s

as, ?(=vi

s h a l l be bounded.

W e r e c a l l t h a t y, and ployed i n Sect.2.

7, are t h e rectangular Cartesian coordinates em- / 3 U

The f a c t o r of t h e form (9.1) introduced i n t o t h e right-hand s i d e s of q u a t i o n s (2.la) (2.lb) permits constructing t h e f i e l d of displacements satisfyi n g t h e inhomogeneous equations of equilibflum of t h e e l a s t i c body caused by t h e f i e l d s of body f o r c e s i n t h e unbounded e l a s t i c medium, approaching zero as r i n c r e a s e s without l i m i t . The p o i n t M(T), ), i f t h e condition (9.3) i s satis-

a r e bounded, will, as before, b e t h e df Yi'Q dY, p o i n t of a p p l i c a t i o n of t h e u n i t concentrated force. The f u r t h e r constructions of t h e i n t e g r o d i f f e r e n t i a l equations do n o t i n p r i n c i p l e d i f f e r from those considered above. f i e d and i f t h e d e r i v a t i v e s

A p o s i t i v e c o e f f i c i e n t k can always be chosen such that, on t h e boundary of t h e assigned region ( E ) and o u t s i d e t h a t region, t h e p o s i t i v e v a l u e s of t h e components of t h e a u x i l i a r y displacements, s t r e s s e s and body forces, s h a l l n o t exceed prescribed small q u a n t i t i e s . This will ensure t h e focusing p r o p e r t i e s of t h e k e r n e l s of t h e i n t e g r o d i f f e r e n t i a l equations.
A l l above elementary conclusions r e q u i r e no d e t a i l e d proof, s i n c e they result from t h e well-known a n a l y t i c p r o p e r t i e s of t h e p a r t i c u l a r s o l u t i o n s of t h e equations of s t a t i c s of an e l a s t i c medium, corresponding t o t h e a c t i o n of concentraPed forces. They do, however, confirm t h e e x i s t e n c e of i n t e g r o d i f f e r e n t i a l equations of t h e theory of s h e l l s with focusing kernels, which a r e of fundamental s i g n i f i c a n c e in t h e s t a t i c s and dynamics of s h e l l s . Section 10. I n t e g r o d i f f e r e n t i a l Equations Defining Contiguous S o l u t i o n s of t h e BokProblems _ _ of t h e S t a t i c s and Dvnamics of S h e i l s
~

Here we will b r i e f l y c h a r a c t e r i z e another method of s e t t i n g up and solving t h e i n t e g r o d i f f e r e n t i a l and i n t e g r a l equations of t h e s h e l l theory. T h i s method was f i r s t given by u s i n 1946 (Bib1.23e) and w a s f u r t h e r developed l a t e r Several r e p o r t s have been published i n t h e meantime,in (Bib1.23f ), (Bib1.23i). which t h i s method i s applied t o various problems of t h e s t a t i c s of s h a l l + . The problems considered by t h e s e methods a r e s p e c i a l case3 of t h e general problem which can be defined as follows: The s o l u t i o n of a boundary problem of t h e s t a t i c s o r dynamics of s h e l l s i s known. Required, t o c o n s t r u c t t h e solut i o n of a d i f f e r e n t problem c l o s e t o t h e f i r s t one by some c r i t e r i o n . This problem w a s posed by u s previously (Bib1.23e). S o l u t i o n s c l o s e by some d e f i n i t e c r i t e r i o n we term contiguous o r adjacent. I n t h e above works, t h i s problem w a s solved mainly i n one version. The s o l u t i o n of a boundary /3k2 problem of t h e equilibrium of a p l a t e , onto whose middle plahe was mapped t h e middle s u r f a c e o f a s h e l l , was assumed t o be known. Then, a system of integrod i f f e r e n t i a l equations or, i n p a r t i c u l a r , i n t e g r a l equations, w a s constructed

*W e

might mention t h e works (Bibl.17,

31a,b;

345

which yielded t h e s o l u t i o n of an adjacent boundary problem f o r t h e s h e l l . A n exception w a s another study ( B i b L g l b ) , where t h e solution. of t h e boundary problem of s t a t i c s of a c y l i n d r i c a l s h e l l was used f o r solving t h e boundary problem of s t a t i c s of a c y l i n d r i c a l s h e l l with modified boundary conditions. The p r i n a r g means f o r c o n s t r u c t i n g t h e i n t e g r o d i f f e r e n t i a l and i n t e g r a l equations of an a d j a c e n t problem,was t h e a p p l i c a t i o n of t h e Reciprocal Theorem according t o our previous s t u d i e s (Bib1.23e) which we l a t e r developed i n g r e a t d e t a i l (Bib1.23h). The essence of t h e s e c o n s i d e r a t i o n s reduces t o t h e following: It i s u s u a l l y a s s e r t e d t h a t t h e Reciprocal Theorem connects two systems of displacements and t h e corresponding f o r c e s o r s t r e s s e s i n some e l a s t i c body The i n t e r p r e t a t i o n of this theorem can be expanded by consi(11, Sect.12). d e r i n g two systems of displacements i n two d i f f e r e n t e l a s t i c bodies with m t u a l l y connected a r i t h m e t i z a t i o n o f t h e i r i n t e r i o r p o i n t s by systems of coordiThen t h e system of displacen a t e s having t h e same r e l a t i v e dimensionality. ments of t h e second body may be a t t r i b u t e d t o t h e first body, by determining t h e e x t e r n a l f o r c e s corresponding t o t h e s e displacements from t h e equations of t h e theory of e l a s t i c i t y . This treatment of t h e theorem of work and r e c i p r o c i t y has proved f r u i t f u l , y i e l d i n g new approaches t o t h e s o l u t i o n of problems, n o t only of t h e mechanics of s h e l l s b u t a l s o of t h e s t a t i c s and dynamics of one-dimensional and threeIt a l s o bedimensional problems of t h e theory o f e l a s t i c i t y and p l a s t i c i t p . came p o s s i b l e t o separate, from t h e displacements o f p o i n t s of t h e middle surface, t h e " p l a t e terms" from t h e terms t h a t depend on t h e curvature of t h e midd l e surface. The s e p a r a t i o n of t h e "plate terms" has helped t o solve a number of problems, mainly on t h e equilibrium o f c y l i n d r i c a l s h e l l s , although t h e method has been developed f o r s h e l l s of a r b i t r a r y form (Bib1.23g-i). Limited space prevents u s from examining t h i s method in more d e t a i l ; we have, t h e r e f o r e , found it more expedient t o focus t h e reader's a t t e n t i o n on t h e new methods of a p p l i c a t i o n o f t h e Reciprocal Theorem, l e a d i n g t o t h e construct i o n of equations w i t h focusing kernels. Several statements should be made i n conclusion: _ .

1 . The described method can be f u r t h e r improved by introducing focusing f a c t o r s of t h e form of eq.(q.l) i n t o t h e system of a u x i l i a r y displacements. O n approximate replacement of t h e system of i n t e g r o d i f f e r e n t i a l equations of equil i b r i u m of a s h e l l by a system of a l g e b r a i c equations, this will permit u s t o f 3 0 decrease t h e number of unknowns in each equation.
2. The method of c o n s t r u c t i n g equations t h a t determine t h e adjacent solut i o n s permits c o n s t r u c t i n g a chain o r c o n t i n u i t y of s o l u t i o n s i n which each sol u t i o n r e s u l t s from t h e preceding solution. This, of course, r e q u i r e s a l a r g e amount of work.

3. The system of i n t e g r a l equations o f equilibrium of t h e theory o f s h e l l s ,

Cf., f o r example (Bib1.23c), as well a s t h e paper by N.A.Kil'chevskiy and L.V.Lirsa, and t h e paper by R.A.Mikhaylenko i n Izv. Kiev. Politekhn. Inst., Vo1.31, 1961.

346

obtained by transformation of t h e i n t e g r o d i f f e r e n t i a l equations, can be a system of F'redholm i n t e g r a l equations of t h e second kind. This system, however, may have no unique solution, i f t h e conditions of Fredholmts t h i r d theorem are satisfied. A n example of this case i s given elsewhere (Bib1.23j).

1 . Concluding Remarks on t h e I n t e g r o d i f f e r e n t i a l and I n t e g r a l Section 1 Equations of t h e S t a t i c s and Dynamics of S h e l l s

W e a r e here giving general conclusions on t h e r o l e played by t h e results of t h e i n v e s t i g a t i o n s given i n Chapter V f o r t h e general theory of s h e l l s . Here we must d i s t i n g u i s h between t h e purely t h e o r e t i c a l value and t h e applied value of t h e methods considered. Let u s f i r s t c h a r a c t e r i z e t h e t h e o r e t i c a l value of t h e method of i n t e g r o d i f f e r e n t i a l equations r e s u l t i n g from t h e Reciprocal Theorem.
W e have shown t h a t t h e a p p l i c a t i o n of t h e Reciprocal Theorem supplements t h e s o l u t i o n methods f o r t h e problem of reducing t h e three-dimensional problems of t h e theory of e l a s t i c i t y t o t h e two-dimensional problems of t h e theory of s h e l l s by a s u b s t a n t i a l l y new method.
One of t h e advantages of t h i s method, i n our opinion, i s t h a t i t s applicat i o n does not r e q u i r e t h a t t h e components of t h e body and surface f o r c e s app l i e d t o t h e s h e l l be d i f f e r e n t i a b l e , i n c o n t r a s t t o t h e methods i n d i c a t e d a t t h e beginning of Chapter 1 1 1 , which r e q u i r e s a t i s f a c t i o n of t h e conditions t h a t t h e v e c t o r components of t h e prescribed f o r c e s be d i f f e r e n t i a b l e . In (111, Sect.19) we note t h a t t h e s o l u t i o n of t h e equations r e s u l t i n g from t h e general equation of dynamics l i k e w i s e requires t h a t t h e v e c t o r components of t h e prescribed f o r c e s b e d i f f e r e n t i a b l e o r t h a t t h e theory of generaltced f u n c t i o n s be applied. These d i f f i c u l t i e s a r e eliminated when t h e methods given in this Chapter a r e applied. The smoothing influence of i n t e g r a t i o n permits us, witho u t a n a l y t i c complications, t o consider a s h e l l loaded by concentrated forces. I n t h i s connection, we must emphasize t h e fundamental d i f f e r e n c e between t h e methods considered i n Sects.1 9 of this Chapter and t h e methods mentioned i n Sect.10. This d i f f e r e n c e c o n s i s t s i n t h e f a c t t h a t t h e method i n Sect.10 i s n o t an independent method of reduction, b u t i s based on a p p l i c a t i o n of t h e results of a preliminary reduction.

Several works have r e c e n t l y been published on new methods for s e t t i n g up/344 t h e i n t e g r a l equations of t h e s t a t i c s and dynamics of s h e l l s . These works give methods t h a t permit replacement of t h e system of d i f f e r e n t i a l equations of equil i b r i u m o r motion of an element of a s h e l l , compatibly with t h e boundary cond i t i o n s , by equivalent s y s t e m s of i n t e g r o d i f f e r e n t i a l o r i n t e g r a l equations-!!-.

It i s c l e a r These methods are r e l a t e d t o t h e method given i n Sect.10. t h a t t h e results of t h e replacement of t h e system of d i f f e r e n t i a l equations by a n equivalent system of i n t e g r a l equations w i l l n o t e l i m i n a t e t h e e r r o r i n t r o duced i n t o t h e system of d i f f e r e n t i a l equations by t h e a p p l i c a t i o n of v a r i o u s
+$

C f . A.A.Berezovskiy, I n t e g r o d i f f e r e n t i a l Equations of t h e -Nonlinear Theory of F l a t Thin S h e l l s , Ukr.Matemat. Zhurnal, Vol.XI1, 1960; and (Bibl.19).

347

11111 I

simplifying assumptions, f o r example, t h e Kirchhoff-Love assumptions, V.Z.Vlas o v t s simplifying assumptions of t h e t e c h n i c a l theory of s h e l l s , and others. For this reason, t h e cognitive value of t h e s e methods i s lower than t h a t of t h e general method considered in Sects.1 9, which autonomously s o l v e s t h e reduct i o n problem.

The method considered i n Sects.1 9 permits o f f u r t h e r g e n e r a l i z a t i o n s , based on t h e d i s t r i b u t i o n of s i n g u l a r i t i e s of t h e displacements of t h e a u x i l i a r y problem, not on a l i n e b u t i n p a r t of t h e volume of t h e s h e l l , I n t h i s case, we can c o n s t r u c t v a r i o u s generalized averages of t h e components of t h e displacement v e c t o r and t h e components of t h e s t r a i n and s t r e s s tensors, t h u s e l i m i n a t i n g t h e complications connected with t h e preliminary approximate repres e n t a t i o n s of t h e displacement v e c t o r components by eqs. (3.9). W e w i l l d i s c u s s now t h e p r a c t i c a l value of t h e above methods. The method based on t h e a p p l i c a t i o n of i n t e g r o d i f f e r e n t i a l and i n t e g r a l equations with focusing kernels, a s i s obvious from t h e content of Sects.1 9, i s an e f f e c t i v e means f o r a numerical s o l u t i o n of t h e boundary problems of t h e s h e l l theory, I n e x a c t l y t h e same way, t h e methods given in Sect.10 can be used a s t h e b a s i s f o r a numerical s o l u t i o n of t h e s e boundary problems. It seems t o u s t h a t t h e methods considered a r e more adaptable t o t h e a p p l i c a t i o n of numerical computational methods than systems of d i f f e r e n t i a l equations, s i n c e t h e replacement of t h e i n t e g r a l by a f i n i t e sum introduces a smaller e r r o r than t h e replacement of der i v a t i v e s by a r a t i o of f i n i t e d i f f e r e n c e s . The inconvenience connected with t h e n e c e s s i t y of extending t h e sums, approximately replacing i n t e g r a l s , over a l l nodes of t h e coordinate n e t of t h e middle surface, which l e a d s t o equations with a l a r g e number of unknowns, i s eliminated by t h e i n t r o d u c t i o n of focusing kernels. Of course, a l l numerical methods r e q u i r e t h e use of computers,

The p r i n c i p a l shortcoming of t h e study i n Chapter V i s t h e l a c k of calcue have confined /345 l a t e d a n a l y t i c expressions and Tables of focusing kernels. W o u r s e l v e s t o a d e s c r i p t i o n of t h e t h e o r e t i c a l p r i n c i p l e s of t h e method, s i n c e construction of t h e a n a l y t i c expressions f o r t h e k e r n e l s and t h e corresponding Tables i s very t e d i o u s and would t a k e a r e l a t i v e l y long time, These d a t a w i l l be published i n t h e next p a r t of t h i s study,

It i s c l e a r t h a t a r e a l i s t i c study of focusing k e r n e l s might compel t h e i n t r o d u c t i o n of p a r t i c u l a r c o r r e c t i o n s t o t h e above-given methods of t h e i r cons t r u c t i o n , b u t t h e r e can be no doubt t h a t this would not a f f e c t t h e fundamental p r i n c i p l e s of t h e method i t s e l f .

A shortcoming inherent i n t h e e n t i r e study i s t h e f a c t t h a t we neglected t h e d i s s i p a t i v e f o r c e s of various o r i g i n generated during t h e motion of t h e elements of t h e s h e l l . W e d e l i b e r a t e l y adopted t h i s extensive s i m p l i f i c a t i o n , s i n c e our o b j e c t w a s t o s e t f o r t h t h e p r i n c i p l e s of t h e a n a l y t i c theory of s h e l l s , and e i t h e r t h e i n t r o d u c t i o n o r t h e n e g l e c t i n g of d i s s i p a t i v e f o r c e s does not go t o t h e foundation of t h e theory.
W e r e c a l l t h a t , i n many papers on t h e dynamics of s h e l l s , t h e e f f e c t of t h e d i s s i p a t i v e f o r c e s has likewise n o t been subjected t o i n v e s t i g a t i o n . It was noted i n t h e s e cases t h a t t h e i n f l u e n c e of t h e d i s s i p a t i v e forces, a s shown

by experiment, causes a r a p i d damping of f r e e o s c i l l a t i o n s (Bibl.12). This inconsistency, which sometimes appears i n t h e dynamics of s h e l l s , i s elimina t e d in a number o f more recent s t u d i e s , i n c l u d i n g t h e B o l o t i n monograph (Bib1 2c )

The method o f i n t e g r a l and i n t e g r o d i f f e r e n t i a l equations i n t h e t h e o r y of s h e l l s has n o t received general recognition. As A.I.Lurtye so v i v i d l y p u t s it, i t s llcompetitivenessn s t i l l r e q u i r e s confinnation.

W e assume t h a t t h e development of t h e theory of equations with focusing k e r n e l s and examples of t h e i r a p p l i c a t i o n t o s p e c i a l problems will y i e l d convincing confirmation'of t h e power of this new method. It i s piXrticularly imp o r t a n t t o c h a r a c t e r i z e t h e f i e l d of s p e c i a l problems i n which t h i s method has obvious advantages o v e r t h e methods of c l a s s i c a l s h e l l theory. This f i e l d e r e c a l l t h a t t h e methods should b e t h a t of p a r t i c u l a r l y complex problems. W of a n a l y t i c mechanics have long been applied t o problems of p a r t i c u l a r d i f f i culty, whose s o l u t i o n could n o t be d i r e c t l y obtained from t h e laws of Newton and t h e general theorems of dynamics. D.Leach w r i t e s t h a t t h e a p p l i c a t i o n of a n a l y t i c mechanics t o t h e s o l u t i o n of simple problems i s as i n e f f i c i e n t as i s t h e use of a n a i r p l a n e t o c r o s s a street-%. This statement i s fully a p p l i c a b l e t o t h e a n a l y t i c mechanics of s h e l l s .
The f i e l d of problems f o r which it i s obviously expedient t o apply t h e new methods w i l l g r a d u a l l y become outlined'n".t.

* D.Leach,

C l a s s i c a l Mechanics. IL, Moscow, 1961,

+e+The first s t e p i n this d i r e c t i o n has been taken.

Cf.,interalia,(Bibl.l7,31b)

349

BIBLIOGRAPHY

I. Monographs and Handbooks

Theory of Anisotropic S h e l l s (Teoriya anizotropnykh obolochek) Fizmatgiz, 1961. 2. Bolotin, V.V.: a ) The Dynamic S t a b i l i t y of E l a s t i c Systems (Dinamicheskaya ustoychivostf uprugikh sistem). Gostekhizdat, 1956; b ) S t a t i s t i c a l Methods i n S t r u c t u r a l Mechanics ( S t a t i s t i c h e s k i y e metody v s t r o i t e l f n o y mekhanike) Gosstroyizdat, 1961; c ) Nonconservative Problems of t h e Theory of E l a s t i c S t a b i l i t y (NekonFizmatgiz, 1961. servativnyye zadachi t e o r i i uprugoy ustoychivosti) 3. Vlasov, V.Z.: a) The General Theory of S h e l l s (Obshchaya t e o r i y a obolochek). Gostekhizdat, 1949; b ) Thin-Walled S p a t i a l Systems ( 'Tonkostenn.yye prostranstvennyye sisteqy). Gosstroyizdat, 1958; c ) Beams, Plates and S h e l l s on an m a s t i c Base ( I n Collaboration with N.N.Leontfyev) (Balki, p l i t y i obolochki na uprugom ochnovanii; sovmestno s N.N.Leontfyevym). Fizmatgiz, 1960. 4. Volfmir, A.S.: F l e x i b l e Plates and S h e l l s (Gikbiye p l a s t i n k i i obolochki). Gostekhizdat , 1956. 5. Golfdenveyzer, A.L. : The Theory of Thin E l a s t i c S h e l l s (Teoriya uprugikh tonkikh obolochek). Gostekhizdat, 1953. P r i n c i p l e s of t h e Theory of Surfaces (Osnovy t e o r i i poverkh60 Kagan, V.F.: nostey). Gostekhizdat, 1947. 7. Kilfchevskiy, N.A.: Elements of Tensor Calculus and i t s Application t o Mechanics (Elementy tenzornogo i s c h i s l e n i y a i ego prilozheniya k mekhanike) Gostekhizdat, 1954. 8. Kochin, N.Ye.: Vector Calculus and t h e F i r s t P r i n c i p l e s of Tensor Calculus Izd. AN (Vektornoye i s c h i s l e n i y e i nachala tenzornogo ischislerfiya)

1 . Ambartsumyan, S.A.:

SSSR, 1951.

9. Lurfye, A.I.:

a ) S t a t i c s of Thin-Walled E l a s t i c S h e l l s ( S t a t i k a tonkostennykh uprugikh obolochek) Gostekhizdat , 1947; b ) S p a t i a l Problems of t h e Theory of E l a s t i c i t y ( P r o s t r a n s t v e n q y e Gostekhizdat, 1955. zadachi t e o r i i uprugosti) 1 0 . Mushtari, Kh.M. and Galimov, K.Z.: The Nonlinear Theory of E l a s t i c S h e l l s Tatknigoizdat , 1957. (Nelineynaya t e o r i y a uprugikh obolochek) 1 1 . Novozhilov, V.V. : Sudpromgiz, 1951; a ) Theory of Thin S h e l l s - (Teoriya tonkikh obolochek) b) Elements o f t h e Nonlinear Theory of E l a s t i c i t y (Osnovy nelineynoy t e o r i i uprugosti). Gostekhizdat, 1948. Some Dynamic Problems of t h e Theory of S h e l l s (Nekotoryp 12. Oniashvili, O.D.: dinamicheskiye zadachi t e o r i i obolochek). Izd. AN SSSR, 1957. Riemannian Geometry and Tensor Analysis (Rimanova geo13. Rashevskiy, P.K.: metriya i tenzornyy analiz). Gostekhizdat, 1953.

350

1 1 . Papers and Books on Special Problems of S h e l l Theory

a . Alumyae,

N . A . : .On t h e Fundamental System of I n t e g r a l s of t h e Equation of Small, Axisyrmnetric, Steady Vibrations of an E l a s t i c , Conical S h e l l of Revolution (0 fundamentaltnoy sisteme i n t e g r a l o v uravneniya malykh o s e s h e t r i c h e s k i k h ustanovivshikhsya kolebaniy uprugoy konicheskoy Izv. AN EstSSR, S e r i y a tekhn. i fiz.-mat. nauk, obolochki vrashcheniya). No.1, 1960. 1 5 0 Akseltrad, E.L.: a) Some Problems of t h e Theory of S h e l l s Inhomogeneous i n Thermoelastic P r o p e r t i e s (Nekotoryye zadachi t e o r i i neodnorodnykh PO termouprugb svoystvam obolochek) Thesis, Leningrad Polytechnic I n s t i t u t e , 1959; b) Equations of Deformation of S h e l l s of Revolution and of Flexure /347 of Thin-Walled Rods under Great E l a s t i c Displacements (Uravneniya def o r m a t s i i obolochek vrashcheniya i i z g i b a tonkostennykh sterzhney p r i Izv. AN SSSR, Otd. tekn. nauk, boltshikh uprugikh peremeshcheniyakh). mekhan. i mashinost., No.4, 1960. 16. Ambartswqyan, S.A. : a) O n Two Methods of Calculating Two Massive Orthotropic S h e l l s ( 0 dvukh Izv. AN metodakh r a s c h e t a dvukh slitrlykh ortotropnykh obolochek) ArmSSR, N0.2, 1957; b) Contribution t o t h e Nonlinear Theory of Plane Orthotropic S h e l l s ; i n Collaboration with D.V.Peshtmaldzhyan ( K nelineynoy t e o r i i pologikh Izv. AN ortotropnykh obolochek; sovmestno s D.V.Peshtmaldzhyanom) ArmSSR, No.1, 1958; c ) Contribution t o t h e Theory of Bending of Anisotropic P l a t e s ( K t e o r i i Izv. A n SSSR, No. 5, 1958; i z g i b a anizotropnykh p l a s t i n o k ) d) O n t h e Dynamic S t a b i l i t y of Nonlinearly E l a s t i c Three-Layered P l a t e s ; i n Collaboration with V.Ts.Gnuni (0 dinamicheskoy ustoychivosti nelineyno-uprugikh trekhsloynykh plastinok; sovmestno s V.Ts.Gnuni). P r i k l . m a t . i mekh. (PMM), Vo1.25, No.4, 1961. The Calculation of S h e l l s (Raschet 17. Vaynberg, D.V. and Sinyavskiy, A.L.: obolochek). Gosstroyizdat, UkrSSR, 1961. O n a Method of Calculating Prismatic S h e l l s (Ob odnom metode 18. Vekua, I.N.: Trudy Tbilis. m a t . Inst., Vol. 21, r a s c h e t a prizmaticheskikh obolochek)

19. Gavelya, S.P. and Kuzemko, A.M.:

On t h e E l a s t i c Equilibrium of a Rigidly Clamped F l a t S h e l l of Constant Curvature w i t h Arbitrary Contour (Ob uprugom ravnovesii zhestko zashchemlennoy pologoy obolochki postoyannoy krivizny s proizvoltnym konturom). Minutes of Conference on t h e Theory of Plates and S h e l l s , KazanT, 1961. 20. Golt denveyeer , A. L .: a) On Reissnerts Theory of t h e Bending of P l a t e s (0 t e o r i i i z g i b a p l a s t i nok Reyssnera). Izv. AN SSSR, Otd. tekhn. nauk, No.4, 1958; b) The Asymptotic P r o p e r t i e s of Eigenvalues i n Problems of t h e Theory of Thin E l a s t i c S h e l l s (Asimptoticheskiye svoystva sobstvennykh znacheniy v zadachakh teorii uprugikh tonkikh obolochek). PMM, Vo1.25, No.&, 1961, The S t a b i l i t y of Bimetallic C y l i n d r i c a l S h e l l s (Ustoy21. Grigolyuk, E.I.: Inzh. sborn., c h i v o s t t b b e t a l l i c h e s k i k h t s i l i n d r i c h e s k i k h obolochek) Vo1.23, 1951. On t h e Basic Relations o f t h e Theory of Thin S h e l l s 22. Darevskiy, V.M.:

1955.

351

(Ob osnovnykh sootnosheniyakh t e o r i i tonkikh obolochek).

P M M , Vo1.25,

No.3, 1961, 2 3 . Kilfchevskiy, N . A . (Kiltchevs'kiy, M.O.): a) Generalization of t h e Modern Theory of S h e l l s (Obobshcheniye sov-

remennoy t e o r i i obolochek). PMM, Vol.2, No.4, 1939; b ) P r i n c i p a l Equations of t h e E l a s t i c i t y of S h e l l s and Working Methods of I n t e g r a t i o n (Osnovni rivnyannya rivnovagi pruzhnikh obolonok i deyaki metodi i k h integruvannya). Zbirnik p r a t s ' I n s t . mat. AN URSR, Nos.4, 5,

c j The Theorem of Work and Reciprocity and t h e Construction of Greents Tensor i n t h e Theory of S m a l l E l a s t o p l a s t i c Deformations (Teorema o vzaimnosti r a b o t i postroyeniye tenzora Grina v t e o r i i malykh uprugoW, V 0 1 . 2 1 , N0.5, 1957; plasticheskikh d e f o m t s i y ) . d ) On Axisymmetric S t r a i n s and E l a s t i c S t a b i l i t y o f a C i r c u l a r Tube Under t h e Action of h n g i t u d i n a l Compressive Forces (Ob osesimmetrichnykh deformatsiyakh i uprugoy ustoychivosti krugloy truby, nakhodyashcheysya M M , V 0 1 . 6 , 1942; pod deystviyem prodol'nykh szhimayushchikh s i l ) . P e ) P r a c t i c a l Methods f o r Equations of Motion i n C y l i n d r i c a l S h e l l s (Nablizheni metodi biznachennya peremishchent v t s i l i n d r i c h n i k h obolonkakh), Zbirnik p r a t s t I n s t . mat. AN U B R , No.8, 1946; f ) Study of Certain Questions of t h e Theory of E l a s t i c i t y (Issledovaniye nekotorykh voprosov t e o r i i uprugosti) Izv, Kiev. politekhn. I n s t . Vol.15, Gostekhizdat UkrSSR, 1954; g) I n t e g r o d i f f e r e n t i a l and I n t e g r a l Equations of Equilibrium of Thin E l a s t i c S h e l l s ( I n t e g r o - d i f f e r e n t s i a l fnyye i i n t e g r a l b y y e uravneniya ravnovesiya tonkikh uprugikh obolochek). PMM, Vo1.23, No.1, 1959; h) Functional Equations of Equilibrium of Thin E l a s t i c S h e l l s (FunktIzv. sional'nyye uravneniya ravnovesiya tonkikh uprugikh obolochek) Kiev. politekhn. I n s t . , V o l . 3 1 , 1961; i) I n t e g r o d i f f e r e n t i a l and I n t e g r a l Equations of Equilibrium of Thin E l a s t i c S h e l l s ( I n t e g r o d i f f e r e n t i a l - und Integralgleichungen f& das . angew. Math. U . Gleichgewicht d b e r e l a s t i s c h e r Schalen). Zschr. f Mechan., Vol.kO, No&, 1960; j ) On Working Rules f o r I n t e g r a l Equations, Constructed on t h e Basis of t h e Theorem of Work and Reciprocity; i n Collaboration w i t h G.I.Tkachuk (Pro deyaki v l a s t i v o s t i i n t e g r a l f n i k h r i v q a n ' , skladenikh na osnovi Prikladna mekhanika, teoremi vzaemnosti r o b i t ; sovmestno s G.I,Tkachuk).

6 1940;

0 0 1 . 5 , N o . 2 , 1959.
Korolev, V.I. : Thin Two-Layer P l a t e s and S h e l l s (Tonkiye dvukhsloyqyye o 1 . 2 2 , 1955. p l a s t i n k i i obolochki), Inzh. sborn., V 25. Lurfye, A.I.: a) The General Theory of Thin E l a s t i c S h e l l s (Obshchaya t e o r i y a uprugikh o l . 4 , N o . 2 , 1940; tonkikh obolochek). PMM, V b) Contribution t o t h e Theory of Thick Slabs ( K t e o r i i t o l s t y k h p l i t ) .

2 4 .

PMM, v01.6, Nos.2, 3, 1942;

c) O n t h e Equations of t h e General Theory of E l a s t i c S h e l l s (Ob uravneniyakh obshchey t e o r i i uprugikh obolochek). PMM, V o l . L ! t . , No.5, 1950. 26. Gol'denveyzer, A.L. and Lnrrtye, A.I.: O n t h e Mathematical Theory of Equilibrium of E l a s t i c S h e l l s (0 matematicheskoy t e o r i i ravnovesiya uprugikh obolochek). PMM, V o l . l l , No.5, 1947. Contribution t o t h e Theory of S h e l l s 2 7 . Mushtari, Kh.M. and Tere ulov, 1.G.: of Moderate Thickness K t e o r i i obolochek sredney tolshchiny). Dok. AN

352

SSSR, Vol.128, No.6, 1959. On t h e E r r o r of t h e Klrchhoff 28. Novoahilov, V.V. and Finkel'shteyn, R . M . : Hypothesis i n S h e l l Theory (0 pogreshnosti gipotea Kirkhgofa v t e o r i i obolochek). PMM, Vo1.7, No.5, 19.43. U n e a r Equations of t h e Dynamics of an E l a s t i c C i r c u l a r 8. Nigul, U.K.: C y l i n d r i c a l S h e l l Free of Assumptions (Lineynyye uravneniya dinamiki uprugoy krugovoy t s i l h d r i c h e s k o y obolochki, svobodnyye o t gipotez) Trudy Tallin. politekhn. Inst., Ser.A, N0.176, 1960. Contribution t o t h e Theory of Vibrations of Thin P l a t e s 30. Petrashen', G . I . : ( K t e o r i i kolebaniy tonkikh p l a s t i n ) . Uch. zap. Leningrad. gos. d v . , No.L!+9 ( s e r . matem. nauk, No.&), 1951. 31. Remizova, N . I . : a) I n t e g r a l Equations of Equilibrium o f Thin E l a s t i c C y l i n d r i c a l S h e l l s ( Integral'rryye uravneniya ravnovesiya tonkikh uprugikh t s i l i n d r i c h e s k i k h obolochek). F'MM, Vol.23, No.3, 1959; b) Contribution t o t h e Development of a Method of I n t e g r a l Equations Applicable t o C y l i n d r i c a l S h e l l s ( K r a z v i t i y u metoda i n t e g r a l nykh Izv. Kiev. uravneniy p r i m e n i t e l t n o k t s i l i n d r i c h e s k i m obolochkam) politekhn. Inst., Vo1.31, 1961. 32. Selezov, I . T . : Study of t h e Propagation of E l a s t i c Waves i n Slabs and S h e l l s (Issledovaniye rasprostraneniya uprugikh voln v p l i t a k h i oboloThesis, I n s t . Mekhan. AN UkrSSR, 1961. chkakh) Some Problems of S t a b i l i t y and Vibration of Anisotropic 33. Khachatryan, A.A.: P l a t e s and S h e l l s (Nekotoryye zadachi ustoychivosti i kolebaniy anizotropnykh p l a s t i n o k i obolochek). Authorts Abstract of Candidate's D i s s e r t a t i o n . Yerevan. politekhn. Inst., 1961. 3 4 . . Fradlin, B . N . and Shakhnovskiy, S.M.: a ) Contribution t o t h e Calculation of a Plane S h e l l , Rectangular in Planview, o f Double Curvature i n t h e Cases of Rigid and Hinged Attachment of i t s Contour (K raschetu pryamougoltnoy v plane pologoy obolochki dvoyakoy k r i v i z n y v sluchaye zhestkogo i sharnirnogo zakrepleniya eye kontura). Izv. vyssh. ucheb. zaved. S t r o i t . i a r k h i t e k t u r a , No.9, 1959; b ) Equations of Motion of Plane S h e l l s , with Introduction of AwdZiary Ioads (Viznachennya peremishchenc pologoi obolonki, shcho znakhodit t sya Prikladna mekhanika, Vo1.6, No.4, p i d dieyu doviltnogo navantazhennya)

'

1960.

353

AUTHOR IILDEX

Akhiyezer,N.I. Aksel rad, E . L Alumyae N A hbartsumyan, S . A

f

,..

Basset, A. Bekkenbakh,E.F. Berezovskiy,A.k. Birger, I .A Blas chke J i Eogolyubov, 1J T I 3olotin,V.V. Brown Bubnov Castigliano Carson Cauchy Chaplygin,S . A Chebyshev,P .Lo Christoffel Clebsch Courant,R. D T A1e& e r t Darevskiy,V . M . Dirac D i r ic U e t Ditkin, V . A Einstein, A. Zpstein,R.S Euler Feinberg, D .V Finkelfshteyn,R.M. Fourier Fradlin,V . N Fredholm Gabelya, S P Galerkin,B .Go Ga.limov,K.Z. Gauss Geckeler, J. Gilbert D Golfdenveyzer,A.L. Goncharov,V .Ya. Gnuni,V T s Green

,.

15, 31ci. Grigolpk,E.I. 225, 3L6 I-Iamilton 31,6 Heaviside 1 1 , 88, 171 Eertz,H. 346, 317 FIil1,R. 169 Hooke 315 Huber 3L4 Il?yushkin,A.A.
1 1 21

225, 347 10, 281 13Li 151 10 181 227, 252, 253
181 181, 189
10

..
.

198
1 1 , 118, 222, 3L6 223 28L

171
32C

85, 86, 90, 296

10 222

. . . ..

255, 256, 272 73 331 IL, 169 87, 347 132 116 321 19 170 257, 262, 263, 331 285, 347 87, 348 lW,1L8, 1L9

Jacobi,K. Kachanov,L.M. Kagan,V.F. Kantorovich,L.V. Karman,T. Kazantseva, G . Ye Kennard,E.H. K i l f chevskiy,N.A. Kirchhoff,G. Kirsa,L.V. Khachatryan,A.A. Kochin,N.Ye. Ko~oukhov,N.V. Korolev,V.I. KraussbF. Krendel1,S. Krylov,A.?J. Kupradze,V .Do Kuzenlko k . ! , I Lagrange

181

3L6
211

,.

3, 12, 88, 89, 10, U,267, 72, 7.4, 76,

12, 3L6, 3L7, 14I-S

..

3L7 285, 3L7

Lame Landau,L. Lantsosh,K. Laplace 318 Lauricella 3L.3 Lavrenttyev, M.A. 31L7 Le C h a t e l i e r 28L Legendre 3L6 Leybenzon, L.S 271 Levi-Civita,T. 101 L i ch D 331 L i f s h i t s , E . 3LE; Love,k. I 231~ Lur ye, A .

275, 332, 33L 321 170, 173 3h2, 3L6, 3L7 73, 85, 2L8, 263 3L2 171, 3L8 3r,6 2 7 1 L 225, 347 168 198 10L 103, 269 347 1 0 , 225, 258, 262, 263 25L
222

15, 290, 34.0 279, 320

e5
291 223 293, 29k 121, 123 w, 255 3L5
222

,.

77, 81, 85, 86, 1 6 ~ 12, 9L, 169, 320, 346 169, 326

x- Page numbers r e f e r t o pagination i n the o r i g i n a l foreign text.

354.

Mellin Mieses lilikaylenko R .A Mikeladze Mitropol*skiy,Yu.A.

321 Trefftz,E.
le1

MOlOtkOVy

LJ

hshtari,Kh.M. ~luskhelishvili,M.I New" Newton Nigul,U.K. IvovozhiLov,v .v Onashvili, 0 . D Ostrogradskiy,ltl.B Panovko Ya. G. Peshtmaldzhan,D .V Petrashen,G,I. Petrenko ,ILP Pfaff Pogorelov,A.V Poisson Prudnikov,k,N. ' Pugachev,V.S. Rashevskiy,P.K, Reissner

3Q Uflyand,Ya,S. -179 'Veblen,O. . 198 Vekua,I,N. 99, 225, Vlasov,V ,Z 12, 88, 89, 122, Voigt . 169 Volfmir,A,S,
I

1 1 , 88, @ , 2 1 0 , 217

S6 Volterra 116 Watson, G.N.

3A5 l\kitt,aker, E . T
12, 59, 73, 87 1 0 , 258, 262, 263, 281

. .
.

5e, 2 G 3

291 291

r.5

1 2 , 3Lb

. . .

198 347 99, 3SL lC3, 120 78 223 85, 86, 90 321
180

Remizova, N , I Ricci Riemann Rite Rodrigues,O. Saint-Venant Salvadori,M.D SchoutenSeidel Selezov,l.T.

hL, 346 121, 170, 1T1, 173, 317 318 9L 321 236, 258, 262,
26L 2l 85, g6, 252, 270

314
268

Shabat,B . V Shaknovskiy, S .Id. Shevchenko,Yu.N. Shtayeman, I .Ya, Sinyavskiy,A .Lo smirnov,v . I Somigliana StruikyD. * Taylor Terguelov,I.G, Timoshenko, S ,P Tkachuk, G. I

94, 103,

120, 318

io?,

31L

291 34e 330 274 285, 31~7 258,' 269 85, 298
288

50, 51, 150, 155 122; 169, 348 103, 170, 201 3h7 355

. .
I .

SUBJEZT INDEX
Algebra, t e n s o r Alternating f i e l d s

Alternation Analysis, t e n s o r 16, 17 Approximation of functions 274 Basis, coordinate 2.4 Basis, l o c a l 18, 20, 289 ?3asis, r ecipro c a l 23, 26 Boundary conditions 115, 240, 245, 252 Branch point 338 Carrying power 20L Carryi-ng power of s h e l l 221 Chain s t r e s s e s i n s h e l l 240, 2W, 244 Charact e r i stic of d i f f e r e n t i a l equations 112 C h r i s t o f f e l synbols of t h e f i r s t kind 38, 57, 64, 65 C h r i s t o f f e l symbols of t h e second kind 37, 38, 65, 255, 256 Coefficient, Fouriers 1.48 .Coefficient o f c o r r e l a t i o n 190, 191 Coefficient, P o i s s o n f s 190, 191, 242 Components of body f o r c e s 263 Components of displacement 113, 268, 281 vector Components of f i n i t e deformation t e n s o r 272 Components of s t r e s s tensor 113, 114 -Components of s t r e s s 268 vector Components o f tensor, contravariant 24, 25, 62 Components of tensor, covariant 27, 62 256, 266 Components o f v e c t o r .Components, t a n g e n t i a l , of s t r a i n tensor 21J_ Conditions of completeness 262 .Conditions of i n t e g r a b i l i t y 295 Conditions of p l a s t i c i t y 188 Conditions o f p l a s t i c i t y , Huber-:vIieses 181

307- Contraction 22&, 230, 232, Coordinates, 235, 236 Coordinates, 33 Coordinates,
Coordinates, Coordinates, Coordinates, Coordinates,

of t e n s o r s arc Cartesian curvilinear N e r excess Gaussian generalized

31

17,

254 286, 288 287, 288,

289
72

151, 330
228, 258, 278, 279

69

Coordinates, Lagrangian Coupling, geometrical Coupling, kinematic C r i t e r i o n of i n s t a b i l i t y C r i t i c a l load Curvature 105, 266, Curvature, p r i n c i p a l Deformation of b e n t shell Deformation, p l a s t i c Delta-function Density o f energy of deformation Density o f k i n e t i c energy Density of Lagrange function Density of s h e l l Derivative, absolute Derivative, covariant 45,

69
277, 2.78, 27.9 277, 279, 282 214, 220

338
261, 265, 270, f f - 238
. >

238

176
132, 262 226 226

226
112, 225, 274, 283 258

u,

329 105

Derivative, f u n c t i o n a l Determinant, fundamental 19, 63 Determinant of system 323 D i f f e r e n t i a l of tensor, abs o l u t e 39, 40 D i f f e r e n t i a l of vector, absolute 36, 37, 39 Direction cosines 289 Displacement f i e l d 121, 227, 252 E l a s t i c constants 329, 339 Elastic l i m i t 204

Page numbers r e f e r t o pagination i n t h e o r i g i n a l foreign text.

356

ELement of a r c Element of area Element of s h e l l

ELement of volume Elementary solutions of t h e three-dimensional problems of e l a s t i c i t y theory Elementary work of deformation 2 2 6 , Energy, k i n e t i c
Ehergy of acceleration Energy of deformation Energy of deformation, quasi-spe c i f ic Ehergy of deformation velocity Energy, s p e c i f i c Equations, biharmonic Equations, Lagrange, of t h e second kind Equations, ~am4 Equations, nonlinear La" &pations of v a r i a t i o n Error, non-removable Existence of t h e boundary problem Finite differences Flexural s t r e s s e s i n shell Flexure, strong Force dipole Force i n t h e s h e l l Forces, body

126, 261 122, 126, 308 1 0 6 , 2 2 6 ,2 5 8 , 260, 266, f f . 318

2 8 6

2 2 9 . ,238, 277 271 2 2 6 Function , approximation 194

Function, generalized Function, Heaviside Function, Lagrange

Formula, Greent s 285 Formula, i n t e r p o l a t i o n 2 7 3 , 276, 2 7 8 Formula, i n v e r s e transf o m t i o n B7 Formula of mechanical quadratur.es 322 Formula, OstrogradskiyGauss 336 Formula, reduction 1 0 2 , 103, 105, 1 0 3 Formula, Riemann4ellin 321 Formula, Taylor 62, 63 Functional 257, 2 5 8 , 262,

268 2 5 4 133, 340, 343 151, 230 225, 229, 2 3 2 ,

251, ff

2 8 1

62, 2 0 0 ,2 8 0 66, 254

1 2 4 , 129

Function, transcendental 321 Hy-potheses, KirchhoffLove 86, 87, 165, 263, 270 Index, dummy 1 8 I n e r t i a of r o t a t i o n of 120 s h e l l element I n e r t i a l term 334, 138 I n i t i a l conditions 1 1 8 , 130, 267,

253

2 7 1 , 277

I n i t i a l v e l o c i t y of 269, 285 deformation 2 7 1 311 313 Integral, contour Integral, curvilinear 2 5 4 , 310 I n t e g r a l equations of 24Js h e l l theory 285, 310 338 290 I n t e g r a l , improper 295 163, 165 I n t e g r a l transformation 321 1 0 9 ,2 5 4 , 262, I n t e g r o d i f f e r e n t i a l equat i o n s of s h e l l theory 285, 306, 309, 2 6 3

Forces, concentrated Force s d i s s i p a t i v e Forces, generalized conservative Forces, generalized nonconservative Forces, quasi-body Forces, surface

103, 286, 309 345

2 8 1
,

326
Invariant of coordinate 2 4 transformation Invariant of deformation tensor ( l i n e a r ) 58 I t e r a t i o n process 270 I t e r a t i o n process, GaussSeidel 314, 325 Kernel, focusing 298, 311, 324 Kinematic conditions 262 Kronecker d e l t a 2 6 , 62, 251 ~ame' c onstants 58, 226 Laplace-Carson transformation 320

2 8 1

2 7 3
loq, 2 5 4 ,2 5 7 ,

262, 2 6 3

Form, fundamental quadratic Form, P f a f f F o d a , d i r e c t transformation

1 8 7 8

287

357

Lzw, generalized Law, Hooitef s

57 168, 171, 173

227, 252, f f

Law, Newtonfs Third Law, nonlinear Law, Voigt-Xurnaghan Linear theory of s h e l l s Line, coordinate
Line of curvature Load, a d d i t i o n a l Load d e n s i t y Load, focusing Load, p r i n c i p a l Local p e r t u r b a t i o n s Longitudinal shock t o rod Lowering of i n d i c e s Nethod, d i s c r e t e continuum Method, of Bubnov and Galerkin Kethod of eqLtivalent linearization Ilethod of l e a s t squares Method o f networks Kethod of p e r t u r b a t i o n s Method of reduction of Cauchy and Poisson Xethod of successive approximat i o n iiiethod, Eonte Carlo Iiiethod, RitzTs

71 59, 6 0 60
237 20, 25L, 260, 261, .288 21, 25h 299 290, 292 290, 307, 310,

311
299 290 271 32

Operator, TTforcefr lu9, 110 Operator, Laplace 67, 279 Operator of p a r a l l e l displacement 135 Optimum system 180 Operator, wave 112, 279, 263 Original 326 0sc i l l a t i o n s, nonstationary 320 Oscillations, stationary 236, 319 Faraboloid 312 P a r a l l e l displacement of tensor 1J-r 42, 43, 45,

165, 117

326
28L

339
267 275 251, 253

85, 90
1 0 1 , 262, 28C

315
236, 258, 262, 264

Hethod, rtsemi-inverserf 170 Kethod, Somigliano 298 Ketric k i the s h e l l 20, 21 Wetric of deformed medium 61 I i e t r i c of space 19, U3 Noment, bending 248 Moment, t w i s t i n g 165 Natural boundary conditions 161 Katural i n t e r v a l of approximation 245 !:odes of network 277 Wonlinear i n t e g r o d i f f e r e n t i a l 331 equations of s h e l l theory i\!ormal, p r i n c i p a l 268 Operator calculus 321 Operator, dii'f e r e n t i a l 1C9, 264

P a r a l l e l d isplscement o f vector Path of coriiparison Path or" r e a l motion Pernutation of i n d i c e s Plate Point of d i s c o n t i n u i t y of' t h e f i r s t kind Polynomial, i n t e r polation Polynomial, Legendre P r i n c i p l e , Castigliano Principle, DTAlembertLagrange P r i n c i p l e , G a m s , ol" l e a s t constraint P r i n c i p l e , Le ChatelierE raun P r i n c i p l e of po ss i b l e displacements Principle of variations P r i n c i p l e , OstrogradskiyH a m i lt o n Problem, dynamic Problem, externs1 Problem, first boundary Troblem, i n t e r n a l Problem, secocd boundary Problem, s t a t i s t i c a l Problem, three-dimensional o f e l a s t i c i t y theory 253, Problem, two-dimensiocal ol" s h e l l theory Problerr?, v a r i a t i o n a l Product of t e n s o r s Programing PseGdos c a l a r

117 131, 261

291

33
240, 265 292 312 293, 294

171

1-4, 169

14, 222
222

330
281, 282 239 72 115, 278 72 116, 279 239, 21L4

118

267, 27IL

253
269

31.
28L

36

358

36 Pseudovector 232 Quadratic approximation Quantity, s c a l a r , see S c a l a r Quamtity, tensor, see Tensor Quantity, vector, see Vector Radius of curvature 311 Radius of curvature, p rincj .p a l 21 L9, 5 C Radius vector of a point Ra.ising of index 32 25, Rank o i ' tensor Recurrefit dependence lG7 Regior, o f approximation, choice of 216, 251, 276, 277 Region of p l a s t i c i t y 1 8 8 (boundary) Region o f s t a t i c 221, 222, 22L instability Representation o f a 326 function 320 Resonance l9G R i g i d i t y of m a t e r i a l Rod 276 Saint-V enant s compatibility conditions of deformation 5G, 57, 171, 173, 252 Scalar 24, 27, 29, 136 2L, 3 6 Scalar, ab s o l u t e 3 c d a r field 214 S c a l a r product OP v e c t o r s 31, 291, 304' S e r i e s , Fourier 141, 1L8, 149 S e r i e s , MacLaurin t e n s o r 169, 326 S e r i e s , Taylor 5C, 51, 1 5 G , 155, 283 Shearing s t r e s s e s i n s h e l l 120 S h e l l , c i r c u l a r conical 22 Shell, cylindrical 22, 2h8 Shell, homogeneous 225, 232, 23t, 252 Shell, inhomogeneous 229 Shell, isotropic 225 Shell, layered 7c, 225, 229, 232, 252 S h e l l , moderately thick, and t h i c k 87 S h e l l of constant t h ickne ss 261, 287 S h e l l of quasi-variable thickness 221 270 S h e l l , plane 321 Shell, t h i n
f

Singular l i n e Singular poirit Space, Euclidean Space, non-Euclidean Stress field

286, 29L 286, 289 L2, L3, 56

r 3 176, 193, 203,

227, 252

Surface, base
Surface, Surface, Surface, Surface, boundary contour coordinate niddle

lOL, 226, 229, 237

1c2, 255 253, 255, 267 287 26C, 26t3, 270, 28 7 2 G L

Supporting power Spm"trization 33 System of m a t e r i a l p o i n t s 130 Tangential p a r t of t h e s t r a i n tensor 1 C 4 , 105 Tensor, antisymmetric 33, 34, 335 Tensor, curvature (TCicmanri-Christoffel)L2, Lb3, 91, 105 Tensor, s t r a i n 2G2, 217, 228, 235, ff 58, 59 Tensor e l a s t i c i t y Tensor f i e l d LC Tensor, f i n i t e deformation 55, 161, 332 Tensor, Greenfs 3 1~7 Tensor, metric 29, 30, 25L Tensor, mixed 31 Tensor, m u l t i p l i c a t i v e 31, 35 Tensor, small deformation 53, 335 Tensor, stress 200, 228, 235, 238 Tensor syrmnetric 33 Tests o f s h e l l s 210, 217 Theorem of Fredholm (third) 3L 3 Theorem of i n t e g r a l mean U.6, 296, 297 Theorem, OstrogradskiyGauss 125 Theorem o f r e c i p r o c a l work 79, 80, 285, 298,ff Theorem of r e c i p r o c a l work, generalized 286 Theory of Newtonian 265, 289 pot e n t i a 1 Thickness of reduced shell 2L4 Torque 165

359

Triangulation n e t 275, 328 True motion of a system 271 Unbounded e l a s t i c medium 286, 299 Uniqueness of boundary problem 214, 269, 285 Variation 155, 156 Vector 2-4, 27, 29 Vector, contravariant 288 Vector, displacement 52, 253, 267 Vector, p o l a r 36 Vector product, see Vector Vector product of v e c t o r s 35 Vector, s t r e s s 163 Velocity, generalized 228, 229 Velocity, instantaneous angular 39 V i r t u a l work 147 Wave processes i n p l a t e s and shells 173 Young's modulus 58, 203

by 0 . V .

Translated f o r the National Aeronautics and Space Administration Leibiger Research Laboratories, Inc.

The aeronautical and space activities of the United Stdes shall be conducted so as to contribute . . . to the expansion of human knowledge of phenomena in the atmosphere and space. The Administration shall provide for the widest practicable and appropriate dissemination of information concerning its activities and the results thereof.

-NATIONAL AERONAUTICS AND

SPACE

ACT OF 1958

NASA SCIENTIFIC A N D TECHNICAL PUBLICATIONS

TECJ3NICAL REPORTS: Scientific and technical information considered important, complete, and a lasting contribution to existing knowledge. TECHNICAL NOTES: Information less broad in scope but nevertheless of importance as a contribution to existing knowledge. TECHNICAL MEMORANDUMS: Information receiving limited distribution because of preliminary data, security classification, or other reasons. CONTRACTOR REPORTS: Technical information generated in connection with a NASA contract or grant and released under NASA auspices. TECHNICAL TRANSLATIONS: Information published in a foreign language considered to merit NASA distribution in English. SPECIAL PUBLICATIONS: Information derived from or of value to NASA activities. Publications include conference proceedings, monographs, data compilations, handbooks, sourcebooks, and special bibliographies. TECHNOLOGY UTILIZATION PUBLICATIONS: Information on technology used by NASA that may be of particular interest in commercial and other nonaerospace applications. Publications include Tech Briefs; Technology Utilization Reports and Notes; and Technology Surveys.

Details on the availabilify of these publications may be obtained from:

SCIENTIFIC AND TECHNICAL INFORMATION DIVISION

NATlO NA L AERONAUTICS AND SPACE ADM I NISTRATION

Washington, D.C.

PO546