Beruflich Dokumente
Kultur Dokumente
By
NANCY BRAZELL BROOKS
and
N. M. NEWMARK
Technical Report
to
OFFICE OF NAVAL RESEARCH
Contract N6ori-071(06), Task Order VI
Project NR-064-183
UNIVERSITY OF ILLINOIS
URBANA, ILLINOIS
mI
THE RESPONSE OF SIMPLE STRUCTURES
TO DYNAMIC LOADS
by
Technical Report
to
April 1953
University of Illinois
Urbana, Illinois
i.
CONTENTS
Page
uIST )F FIGURES
ACKNOWLEDGMENT
I. INTRODUCTION
APPENDIX
LIST OF FIGURES
ACKNOWLEDGMENT
I. INTRODUCTION
A. INTRODUCTORY STATEMENT
pends upon the characteristics of the impulsive excitation and upon the physical
keeping with the accuracy of the analytical procedure, the use of rigorous
the load and the structure involved. However, there are always certain in-
and the load-deformation characteristics of the material from which the structure
is made. No matter how "exact" is the analytic procedure for determining re-
ture to various patterns of impulsive load is required before one can investi-
gate the effect on the response of deviations in the forcing function parameters
For this fundamental study the force functions and the structure con-
sidered are ones which can be defined by simple mathematical expressions. The
impulsive load patters include those blast and impact disturbances described
parameters. For a specific impulse pattern the average applied force and the
duration of the load are the force parameters considered. The mass of the
structure, the rigidity, and the elasto-plastic properties of the material from
which the structure is made comprise the characteristics of the structure. Com-
of these parameters. To this end, the results of the analysis used are pre-
employed for determining the response of the system, its maximum response, and
B. NOTATION
The terms used are defined when they first appear in the text. They
K = spring stiffness
M = mass
Xm
X = displacement of mass
t = time
A. DEFINITION OF PROBLEM
a mass supported by a spring which has no mass as shown in Fig. 1. When sub-
of motion:
it
the mass; Q(X), the resisting force in the spring; and P(t), the impulsive
at t = 0, X X= , (2)
resistance until maximum displacement is reached. (Fig. 2). The elastic re-
spring. When the response is greater than the yield deflection, the spring
resistance is constant:
3. Forcing functions.
P1 being the average applied load and tl, the duration of the applied load.
(Fig. 3-b). In addition to these two impulse patterns, the triangular forc-
ing functions with a terminal and an intermediate maximum force are considered.
ing function:
where at I defines the time at which the maximum load occurs (0azl). The
1. General response.
X/Xy = Acoscut + Bsinwt + P(t)/ P 1 , for XSXy and P(t) linear; (9-a)
frequency of the system, and P is the ratio of the yield strength to the
CU = K /M(10)
13 = y/PI (11)
2. Maximum response.
By setting the velocity of the mass equal to zero, the time, tm, at
The graphs (Figs. 4-7 inclusive) which represent the maximum re-
sponse furnish a better picture of the system's behavior. For various ratios
of yield strength to average applied load, 0, the maximum response with re-
load and the period of the structure, tl/T. These charts show the relations
between the duration of the load, tl, the time to attain yielding, ty, and
the time of maximum deflection, tm, An accurate estimate of the maximum re-
able from these response graphs provided the forcing function and the
The type of impulsive excitation may, or may not, affect the maximum
response of the structure. In the so-called impulse region when the duration
interest to note that the pattern of load does not significantly affect the
maximum response, other things being equal. (Fig. 8). In contrast to this,
the impulse pattern may greatly influence the maximum deflection when the
load is applied very slowly relative to the system's period. (Fig. 9).
8.
Of the forcing functions studied, the triangular pulse with an initial peak
force produces the most critical response in the long-duration region; the
to what extent a change in one parameter, the duration of load t1 for instance,
affects the maximum response. In Fig. 5-a, which represents the response of
the structure to an initial peak force, for P equal to 1.0 a twenty percent
there are two cases of particular interest: one in which the duration of the
applied load is less than the time at which the maximum response occurs; the
other in which the load terminates long after the maximum response is attained.
history of the system's behavior before the velocity of the mass is appre-
equal to the yield deflection, and the mass is at rest when the load is applied:
X = Xy ; (2-a)
x =o. (2)
9.
The general solution to this problem is of the form of equation (9-b), where,
as before, the constants C and D are determined from the initial conditions
for velocity and displacement. The maximum response is found in the same
manner as previously described for the more rigorous solution. These ex-
pressions for maximum response are included in the appendix. How closely
this approximate response agrees with the actual response in the impulse
region is illustrated in Fig. 10, for the case of the step-pulse function.
Similarly, for the various triangular forcing functions the agreement be-
When the maximum deflection occurs long before the load ceases, the
time at which the response is a maximum may not significantly exceed the
time at which the spring first yields. For this reason, neglecting the elastic
response of the system is not an accurate assumption for this case. A better
"exact" solutions for maximum displacement allowing the duration of the load,
tl, to go to infinity. The expressions which result define the maximum re-
sponse for a duration of load which is long in comparison with the period of
the structure. The approximate solutions thus obtained are exact for the
the maximum response for the triangular pulse functions as shown in Fig. 11.
A. DETERMINATION OF SENSITIVITY
may be defined as the change in maximum response due to a change in the para-
to that parameter represents the influence factor sought. Then the summed
Dividing this expression by the maximum displacement, Xm, gives the dimen-
sionless relation:
Xm
-y= C AP1 + CQQy + Ct A l +
CK--K- + C1-4 (12-a)
where Cp, CQ, Ct., CK, dnd CM are a measure of the sensitivity of the response
of the structure:
0= 6Xm Pl (13)
F1 Xm'
and similarly for the other parameters. These influence factors are charted
in Figs. 12-15 for the step-pulse function and the initial-peak triangular
forcing function. For each type of impulse pattern studied the analytic
expressions for the influence factors are included in the appendix. Because of
the dimensionless form of the formulae for maximum response, certain relations
This allows the influence factors Cp, CK, and CM to be defined in terms of
CQ and Ct .
11.
expressions given in the appendix may be used to find the change in maximum
response and the corresponding influence factors are obtained either from
the graphs or from the analytic expressions. When substituted into equation
(12-a), these quantities and the magnitudes of the relative changes whose
the parameters, equation (12-a) may be used with a certain degree of accuracy
for finite changes in the parameters. It can be shown that the error in-
error; that is, by considering the variations in ten percent steps, or less,
and replacing the old structure by the new at each step, the change in maxi-
C. SIGNIFICANCE OF RESULTS
The graphs for the influence factors illustrate the order of impor-
tance of the changes in the various parameters. In the impulse region the
average applied load is generally the most critical in its effect on the
12.
maximum response with the duration of load second in importance. The average
applied load again has the greatest influence on the maximum response in the
the influence factors. In the long-duration region the influence factor for
average load, for instance, may be as high as 8 or more. Then too, a varia-
The duration of load, for example, has little effect in the long-duration
regions for 0 greater than 5.0. The range over which the influence factors
function parameters. If these quantities are not precisely known, the effect
pressions for maximum response given in the appendix are accurate for
estimating purpose in the impulse region only. In certain cases the re-
often a matter of arbitrary selection. When this condition exists, the use
mate procedures which are in keeping with the quality of the data available.
14.
p. 159.
1. Step pulse.
(A0. o$ tn to 6i.
(M). si ,
wi.m= 0.(e.a)tr
cq ni , ..
C t /.
(4). xm X, o. iain'lz
(). C a.
(4). CCs
(4). -(4)
y
!1 (0.e +Im.= .
16.
Si-
(4). C ,6-I
i6;i, 6r -.
(U.aos et - -6
(4).CX -,+ _ -
m ct 4,;,-,,,.).
(). CQ:o.
( o *
4). + 4 fay
17.
(M). Cq:o
W. C : _ - des coi
( ). XmX : m (.mo)
QXm /X
('). Cg . -
-- -L (cot
,f,aa J, s4- 1i-no,,,1,).
(5) . Cq mI.
.g .
,-xdo
t. sca)..~t
1 4l no +
().C , ,
*().
04isI. 4 1 i - n r.
0).eI- iz- oSinct
18.
(i). X/Xe~:
.,-f. - c4~ta,
(W. 1 Is/A
(9). CM IZ
0f)- eat,,
(S5) ct % -I
xm + A idokA12,+ 1-ccscatTJ
Q &I,
(Z).C ,w 0.
(4). C. L~
(B). o% 4 c.2% r . -
(I). ~. 1 : ocoic4I
.~(i-Oi~g~ - i-s(1 co1~t
1
N-K /X 4
M(5)<.C< It -e .1%L
evein(vh
"~w+ +v J_ (Wr--a.Jtv)" (1
jb.&+'a 'V (t.A -od/ 4)I a +
(0) Gin~
at, 4l' cot1
22.
(4). Cq'dl
an. Ct Vi (a tol
0+ 4.4 r J
+ 'AL
-41. ctIA-e
I / + II.AC(.c~)
(41 Ct OI=
(D.~c.4 t,4 6i
%4t i I i W
(I)
23.
-+ C("t,-
j(4).t i- .ol i t o t
j''I
1. Step pulse.
(A). Impulse region.
Xr,,x + II
2. Intal-peac triangular pulse.
(a). Xei/X
1 ' )[edtI )J
error - C p (1)
Cp
where Cp
Axm p
and C = lin
C p#O ap In
a e b . for a a b
Cp
or a a 1 b
a+l b +l (16)
Cp
This restriction limits the magnitude of the change in the parameter under
consideration to:
where a and b define the desired accuracy, Cp is the value for the influence
factor in the region under study, and AXm/Xm is the allowable relative
__________ P(t)
P (t)
xyx Xm X
P,
aSTEP-PULSE
P
I P,
b. INITIAL-PEAK
ti
2 1P -- -- -- -
cTERMINAL-PEAK
d. INTERMEDIATE -PEAK
14-
to -
NN
40
0 0
002Q4 / 0.L
'J -it7-it
li X1 i I --
110Ii
I- - --
pop -
g/As
-~~ 14t'
~
Nie-
Mr0.
/lox
MNO
-4 -
6r*.- 16P -
T-MUA PU-LSE
-0 ------ - -
300
-~~ -_4 --
--- 00
400
000
0.00
TRIANG! R PULSE /
-34.
ic:. I CL
-- I-
.---- I~ V t-it
- - -- - - t--~ -~ - jC-
- - -- -- - - - -. ~-- -- -CI
- ,- - - > ~ - , E\
-h I {4.-
-K,- -
IX* -4-a.-
-1-l
CL
C.
*iK-u I
do
qu
A X/ w
-UT-.- -- ~1* --
- - - - - - - ~ .-.-...s~--.- -- - - -
- - -~-.~4~.---4.~-.-- .~ j.--4 - - - -
21 111 IIZ:
-- - - .-.-.
7ff
- - - - - ---. ~ * -. +~---- -
1~~
- i----. * * -K
-~ j- 0
-- - -- *-................1~~ 4
- i.bt.L4- ~ -*. -
I,, ________
-i- -i-I -
- -
* -- -,----~
*7~i~~ --
-
~ ~~-{.........-
-~
hi
I-
-.
*~ -
. -I - -
- - II-,
- hi
I - -
.L2
- .4 ~..*...
-
-*-**t--.-~ - - - - -
0
2
C
2
-- - 37.
O.-.
-- - 4 --.- ~-Ink.S
.4 44Mom
(I - - I0
4 - -- --.-- +-.- - - i0
-~~ ~~~~
- -~~~ - -4 -O -- - - --.- 4.
I E
so --- - AV4S,-
-- 7 010
70----
60 _ _1 ~ ~ -
so- - -
(o/
'II
J-1B
I~ 0
Z 4)
CL
w0
0 c0
o~ ta 0
60 0
TO
40
i'~ ttI
IZ C
I ~ I-- S--------- .C
1
7. ....
....
x /OX
4---
/100
f 4-
co 0K .4.
FW 12 o-
70.T
AVRG
0 .4 t. 0.
LODAN
.4 1.0I
ILDLA
INFUEC R FO TP ./C IUS
- - - - - I
- -~ -. ~3.
I - ---- ~~ -
- ~*-h--+----I-*-
- - - - --
-. -- r- em
0
h.
-i--- v.
-A----
-
.- - -- ------4--
-- -4 'mm
6 -- - I
................................................................................-.
e __~.~ .~. w
. -J
- - - .~. . ~ Ba.
2
I ___
- - -t .4..~. -~ - -.
- - ... 4......................----.- hi
I
-ma
.~ I .. 4
* a -
I I
S. 5 5 .4 -u ~
410 03.
:1.
-3 .51.0
.44 A
0 1.0 0w
l 0.4 fO/A06a LO
CM CK c, ,T0. .
.4.
-1.
4 .1.4. 0
5.0 0
Os 96 0T 10
M1 14-0 AVERAGE LOAD AM6 YIELD LOAD INFLUENCE
FACTOR FOR NTL#U. PEAK TWAOILM PULSE
* - - .- - - - _ _ _ _
- Si.
I: -LO.0 _______
0_
-. 3 .7 .6 _ _ _ _
.8 . .4
0 1.0 00 *04 .
0~~~I 6L -. -.
cm Cot 04 #I/ T
w
0
I-
U
IL
U)
-M
0)
IW
L
0)
.j9 n
IRI
a.
Department of Defense
Other Interested Government Activities
GEINRAL
ARI-I
NAVY
TI'H
R GOVE Z AGENCIE S
Dr. L. Fox
TASK VI PROJECT - C. E. RESEARCH STAFF Mathematics Division
National Physical Laboratory
W. H. Munse Teddington, Middlesex
England (1)
J. E. Stallmeyer
Professor W. J. Duncan, Head
R. J. Mosborg Dept. of Aerodynamics
College of Aeronautics
W. J. Austin Cranfield, Bletchley
Backs, England (1)
L. E. Goodman
Professor R. V. Southwell
C. P. Siess Imperial College of Science
and Technology
Research Assistants (10) Prince Consort Road
South Kensington
Files (5) London S.W. 7, England (1)
Reserve (20)
AV