Movement Equations 5 Dynamics of Sets of Solids

Movement Equations 5
Non-deformable Solid Mechanics Set

coordinated by
Abdelkhalak El Hami
Volume 5
Movement Equations 5
Dynamics of a Set of Solids
Georges Vénizélos
Abdelkhalak El Hami
First published 2019 in Great Britain and the United States by ISTE Ltd and John Wiley & Sons, Inc.
Apart from any fair dealing for the purposes of research or private study, or criticism or review, as
permitted under the Copyright, Designs and Patents Act 1988, this publication may only be reproduced,
stored or transmitted, in any form or by any means, with the prior permission in writing of the publishers,
or in the case of reprographic reproduction in accordance with the terms and licenses issued by the
CLA. Enquiries concerning reproduction outside these terms should be sent to the publishers at the
undermentioned address:
ISTE Ltd John Wiley & Sons, Inc.

27-37 St George’s Road 111 River Street
London SW19 4EU Hoboken, NJ 07030
UK USA
www.iste.co.uk www.wiley.com
© ISTE Ltd 2019

The rights of Georges Vénizélos and Abdelkhalak El Hami to be identified as the authors of this work
have been asserted by them in accordance with the Copyright, Designs and Patents Act 1988.
Library of Congress Control Number: 2019946982
British Library Cataloguing-in-Publication Data

A CIP record for this book is available from the British Library
ISBN 978-1-78630-036-2
Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
Table of Notations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
Chapter 1. Set of Solids with Neither

Loops Nor Branches. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1. Identifying a chain of solids with neither loops nor branches . . 1
1.2. Applying the fundamental principles of mechanics . . . . . . . . 2
1.2.1. Principle of effort generators . . . . . . . . . . . . . . . . . . . 3
1.2.2. Principle of effort receivers . . . . . . . . . . . . . . . . . . . 4
1.2.3. Applying the fundamental principle of dynamics. . . . . . . 4
1.2.4. Theorem of mutual actions . . . . . . . . . . . . . . . . . . . . 7
1.2.5. Summary of equations obtained . . . . . . . . . . . . . . . . . 8
1.3. Study of the movement of a chain of solids
(case of three solids) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.3.1. Applying the fundamental principle of dynamics. . . . . . . 8
1.3.2. Solidifying parameters . . . . . . . . . . . . . . . . . . . . . . 9
1.3.3. Movement equations . . . . . . . . . . . . . . . . . . . . . . . 11
1.3.4. Determining the link unknowns . . . . . . . . . . . . . . . . . 18
1.4. Links between solids . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.4.1. Link associated with the point contact of two solids . . . . . 18
1.4.2. Link torsor associated with the line contact of two solids . . 25
1.4.3. Link torsor associated with the surface
contact of two solids . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
1.4.4. Fundamental links between two solids in contact . . . . . . 32
vi Movement Equations 5
Chapter 2. Vibration Mechanics of Systems of Solids . . . . . 35

2.1. Movement equations of a set of solids . . . . . . . . . . . . . . . 35
2.1.1. Configuring and situating a set
of solids in a Galilean frame . . . . . . . . . . . . . . . . . . . . . . 35
2.1.2. Velocity distributors of n solids . . . . . . . . . . . . . . . . 37
2.1.3. Torsors associated with loads and efforts . . . . . . . . . . . 38
2.1.4. General equation of dynamics derived
from the fundamental principle . . . . . . . . . . . . . . . . . . . . . 39
2.1.5. Applying analytical mechanics of movement . . . . . . . . 39
2.2. Linear oscillatory systems with n solids . . . . . . . . . . . . . . 42
2.2.1. Setting the problem as an equation . . . . . . . . . . . . . . . 42
2.2.2. Equilibrium of a set of n solids . . . . . . . . . . . . . . . . . 46
2.2.3. Oscillations of a set of n solids . . . . . . . . . . . . . . . . . 47
2.2.4. Vibration eigen modes of a set of n solids . . . . . . . . . . 48
2.2.5. Influence of the initial conditions of the problem . . . . . . 53
2.3. Studying the vibrations of a continuous set
by passing to the limit . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
2.3.1. Taking the boundary conditions
into account at any instant . . . . . . . . . . . . . . . . . . . . . . . . 59
2.4. Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
2.4.1. Exercise 1: movement equations – equilibrium . . . . . . . 62
2.4.2. Exercise 2: movement around an equilibrium position . . . 75
2.4.3. Exercise 3: dynamics of an RTT robot
(one rotation + two translations) . . . . . . . . . . . . . . . . . . . . 85
Chapter 3. Vibrations with N Degrees of Freedom . . . . . . . 97

3.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
3.2. Homogeneous system – free vibrations (f1 = f2 = 0) . . . . . . . 99
3.2.1. Without damping (cij = 0) . . . . . . . . . . . . . . . . . . . . 99
3.2.2. Solving the system (Σ) . . . . . . . . . . . . . . . . . . . . . . 100
3.2.3. Damped free system . . . . . . . . . . . . . . . . . . . . . . . 108
3.3. Response on the time domain of an excited system . . . . . . . 112
3.4. Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
3.4.1. Exercise 1: eigen modes of a system with 2 DOF . . . . . . 113
3.4.2. Exercise 2: free and forced oscillations
of a conservative 2-DOF system . . . . . . . . . . . . . . . . . . . . 118
3.4.3. Exercise 3: calculation/test correlation . . . . . . . . . . . . 125
3.4.4. Exercise 4: damped system with a single excited mode . . 128
3.4.5. Exercise 5: system excited by the base . . . . . . . . . . . . 137
Contents vii
Chapter 4. Modal Analysis of N Degrees of Freedom . . . . . . 145

4.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
4.1.1. Normal modes . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
4.2. Response in the frequency domain of a
conservative structure subjected to a harmonic excitation . . . . . . . 146
4.3. Response of a structure with proportional
viscous damping to a harmonic excitation . . . . . . . . . . . . . . . . 150
4.4. Frequency response of a structure with
proportional hysteretic damping . . . . . . . . . . . . . . . . . . . . . . 153
4.5. Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
4.5.1. Exercise 1: receptance matrix of a conservative structure . . 155
4.5.2. Exercise 2: receptance matrix of a structure
with proportional viscous damping . . . . . . . . . . . . . . . . . . . 164
4.5.3. Exercise 3: case of a non-diagonal mass matrix . . . . . . . 170
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
Preface
This book is the fifth and last volume in the Non-deformable Solid
Mechanics series. As with the first four volumes, its content is
inspired by the mechanics teachings offered at the Conservatoire
National des Arts et Métiers (CNAM) in France.
The authors, Georges Venizelos in Paris and Abdelkhalak El Hami

in Rouen, took part in CNAM engineering training in their continuing
education or apprenticeship training. They wish to pay tribute to their
colleague, Michel Borel, equally passionate in teaching mechanics,
who participated in creating the first four volumes but who passed
away, alas, too early, before the creation of this fifth volume.
This book, Movement Equations 5, focuses on the dynamics of sets

of non-deformable solids, thus completing this series. In this volume,
appropriate mathematical tools (torsor calculus and matrix
calculation) are used to obtain the movement equations of a chain of
solids and solve them in order to obtain the information necessary for
the design of mechanical systems.
Chapter 1 presents the direct application of the fundamental

principle of dynamics to the movement of a chain of solids, as
developed in the previous volumes of this series, assuming that each
solid is considered non-deformable. Yet, when these solids are
presented in the form of loops or branches, these are generally subject
x Movement Equations 5
to structural strains like tensile, compression, bending or torsion

during movement.
Having applied the fundamental principle of dynamics to the

movement of a chain of solids, and obtained the movement equations
of a linear system of solids, in Chapter 2 we study the vibration
behavior of a system of n mass-spring-dampers. By passing to the
limit, we describe the study of the vibrations of a continuous
(deformable) system.
Chapter 3 presents the study of the vibrations of a rigid solid, those

of a set of several solids and those of a deformable solid. It is based on
knowledge of variations with respect to time, or frequency, of several
parameters. The variations in these parameters with respect to time
tend to be coupled. The movement equations of the mechanical set
considered then constitute a second-order differential equation system.
The vibratory behavior of a deformable structure or of a set of

solids connected by elastic links depends on the mechanical
characteristics of the system and the amplitude and frequency of
excitation.
In Chapter 4, we look to obtain the response of an excited system

with respect to the excitation frequency.
Georges VÉNIZÉLOS
Abdelkhalak EL HAMI
August 2019
Table of Notations
M material point
t time
m( S ) mass of a solid ( S )
δ ij Kronecker symbol
ε ijk alternate symbol of 3rd order

V vector
  
( λ ) ≡ ( xλ yλ zλ ) basis
  
λ ≡ Oλ xλ yλ zλ frame
ψ ,θ ,ϕ Euler angles, respectively precession,

nutation and natural rotation
   
Π ( V ,W ) plane of both vectors V and W
 
  plane of both vectors V and W
Π ( O V ,W )
passing through point O

  angle of the two vectors, oriented
V ,W  
from V towards W
 
V norm of vector V
xii Movement Equations 5

  scalar product of the two vectors V
V ⋅W 
and W

  vector product of the two vectors V
V ∧W 
and W
 polar unit vector in cylindro-polar
u(α )
coordinates
 polar unit vector in spherical
k (α,β )
coordinates
 vector rotation of angle α around
Ru , α ≡  u α  
the axis defined by the vector u

OM vectorial bipoint or vector
 situation vector of the point O S in
Oλ O S
the reference frame λ
  velocity at instant t of the material

v( λ ) ( M t ) or v( λ ) ( M ) point M throughout its motion in
the frame λ
  acceleration at instant t of the

J ( λ ) ( M t ) or J ( λ ) ( M ) material point M throughout its
motion in the frame λ
 rotation vector or rotation rate of

ωS λ solid ( S ) in its motion in relation
to the frame λ
 drive velocity of material point M

ve M( λ
μ ) in the relative motion of the frame
μ in relation to the frame λ
drive acceleration of the material
 point M in the relative motion of
Je M ( λ
μ ) the frame μ in relation to the
frame λ
Table of Notations xiii
Coriolis acceleration applied to the


Jc M( λ
μ ) material point M during the
relative motion of the frame μ in
relation to the frame λ
d ( λ )  derivative in relation to time of the


V vector V in the frame λ
dt
  torsor characterized by its two

{T }P =  s {T } M P {T } 
reduction elements at point P
  scalar invariant of torsor {T } ,

I = s {T }⋅M P {T }
independent of point P
{T1 } ⊗ {T 2 } product of two torsors
velocity-distributing torsor or
  kinematic torsor associated with the
{ S }P
λ
=  ωSλ v( λ ) ( P ) 
  motion of material point P of solid
( S ) in the frame λ
kinetic torsor associated with the
{ pS }
λ motion of solid ( S ) in the frame
λ
dynamic torsor associated with the
{A S } λ motion of solid ( S ) in the frame
λ
inertial operator of the solid ( S )
I OS ( S m ) supplied with the measure of mass
m expressed in a frame joined to
the solid
inertial drive torsor of the solid
{As e g
λ } ( S ) in the relative motion of λ
in relation to g
xiv Movement Equations 5
inertial Coriolis torsor of the solid

{As c λg } ( S ) in the relative motion of λ
in relation to g
Q1 , , Q6 canonical situation parameters
{Δ} torsor of known efforts
torsor of gravitational efforts acting
{π → S } on the solid ( S )
acceleration of Earth’s gravity
g  9.80665 m.s −2 ( 9.81 on average )
according to the place and altitude
of the body which is subject to it
torsor of link applied to the solid
{L → S }
(S)
power developed by the set of
(g)
P (F → S ) forces F acting on the solid ( S )
throughout its motion
partial power, relative to the
variable Qα , developed by the set
Πα( g ) ( F → S )
of forces F acting on the solid ( S )
throughout its motion
kinetic energy of the solid ( S ) in
T(λ) ( S ) its motion in relation to the frame
λ
Lagrange equation relative to the
( Lα ) variable Qα
When the situation of the solid ( S ) in the frame λ is
represented by the parameters Qα , we state : { λS } = { Sλ,α } Qα ′
where
Table of Notations xv
 
{ S ,α }O
λ
s
=  αδ ( λ , S ) α d O S  partial distributing torsor relative to
 
the variable Qα
partial rotation rate relative to the

 variable Qα , component of the
α
δ ( λ, S ) variable Qα ′ in the rotation rate
 
ωSλ = αδ ( λ , S ) Qα ′
component of the variable Qα ′ of

 the velocity vector of point O S ,
α
dOS
expressed under the form
 
v( λ ) ( O S ) = α d O Qα ′
S
 canonical linear component of the

sα dynamic resultant
 canonical quadratic component of
sαβ dynamic resultant
 canonical linear component of
mQα dynamic moment at Q
 canonical quadratic components of
mQαβ dynamic moment at Q
Eαβ canonical quadratic component of

kinetic energy

sΔ resultant of known exterior efforts
 part of the resultant of known

s Δ∗ exterior efforts independent of time
 moment at Q of exterior known
M QΔ
efforts
 part of moment at Q of known
M QΔ∗
exterior efforts independent of time
xvi Movement Equations 5
 part of the resultant of known

σ (t ) exterior efforts dependent
exclusively of time
 part of the moment of known

μ(t ) exterior efforts exclusively
dependent on time
part of the partial power of the
ΠαΔ∗ known exterior efforts relative to
Qα independent of time
part of the partial power of the
ΠαΔt known exterior efforts relative to
Qα dependent exclusively of time
values at the equilibrium of
( e ) = ( e1 , , e6 ) situation parameters Q1 , , Q6
development limited to the first
D1 [ f ( t ) ]
order of the function f ( t )
infinitesimal variation at first order
ε α ,ε α′ ,ε α′′ of the situation parameter Qα and its
first and second derivative
aαβ , bαβ , cαβ , dα ( t ) algebraic components of the

equations of small movements
torsor components of the equation
{ Aα( e ) } ,{ Bα( e ) } ,{ Cα( e ) } ,{ Δt } of small movements
functional components of the
( A ),( B ),( C )
equation of small movements
matrix components of the equation
[ a ],[ b ],[ c ]
of small movements
  projection basis of the vector
( E1  Ek ) expression of a linear differential
system
Table of Notations xvii
 couple of unknowns that make up

s, L the periodic solution of a
homogeneous differential system
impedance of a system of linear
Z(s) differential equations with constant
coefficients
M (s) inverse of the impedance Z ( s )
Laplace transform of the general
E ( s ) = L[ ε ( t ) ] solution to the small movements of
an oscillator
D ( s ) = L[ d ( t ) ] Laplace transform of the forced

excitation imposed on an oscillator
F ( s ) = L[ f ( t ) ] Laplace transform of the forced

oscillation of the oscillator
phase advance of an exciter signal
φ on the forced response of the
oscillator
amplification factor of the forced
A response in relation to the exciter
signal
pulsation of the oscillating motion
ω or ϖ
of the oscillator
rotation rate of the gyroscopic
Ω
device
ΩT Earth’s rotation rate
1
Set of Solids with Neither

Loops Nor Branches
The direct application of the fundamental principle of dynamics to

the movement of a chain of solids, as developed in the previous
volumes of this series, assumes that each solid is considered as
non-deformable. Yet, when the arrangements of these solids form
loops or branches, during movement these are generally subject to
structural deformation like tension, compression, bending, torsion, etc.
The study of such movements requires recourse to theories on
mechanical structures, reaching beyond the scope of this series.
1.1. Identifying a chain of solids with neither loops nor

branches
In a Galilean frame  g  , we consider a series of n solids (S1), (S2),

…(Si), …(Sn) linked to one another, forming a chain with neither
loops nor branches if
(S1) is linked to  g  and (S2) only;
(S2) is linked to (S1) and (S3) only;
...
Movement Equations 5: Dynamics of a Set of Solids,

First Edition. Georges Vénizélos and Abdelkhalak El Hami.
© ISTE Ltd 2019. Published by ISTE Ltd and John Wiley & Sons, Inc.
2 Movement Equations 5
(Si) is linked to (Si-1) and (Si +1) only;
...
(Sn) is linked to (Sn-1) only.
Let:
– “upstream of (Si)” refer to the subset of solids (Sj) such that
1≤ j≤ i ;
– “downstream of (Si)” refer to the subset of solids (Sj) such that

i ≤ j≤ n ;
– “(Sj-1)” refer to the previous solid (Sj);

– “(Sj+1)” refer to the next solid (Sj).
The solid (S1), with which the frame 1 is associated, is located
with respect to the reference frame  g  using k1(k1 < 6) parameters
1 2 k
q , q ,..., q 1 forming the group ( q1 ) ≡ q1,q 2 ,...,q k1  .
1 1 1 1 1 1 
Each of the solids in the chain is framed and located analogously

with respect to that preceding it. Thus, the solid (Si), with which the
frame i is associated, is located with respect to solid (Si-1), with
i − 1
which the frame is associated, using the ki (ki < 6) parameters
forming the group ( qi ) ≡ q1i ,qi2 ,...,qik1  .
 
1.2. Applying the fundamental principles of mechanics
The fundamental principles governing mechanics take on

particular meaning and importance in the case of chains of solids and
they are worth explaining and commenting on. The case of a set (D)
composed of two solids (S1) and (S2) articulated to each other at a
common point provides an appropriate illustration suitable for
highlighting their application.
Set of Solids with Neither Loops Nor Branches 3
Figure 1.1. Set of two disjoint solids
Let us begin by recalling the two basic principles that are applied
when studying chains of solids.
1.2.1. Principle of effort generators
Two effort generators and , considered disjoint, i.e. such that

, act jointly on the effort receiver according to the
diagram shown in Figure 1.2.
Figure 1.2. Effort generation diagram
The efforts exerted by on are represented by the torsor

, those exerted by on are represented by the
torsor . and those exerted by uniting two effort generators
and on the same effort receiver are represented by the torsor
such that:
1.2.2. Principle of effort receivers
The effort generator acts simultaneously on the two effort receivers

and , disjoint according to the diagram shown in
Figure 1.3.
Figure 1.3. Effort reception diagrams
The efforts exerted by on are represented by the torsor

, those exerted by on are represented by the torsor
and those that exerts jointly on the two effort receivers
and are represented by the torsor such that:
1.2.3. Applying the fundamental principle of dynamics
In the Galilean frame , consider the set (D) = (S1 (S2) To

simplify the expression of the torsor equations that result from
applying the fundamental principle of dynamics to the different
moving bodies that may be considered, the two solids will be noted 1
and 2, respectively. These equations also involve the external
()
environment of these bodies: D represents the external environment
of ( D) , that is, everything within its field of evolution, with the

( )
exception of those constituting D , strictly excluding (S ) and (S2 ) ,
1
which are also disjoint.
We must also consider that the reduction elements of a dynamic

torsor are summations achieved using the particles M of a solid, which
makes it possible to write, insofar as the two solids are disjoint,
= +
1.2.3.1. Applying the fundamental principle to solid 1

As the frame  g  is Galilean, the fundamental principle of
dynamics can be applied to the movement of solid 1 with the equation
= {1 → 1}
Note that 1 is the outside of 1, composed of all that is not , and

we have
1= ⋃2
hence, according to the principle of effort generators
{1 → 1} = { ⋃2 → 1} = { → 1} + {2 → 1}
The fundamental principle of dynamics applied to the movement of

( ) is therefore written as
= {1 → 1} = { → 1} + {2 → 1}
By separating the actions expressed by each of the two torsors into

known efforts and links, we obtain an exploitable expression of the
fundamental principle of dynamics applied to the movement of
= {∆ → } + {∆ → } + {ℒ → } + {ℒ → }
1.2.3.2. Applying the fundamental principle to solid 2

As the frame  g  is Galilean, the fundamental principle of
dynamics can be applied to the movement of solid 2 with the equation
= {2 → 2}
Note that 2, is the outside of 2, composed of all that is not ( ),

and we have 2 = ⋃1
Hence, according to the principle of effort generators
{2 → 2} = { ⋃1 → 2} = { → 2} + {1 → 2}
The fundamental principle of dynamics applied to the movement of

( ) is therefore written as
= {2 → 2} = { → 2} + {1 → 2}
Thus, by separating each of the two torsors according to the known

efforts and the links, we obtain:
= {∆ → } + {∆ → } + {ℒ → } + {ℒ → }
1.2.3.3. Fundamental principle applied to the set (D)

As it is assumed, in mechanics, that effort torsors may be separated
according to the physical laws (gravitation, electromagnetism and
contact forces), to which a mechanical set is subjected, the following
expression expresses in short the application of the fundamental
principle of dynamics to the movement of (D) in the Galilean frame
g
={ → },
where = 1⋃2.
According to the principle of effort receivers
{ → }={ → 1⋃2} = { → 1} + { → 2}
⟹ ={ → 1} + { → 2}
that is, after separating into known efforts and links:
= {∆ → } + {∆ → } + {ℒ → } + {ℒ → }
1.2.4. Theorem of mutual actions
The sum of the two equations obtained by application of the

fundamental principle of dynamics to each solid ( ) and ( ),
respectively, gives us:
+ ={ → 1} + {2 → 1} + { → 2} + {1 → 2}
Consequently
= + ={ → 1} + {2 → 1} + { → 2} + {1 → 2}
With, in addition:
={ → 1} + { → 2}
The combination of these different equations results in the

expression of the theorem of mutual actions that two solids exert on
one another
{2 → 1} + {2 → 1} = {0}
If the separation into known efforts and links, conducted

previously, is applied to this relation, this expression is then written as
{∆ → } + {∆ → } + {ℒ → } + {ℒ → } = {0}
In principle, whatever the physical law involved in the given or

calculable actions exerted by solid ( ) on solid ( ), we can write:
{∆ → } + {∆ → } = {0}
We therefore deduce reciprocity in the link efforts that are exerted

between the two solids
{ℒ → } + {ℒ → } = {0}
1.2.5. Summary of equations obtained
For the set ( ) = ( )⋃( ), to recap:

– two independent torsor equations, i.e. 12 scalar equations;
– three independent link torsors, unknown, {ℒ → }, {ℒ → },
{ℒ → }, representing 18 scalar unknowns;
– p location parameters.
The 12 scalar equations identified above include 18+p unknowns.

As a result, in a system corresponding to a set of two solids, there is a
deficit of
18+p-12=6+p equations.
In the specific scenario where one of the three link torsors is null,
the deficit is reduced to p equations.
1.3. Study of the movement of a chain of solids (case of

three solids)
1.3.1. Applying the fundamental principle of dynamics
In order to illustrate the principles that apply to the study of the

movement of several solids interconnected by means of an articulated
device according to the rule presented above, we will explore further
the case of three solids constituting a chain comprising neither loops
nor branches, therefore meeting the following conditions:
( )⋂( ) = ∅
( ) = ( )⋃( )⋃( ) where ( )⋂( ) = ∅
( )⋂( ) = ∅
The torsor equations translating the movement of each solid in the

pseudo-Galilean frame < > are written as
∗( )⇒ = {1 → 1} = { → 1} + {2 → 1}
∗( )⇒ = {2 → 2} = { → 2} + {1 → 2} + {3 → 2}
∗( )⇒ = {3 → 3} = { → 3} + {2 → 3}
The three solids are disjoint: given the composition of the chain
and applying what was shown in section 1.2, according to the theorem
of mutual actions, we can write that
{2 → 1} + {1 → 2} = {0}
{3 → 2} + {2 → 3} = {0}
We can furthermore deduce the torsor equation of the fundamental

principle of dynamics by applying it to the mechanical set ( )
= + + ={ → 1} + { → 2} + { → 3}
The external and the link efforts acting on each solid come under
the various laws of forces already mentioned, making it possible to
separate the right-hand terms of the torsional equations according to
these laws, applying the fundamental principle of dynamics to each
solid, giving:
= {∆ → } + {ℒ → } − {∆ → } − {ℒ → }
= {∆ → } + {∆ → } + {ℒ → } − {∆ → } − {ℒ → }
= {∆ → } + {∆ → } + {ℒ → }
where the link of ( ) with ( ) leads to the link with < >.
1.3.2. Solidifying parameters
In cases where links are independent of time, the velocity

distributing torsors of the movement of ( ) with respect to < >, of
( ) with respect to < 1 > and of ( ) with respect to < 2 > are
expressed, respectively, as
= ( ), = 1, … , ∈( )
1 2.
1 1
= ( ), = 1, … , ∈( )
2 2.
2 2
= ( ), = 1, … , ∈( )
3 3.
where ( ), ( ) and ( ) are without the parameter groups

introduced in section 1.1 to identify the solids of a chain. It should be
noted in particular that:
1 1
= + = +
2 2 1 1. 2.
As the parameters are not involved in the movement of ( )

with respect to < >, we have = {0}
1.
1 1
⇒ = + =
2. 2. 1. 2.
Likewise, the parameters are not involved in the movement of

1
( ) with respect to < 1 >, and as a result = {0}
2.
1
⇒ = + =
2. 2. 1. 1.
hence
= +
2 2. 2.
The same approach allows us to demonstrate that the relation
2 1
= + +
3 3 2 1
2 1
= + +
3. 2. 1.
can be written more generally as we are faced with a chain of solids

with neither loops nor branches.
= + +
3 3. 3. 3.
We can observe that the parameters are involved in the

movements of ( ), ( ) and ( ) with respect to < >; indeed, if we
block parameters and such that ( ) and ( ) are fixed with
respect to ( ), the only variation in the parameters causes the
location of the set (D) to evolve as if it were a single solid. We can
therefore say that the parameters are solidifying for the set
( ) = ( )⋃( )⋃( )
Likewise, the parameters are involved in the movements of

( ) and ( ) with respect to < > . If we block parameters as
well as , only the variation in parameters can be witnessed
with respect to < > or < 1 >, in the movement of set ( )⋃( ),
which behaves like a single solid. We can therefore say that the
parameters are solidifying for the set ( )⋃( ).
1.3.3. Movement equations
Applying the fundamental principle of dynamics to the movements

of the three solids, ( ), ( ) and ( ) with respect to < > results in
6 × 3 = 18 equations, consequences of the three torsor relations
introduced at the end of section 1.3.1.
Yet, these scalar relations contain link components in their

right-hand terms. Moreover, they do not express the energy dimension
of the movements, and as a result we cannot directly take advantage of
this aspect of the movements studied such as, for example, the energy
formulation of perfect links to obtain movement equations for the set
(D).
We therefore turn to an analytic expression of this movement, and

since we obtain as many analytical mechanics equations as there are
situation parameters to express each of the three solids with respect
to < >, that is to say, + + equations, they should be
chosen carefully, given the solidifying nature of a number of these
parameters.
1.3.3.1. First set of equations

The analytical mechanics equations relating to the location
parameters of the three solids ( ), ( ) and ( ) result from the
following three torsor relations:
( ): =⋯
1.
… ⨂[{∆ → } + {ℒ → } − {∆ → } − {ℒ → }]
1.
( ): =⋯
2.
… ⨂[{∆ → } + {∆ → } + {ℒ → } − {∆ → } − {ℒ → }]
2.
( ): =⋯
3.
… ⨂[{∆ → } + {∆ → } + {ℒ → }]
3.
And as these parameters are solidifying for the set
( ) = ( )⋃( )⋃( ) with respect to < >, we have
= =
1. 2. 3.
⇒ ⨂ + + = =⋯
1. 1.
.... ⨂[{∆ → } + {ℒ → } + {∆ → } + {∆ → }]
1.
relation that can be expressed in terms of the kinetic energy in the

form
( ) ( ) ( ) ( )
− + − …
( ) ( )
…+ − …
( )
…= ∏ (∆ → ) + ∏ (ℒ → ) + ∏ (∆ → ) + ∏ (∆ → )
As the three solids are disjoint and the kinetic energy is a

summation for the set of particles of the field considered, we can also
write
( )= ( )+ ( )+ ( )
and for each parameter , we obtain one of the equations.
( ) ( )
− =∏ ( → )
( )
where the only link that appears is that of ( ) with < >. If this link
is perfect, that is, if ∏ (ℒ → ) = 0, the equations above constitute
a first set of movement equations.
1.3.3.2. Second set of equations

parameters of the two solids ( ) and ( ) result from the
following three torsor relations:
( ): =⋯
1.
… ⨂[{∆ → } + {ℒ → } − {∆ → } − {ℒ → }]
1.
λ
(S ): =⋯
2. α
λ
… ⨂[{∆ → } + {∆ → } + {ℒ → } − {∆ → } − {ℒ → }]
2. α
λ
(S ): =⋯
3. α
λ
… ⨂[{∆ → } + {∆ → } + {ℒ → }]
3. α
where
= {0} and =
1. 2. 3.
making it possible to write:
⨂ + = ⋃ =⋯
2. 2.
.... ⨂[{∆ → } + {∆ → } + {ℒ → } + {∆ → }]
2.
or, in kinetic energy terms:
( ) ( ) ( ) ( )
− + − …
( )
…= ∏ (∆ → ) + ∏ (∆ → )+ {∆ → }+
2.
{ℒ → }
2.
By developing the latter relation, we obtain the following results:
1.3.3.2.1. Interpreting left-hand terms

As ( ) is only a function of the parameters , we therefore
have
( ) ( )
= 0 and =0
The left-hand term of the equation in question is thus written as
( ) ( )
−
1.3.3.2.2. Interpreting right-hand terms

Consider the terms of the expression
{∆ → }+ {ℒ → }
2. 2.
The efforts expressed by torsors {∆ → } and {ℒ → } are only

exerted in the relative movements of the solids ( ) and ( ). They
only put a power in play in these movements.
( ) (∆
→ ) = 1 {∆ → }
2.
1
= {∆ → } =∏ (∆ → )
2.
1
As = {0} and =
1. 2. 2.
we obtain
(∆ → )= {∆ → }= (∆ → )
2.
This result is only applicable because the chain of solids has

neither loops nor branches. We obtain a similar result for the link
torsor {ℒ → }.
The second set of equations, relating to the parameters , is then

written as
( ) ( )
− …
…= ∏ (∆ → ) + ∏ (∆ → ) + ∏ (∆ → ) + ∏ (ℒ → )
We thus have equations, in which the only link involved is that

between ( ) and ( ). When this link is perfect, in other words when
(ℒ → )= (ℒ → )=0
we then obtain movement equations.
1.3.3.3. Third set of equations

parameters of the solid ( ) result from the following three torsor
relations:
( ): =⋯
1.
… ⨂[{∆ → } + {ℒ → } − {∆ → } − {ℒ → }]
1.
( ): =⋯
2.
… ⨂[{∆ → } + {∆ → } + {ℒ → } − {∆ → } − {ℒ → }]
2.
( ): =⋯
3.
… ⨂[{∆ → } + {∆ → } + {ℒ → }]
3.
where:
= = {0}
1. 2.
making it possible to write:
( ) ( )
− …
…= ∏ (∆ → )+ {∆ → }+ {ℒ → }
3. 3.
By developing the latter relation, we obtain the following results.
1.3.3.3.1. Interpreting the left-hand term

As ( ) and ( ) are not functions of the parameters , we
can write
( ) ( ) ( ) ( )
− = −
1.3.3.3.2. Interpreting right-hand terms
Consider the terms of the expression
{∆ → }+ {ℒ → }
3. 3.
The efforts expressed by torsors {∆ → } and {ℒ → } are only

exerted in the relative movements of solids ( ) and ( ). They only
put in play a power in these movements.
2
( ) (∆
→ ) = 2 {∆ → }= {∆ → }
3 3.
1
As = {0} and = {0}
2. 1.
2
{∆ → }= {∆ → } = ∏ (∆ → ) = ∏ (∆ → )
3. 3.
We obtain
2
{∆ → }= {∆ → }= (∆ → )= (∆ → )
3. 3.
The term {∆ → } is a power coefficient that has its place in

3.
an analytical mechanics equation, as well as {ℒ → }.
3.
We then obtain movement equations of the form
( ) ( )
− = ( → )
with the only link terms that they contain being those relating
to {ℒ → }. If the latter is perfect, that is if ∏ (ℒ → ) = 0, these
relations are counted among the movement equations.
1.3.4. Determining the link unknowns
In the case of all-perfect links, the + + analytical

mechanics equations form + + movement equations to
determine, as a function of time, the ( + + )
parameters , , .
The 18-( + + ) equations provided by the direct

application of the fundamental principle of dynamics associated with
perfect link conditions enable the link unknowns to be determined.
1.4. Links between solids
Consider two solids ( ), with surface contour Σ , to which the

frame < 1 > is attached, and ( ), with surface contour Σ , to which
the frame < 2 > is attached.
1.4.1. Link associated with the point contact of two solids
The solid ( ) is in intermittent contact with the single solid ( )

at a point M of its surface contour Σ .
Sett of Solids with Neither
N Loops Nor
N Branches 19
F
Figure 1.4. Po
oint contact of two solids
is the unitary vector norm

mal to the insstantaneous common plaane
Π at M tangent to thhe respectivee surface con
ntours Σ andd Σ of the tw
wo
solids ( ) and ( ), directeed from ( ) to ( ). The veloccity
distribuuting torsor of ( ) wiith respect to t ( ) has the followiing
reduction elements at M:
1
ℳ
2
involviing, by princiiple, six paraameters.
Durring the relative displaacement of the two soolids onto oone

anotherr, the point ofo contact iss mobile on each of them m. Under theese
1
conditions, if ℳ has the dimensions
d of
o a velocity,, this term onnly
2
expressses the instaantaneous sppeed with respect to ( ) of the pooint
of ( ), which coinncides, at thiss instant, witth the point of contact; thhis
vector therefore does not obey o the velocity
v com
mposition laaw
applicaable to a solidd.
t case where contact at M is perrsistent, which amounts to

In the
t componeent of ℳ 1
saying that at thee moment considered, the
2
followiing the insttantaneous normal
n is null, the additional
a liink
relationn provided byy this condition is
1
.ℳ =0
2
and redduces the num
mber of degrrees of freedo
om of ( ).
20
0 Movement Equations
E 5
The reducttion elementss at M of thee link torsor {ℒ } associatted with

thhe contact acttions exertedd by ( ) on ( ) are
{ℒ } = [ ⋮
annd introduce six scalar unnknowns.
Let us exaamine the quuestion of po

oint contact in the speciific case
whhere the fram
me < 1 > is Galilean. Th
he torsor equaation
{ } = { Δ} + { ℒ }
thhat results from

fr the application off the fundam
mental princciple of
dyynamics to thet movemeent of ( ) with respecct to ( ) giives six
sccalar equationns.
With a scaalar equationn characterizzing the linkk, we therefoore have

seeven equatioons for 12 unknowns, hence a deficit of fivee scalar
eqquations.
Moreover, if the link torsor

t has a null momennt at M, a coondition
thhat results in three additioonal scalar equations,
e the deficit is thhen just
twwo scalar equuations.
Let us introoduce a new

w condition to
t attempt too solve the pproblem,
exxpressed by the
t laws of dry
d friction.
Figure 1.5. Point contact

c at M
The resultant of the link efforts is projected in the form
= ( . ) + ⋀( ⋀ )
where:
– = ( . ) defines the normal contact action;

– = ⋀( ⋀ ) defines the tangential contact action.
The power developed by the link efforts is given by:
1 { } 1
ℒ = ⨂ ℒ = .ℳ
2 2
1
ℒ = [( . ) + ⋀( ⋀ )]. ℳ
2
1
where . ℳ = 0 as we consider the contact to be persistent given
2
that there is no displacement of point M along the normal, hence:
1 1
ℒ = [ ⋀( ⋀ )]. ℳ = .ℳ
2 2
The configuration of this type of point contact can present several
aspects, which are explained below.
1.4.1.1. Configuration 1
When the solid ( ) is not bonded to ( ), the relation
. >0
indicates that, with the two vectors forming an acute angle, the normal
contact action exerted by ( ) on ( ) is directed from ( )
toward ( ).
Cancellation of the normal contact action leads this contact at M to

be interrupted by unbonding.
When the sliding velocity of ( ) on ( ) at M is null, that is,
1
ℳ =0
2
the tangential component of the link efforts is small compared to the
normal component, which is reflected by the relation
<
where:
– f is the mutual-friction coefficient of ( ) and ( ), which

depends on the nature of the materials, their surface condition and
their temperature.
– =| . |= .
– = [ ⋀( ⋀ )] = ( ⋀ )
– . < ( ⋀ )
If we examine the summary of equations available to address the

1
question, introducing the relation ℳ = 0 provides the two
2
missing scalar equations to reduce the deficit to zero.
1
As ℒ = , the power developed with respect to the
.ℳ
2
Galilean frame < 1 > by the link associated with the contact actions
that ( ) exerts on ( ) is null. This relation makes it possible to
determine the contact action . This then involves verifying the two
inequalities
. >0
. < ( ⋀ )
and when the next equality is verified
. = ( ⋀ )
There is a possibility of sliding at M.
Let us now consider the case where
1
ℳ ≠0
2
a condition that leads to the three relations
(1) =
(2) 1
⋀ℳ =0
2
(3) 1
.ℳ <0
2
According to relation (2) above, the tangential contact action and
the sliding velocity at M of ( ) on ( ) are colinear. This relation is
written as
1 [
ℳ ⋀ ⋀( ⋀ )] = 0
2
1 ( 1
ℳ . ⋀ ) − ℳ . ( ⋀ )=0
2 2
1
As we have the condition ℳ . = , we deduce from the
2
preceding relation that
1 (
ℳ . ⋀ )=0
2
According to relation (3), the tangential contact action and the
sliding velocity at M of ( ) on ( ) are opposite. This relation is
written as
1
⋀( ⋀ ). ℳ <0
2
1
⇒ [ − ( . ) ]. ℳ <0
2
With the same condition as before, we deduce that
1
.ℳ <0
2
That is to say that the power developed with respect to the Galilean
frame < 1 > by the link torsor associated with the contact actions at
M of ( ) and ( ) is negative.
1
Thus, when we have ℳ ≠ 0, we have the three relations
2
. <
( ⋀ )
( ⋀ ). ℳ 1 = 0
2
1
.ℳ <0
2
The first two provide the two missing scalar equations to reduce
the equation deficit to zero. Having then determined the contact
action , it remains for the inequality . > 0 to be verified.
1.4.1.4. Perfect links

In the case of a perfect, frictionless link (f = 0), the tangential
component of the contact action is null. This action is then expressed
by
= where >0
The value 0 of the two components of the tangential contact action

then provides the two equations enabling the deficit to be reduced to
zero. Moreover, the power developed by the link is null.
N Loops Nor
N Branches 25
1.4.2. Link torso

or associated with th
he line contact of tw
wo
s
solids
1.4.2.1
1. Arbitrary line contactt
wo solids occurs along the
Thee contact bettween the tw t curve’s aarc
(C).
Figure 1.6. Arbitrary line contact
Associated withh point M off the contactt curve (C), whose vicinnity
on this curve (Figuure 1.6) is deefined by thee elementaryy arc ℓ( ),, is
the elem
mentary forcce
( ) ℓ( )
where ( ) is a linnear force deensity.
Thee link torsorr {ℒ } between the two solids in contact has tthe
followiing reductionn elements att Q:
{ℒ } = ( ) ℓ(( )
∈( )
ℳ {ℒ } = ⋀ ( ) ℓ( )
∈( )
Six scalar compponents of thhe link torso

or are, generrally speakinng,
unknowwn.
26
E 5
1..4.2.2. Straiight line con

ntact
Figure
e 1.7. Straightt line contact
The segmeent (C) of thee line contactt belongs to the

t common tangent
pllane ( | . ).
) The normaal penetration n velocity is null at any ppoint on
thhe contact segment. As a result, for f the two points, M and Q,
beelonging to thhe contact seegment, we can
c write
1 ( )( )+
.ℳ =0 ⇒ . ⋀ =0
2
1 ( )( )+
.ℳ =0 ⇒ . ⋀ =0
2
From the term-by-term
m difference for the last two
t expressiions, we
obbtain
. ⋀ =0
By writing = , the
t relation
. ⋀ =0
is satisfied regardless of the point M located on (C) if
. ⋀ =0 ⇒ ( ⋀ ). =0 ⇒ . =0
That is to say, the rotation rate vector of the solid ( ), with

respect to ( ), is collinear to the vector support of the contact line.
In the case of persistent straight line contact, the two relations
1
.ℳ =0
2
. =0
accord four degrees of freedom to the solid ( ).
The reduction elements at Q of the link torsor are given as
{ℒ } = ( )
∈( )
ℳ {ℒ } = ⋀ ( ) = ⋀ ( )
∈( ) ∈( )
As, according to the relation above, . ℳ {ℒ } = 0, the link torsor

involves only five unknowns, its reduction elements at Q can be
written as
{ℒ } = + +
ℳ {ℒ } = +
With the two link equations and L = 0, we have a system of nine

scalar equations for 12 unknowns, that is, a deficit of three equations.
In the case of a perfect link, we have:
( )= ( )
{ℒ } = ( )
∈( )
ℳ {ℒ } = − ( )
∈( )
The link torsor involves only two unknowns and its reduction
elements at Q are
{ℒ } =
ℳ {ℒ } =
With the three additional equations
= 0; = 0; =0
the deficit is reduced to zero.
1.4.3. Link torsor associated with the surface contact of

two solids
1.4.3.1. Arbitrary surface contact

The contact is made along the surface
Figure 1.8. Arbitrary surface contact

N Loops Nor
N Branches 29
Poinnt M, belongging to the contact

c surface (Σ), is chharacterized on
this surrface by an elementary
e s
surface area ( ), whicch specifies its
vicinityy. At this point,
p we associate an elementary
e f
force with tthe
efforts due to the coontact actionn between thee two solids
( ) ( )
where ( ) is a suurface force density.

d
Thee reduction ellements at Q of the link torsor

t are exppressed by
{ℒ } = ( ) ( )
∈( )
ℳ {ℒ } = ⋀ ( ) ( )
∈( )
and theeir componennts are generally unknow

wn.
1.4.3.2
2. Plane surrface contacct
Figure 1.9. Plane

P surface contact
Thee plane surfacce (Σ) belonngs to the tan

ngent plane common
c to tthe
two sollids in contacct.
Witth the normal penetrationn velocity beeing null at any

a point in tthe
contactt area, for thee two points,, M and Q, we
w have the tw
wo relations::
1
.ℳ =0
2
1
.ℳ =0
2
which are summarized in a single equation:
. ⋀ =0
By writing = + , we obtain, successively
. ⋀( + ) =0
. [( + )⋀ ] = 0
.( − )=0
and the latter relation is satisfied, whatever the point M belonging to

(Σ), if we simultaneously have
. =0
. =0
In other words, the rolling rate vector of ( ) on ( ) is null.
In the case of persistent plane surface contact, the three relations
1
.ℳ =0
2
. =0
. =0
result in three degrees of freedom for the movement of ( ).

The reduction elements of the link torsor at Q are
{ℒ } = ( , )
∈( )
ℳ {ℒ } = ( + )⋀ ( , )
∈( )
Six components of this torsor are unknown.
With the three link equations, we attain a system of nine scalar

equations for 12 unknowns, and consequently a deficit of three
equations.
In the case where the link is perfect, we have:
( , )= ( , )
{ℒ } = ( , )
∈( )
ℳ {ℒ } = ( , ) − ( , )
∈( ) ∈( )
The link torsor involves only three scalar unknowns and its
reduction elements at Q lead to:
{ℒ } =
ℳ {ℒ } = +
With the three additional equations
= 0; = 0; =0
the deficit is reduced to zero.

1.4.4. Fundamental links between two solids in contact
The reduction elements at Q of the link torsor {ℒ } associated with

the actions of ( ) on ( ) have the following general expressions:
{ℒ } = + +
ℳ {ℒ } = + +
1.4.4.1. Pivot linkage with axis ( | )

This link only allows one degree of freedom. The velocity
distributor
is accompanied by the partial distributor
If the link is perfect, we have
=0
and the link torsor displays five unknowns.
1.4.4.2. Sliding linkage with axis ( | )

distributor
1
= 0⋮ ′
2
is accompanied by the partial distributor
1
= 0⋮
2.

=0
and the link torsor displays five unknowns.
1.4.4.3. Bolt linkage (sliding-pivot) with axis

This link only allows two degrees of freedom. The velocity
distributor
1
=[ ⋮ ]
2
is accompanied by the two partial distributors

1 1
= ⋮0 ; = 0⋮
2. 2.

= 0 and =0
and the link torsor then displays four unknowns.
1.4.4.4. Helical linkage with axis ( | )

distributor
1
= ′ ⋮ ′
2 2
with the helix pitch only admits a single partial distributor

1
= ⋮ ;
2.
+ =0
and the link torsor depends on five unknowns.

1.4.4.5. Ball-joint linkage of center Q

This link only allows three degrees of freedom. The velocity
distributor
1
= + + ′ ⋮0
2
admits the three partial distributors
1 1 1
= ⋮0; = ⋮0; = ⋮0
2. 2. 2.
= = =0
and the link torsor depends on three unknowns.
1.4.4.6. Fixed-plane linkage with the plane

This link only allows three degrees of freedom. The velocity
distributor
1
=[ +⋮ ′ + ′ ]
2
is accompanied by the three partial distributors
1 1 1
= ⋮0 ; = ⋮ ; = 0⋮
2. 2. 2.
=0 and = =0
and the link torsor then depends on three unknowns.

2
Vibration Mechanics
of Systems of Solids
Having applied the fundamental principle of dynamics to the

movement of a chain of solids and obtained the movement equations
of a linear system of solids, we will now study the vibration behavior
of a system of n mass-springs-dampers.
By passing to the limit, we will describe the study of the vibrations

of a continuous (deformable) system.
2.1. Movement equations of a set of solids
2.1.1. Configuring and situating a set of solids in a Galilean

frame
In the Galilean frame 〈 〉 ≡ ⃗ ⃗ ⃗ , we consider n solids

( ), … , , … ( ) constituting a mechanical set:
( ) = ( )⋃ … ⋃ ⋃ … ⋃( )
We attach a frame to each of these, namely:
( ) ⇔ |⃗ ⃗ ⃗
...

⇔ ⃗ ⃗ ⃗
...
( ) ⇔ |⃗ ⃗ ⃗
These present the following inertia characteristics:

∀ solid = 1, …
– mass ;
– center of inertia ;
− −
– inertia matrix ( ) = − − .
/( ⃗ ⃗ ⃗)
− −
The links between the solids are all assumed to be configurable and
holonomic, which amounts to saying that the reduction elements of
each link torsor are expressed as a function of the only k situation
parameters of the set of solids in the Galilean frame 〈 〉; these links
are expressed as a function neither of time nor of the derivatives of the
k parameters.
If the degrees of freedom of in 〈 〉 are p in number, this

situation being represented by the parameters . , … … , . ; if
is not directly situated with respect to 〈 〉 but with respect to ( ),
with degrees of freedom and the situation parameters
. , … … , . , the number of degrees of freedom of in 〈 〉 is
= +
where r is the number of degrees of freedom of ( ) in 〈 〉.
Likewise, if ( ) is situated with respect to ( ) with degrees

of freedom and ( ) is situated with respect to 〈 〉 with degrees of
freedom, then is situated with respect to 〈 〉 with the ̅ = +
+ degrees of freedom, and so on.
Vibration Mechanics of Systems of Solids 37
When all of the links between the solids are thus identified, we see
the appearance of k parameters , … , … , to situate the chain of
n solids, with ≤ 6 .
2.1.2. Velocity distributors of n solids
The velocity distributor torsor of the movement in 〈 〉 of the solid
= ⃗⋮ ( ⃗)
involves the ̅ degrees of freedom of , hence the ̅ partial

distributing torsors, which generally are not zero; but, with respect to
the k degrees of freedom of the set, the other ( − ̅ ) partial
distributing torsors are null.
When is situated directly in 〈 〉, we obtain from the
partial distributing torsors , where = 1, … .
When is situated via the sequence of solids
( ), ( ) ,…, ( ), ( )
the velocity distributing torsor of its movement is written as
ℎ
= + +⋯ +
ℎ
and as , = , we obtain, for each partial distributing torsor
ℎ
, = , + , +⋯ + ℎ,
,
2.1.3. Torsors associated with loads and efforts
In order to be able to suitably develop the loads and efforts exerted

on a mechanical set ( ) = ( )⋃( )⋃ … ⋃ ⋃ … ⋃( ), the
outside of each of the solids needs to be identified
( ) = ( ⋃ ⋃… ⋃ ⋃. . . ⋃ )⋃ = ⋃
i.e.
{ → } = {ℒ(2 → 1)} + {ℒ( → 1)} + {ℒ( → 1)} + ⋯

+ {ℒ( → 1)} + {Δ(2 → 1)} + ⋯ + {Δ( → 1)}
+ ⋯ + {Δ( → 1)} + ⋯ + {Δ( → 1)}
This expression can be applied to all the solids of the set in the
following summary form:
→ = {ℒ( → )} + {Δ( → )}
+{ℒ( → )} + {Δ( → )}
where
{ℒ( → )} + {ℒ( → )} = {0}
{Δ( → )} + {Δ( → )} = {0}
In the most general case, we therefore witness the appearance of a

( ) ( )
total of unknown torsors, including link torsors {ℒ( →
)} with < and n other link torsors {ℒ( → )}.
However, in the case of a chain of solids with neither loops nor

branches, each solid is only in contact with the one preceding it and
the one that follows it; as a result, numerous link torsors of the
preceding formulas are null.
2.1.4. General equation of dynamics derived from the

fundamental principle
For each solid , the torsor expression of the fundamental

principle of dynamics gives
= {ℒ( → )} + {Δ( → )}
+{ℒ( → )} + {Δ( → )}
( )
There are 6n scalar equations for +6 unknowns, hence a
deficit of + 3 ( − 1) relations.
The previous n torsor equations can be replaced by n other

equations obtained from linear combinations of the first n in order to
assemble the link terms associated with the reciprocal actions of one
solid on another (applying the theorem of mutual actions).
2.1.5. Applying analytical mechanics of movement
To begin with, in order to simplify the writing of the mathematic

expressions to come, let us write
( )
( )=
2.1.5.1. Lagrange equations for each solid
⨂ = ⨂ { ℒ( → )} + ⨂ℒ → …
, , ,
≠
… ⨂∑ ≠ {Δ( → )}+ ⨂Δ →
, ,
Hence the Lagrange equation relating to the variable and

applicable to the solid ( )
( ) ( ) ( ) ( )
− = {ℒ( → )} + {ℒ( → )} …
( )
( ) ( )
…+ {Δ( → )} + {Δ( → )}
i.e., for the solid ( ), ̅ non-zero Lagrange equations, corresponding

to the ̅ partial distributors non-null.
2.1.5.2. Lagrange equations for the set ( )

As kinetic energy is expressed by a summation on all the particles
of a solid, we can write:
( ) = (1) + (2) + ⋯ + ( ) + ⋯ + ( ) = ( )
Hence the Lagrange equation relating to the variable and

applicable to the set ( ):
( ) ( )
−
( )
= ⨂ {ℒ( → )} …
,
+ ⨂ {ℒ( → )} …
,
+ ⨂ {Δ( → )} + ⨂{Δ( → )}
, ,
A number of the terms in the above expression require explanation.
( )
⨂{ℒ( → )} = {ℒ( → )}
,
( )
⨂{Δ( → )} = {Δ( → )}
,
Regarding the two terms
⨂ {ℒ( → )} ⨂ {Δ( → )}
, ,
Taking the first term as an example, we can see that the following
expression appears as
⨂{ℒ( → )} + ⨂{ℒ( → )} …
, ,
… = {ℒ( → )}⨂ − = {ℒ( → )}⨂ …

, , ,
= ⨂{ℒ( → )}
,
and we can then write:
⨂ {ℒ( → )} = ⨂{ℒ( → )}
, ,
The k Lagrange equations ( ) of the movement of the set ( ) in

〈 〉 are written as
( ) ( )
( )
− =
( ) ( )
∑ ∏ {ℒ( → )} + ∑ ∏ {Δ( → )} + …
…+∑ ∑ , ⨂{ℒ( → )}+…
…+∑ ∑ , ⨂{ ( → )} i = 1,…, k
When the links are perfect, the partial power terms of the link
efforts disappear; we then obtain equations where the only unknowns
are the k degrees of freedom of the set ( ). These are the movement
equations of this set.
2.1.5.3. Subsidiary energy–power equations

Written for each solid ( ) of the chain or for the set ( ) that it
represents, these equations of the form
( )
⨂ = =⋯
= ⨂ {ℒ( → )} + ⨂{ℒ( → )} …
+ ⨂ {Δ( → )} + ⨂{Δ( → )}
provide, particularly in the case where the links are perfect and
consequently, the powers that they put in play are null, subsidiary
movement equations, which can advantageously replace the Lagrange
or dynamic equations.
2.2. Linear oscillatory systems with n solids
Here, we will show, based on a particular application, how the

study of the equilibrium and oscillations of a set of n solids is
presented. To do this, we will consider the case of a chain of solids in
straight-line translation by means of prismatic links connected to each
other by n – 1 springs working in tension/compression.
2.2.1. Setting the problem as an equation
The set, guided by the ( ⋮ ⃗) axis, which remains horizontal, is

composed of two main components:
– identical solids ( ), with mass , center of inertia and
length 2a;
Vibration Mec
chanics of Syste
ems of Solids 43
Figure
e 2.1. Solid ( ), a compone
ent of the chaiin
– spprings of natural
n lengthh ℓ and stifffness .
Figure
e 2.2. Spring Ri , a compone
ent of the chaiin
Morreover, the set

s is conneccted at its ex
xtremities O and O’ to tthe
referennce frame < > by tw wo springs according too the diagraam
shown in Figure 2.33, with ⃗′ = ⃗, where L is a given, fixed lengthh.
Fig
gure 2.3. Osciillatory system
m of n solids
We assume thatt the prismattic links thatt assure the sliding

s of eaach
solid onn the horizonntal axis ( ⋮ ⃗) are perfeect.
We set: ⃗= ⃗
Thee velocity-disstributing torrsor of the motion

m of eacch solid has tthe
followiing expressioon:
= 0⃗ ⋮ ′⃗
and for kinetic energy we have:
2. =
The effort torsors acting on ( ) are, respectively,

– the gravity torsor
→ = ⃗ ⋮ 0⃗
where ⃗. ⃗ = 0. The power applied by the gravity torsor is null for the
movement considered ( ) → =0;
– the torsors of the actions of the two springs that frame each solid
ℓ
→ = 1− ⃗ ⋮ 0⃗ = →
where
⃗= ⃗+ ⃗+ ⃗+ ⃗
= − + + ⃗
and
= = − −2 because − >
2
⇒ → = − − −2 −ℓ ⃗ ⋮ 0⃗
ℓ
→ = 1− ⃗ ⋮ 0⃗
= →
where
⃗= ⃗+ ⃗+ ⃗+ ⃗
= − − + − ⃗
and
= = − −2 because − >2
⇒ → = − −2 −ℓ ⃗ ⋮ 0⃗ ;
– the prismatic link torsor that assures the sliding of the solids on
the ( ⋮ ⃗) axis and that is perfect
( )
ℒ→ = 0.
The Lagrange equation corresponding to the motion of the solid

( ) subject to the efforts summarized above is, therefore, given as
=− − −2 −ℓ …
…+ − −2 −ℓ
We will now examine the case of the two end springs of the set,
and .
ℓ
{ → } = 1− ⃗ ⋮ 0⃗ = { → }
⇒{ → } = − ( − − ℓ ) ⃗ ⋮ 0⃗
ℓ ⃗′ ⋮ 0⃗ = {
{ → } = 1− → }
= ( − − −ℓ ) ⃗ ⋮ 0⃗
The linear differential equation system expressing the motion of

the set of these n solids, here the movement equations of this set, is
written as
(1) = −( + ) + + ( −2 )
+ ℓ − ℓ
(2) = −( + ) + +2 ( − )
+ ℓ − ℓ
………………………………………………………………………
( )
= −( + )
+ +2 ( − )
+ ℓ − ℓ
……………………………………………………………………
( − 1)
= −( + )
+ +2 ( − )+ ℓ − ℓ
( )
= −( + )
+ + (2 − )
+ ℓ − ℓ
2.2.2. Equilibrium of a set of n solids
If the system is in a situation of equilibrium, it will be

characterized, for each solid ( ) in the chain, by
= 0, = 0, = and = 1, … ,
hence the equilibrium equations resulting from the above relations are
given as
(1 ) −( + ) + =− ( +ℓ )+ (2 + ℓ )
(2 ) −( + ) + =− (2 + ℓ ) + (2
+ℓ )
…………………………………………………………….
( ) −( + )
+ =− (2 + ℓ ) + (2
+ℓ )
…………………………………………………………….
( −1 ) −( + )
+ =− (2 + ℓ )+ (2 + ℓ )
( ) −( + )
=− − (2 + ℓ ) + (2 + ℓ )
2.2.3. Oscillations of a set of n solids
Around the equilibrium position represented by = and

= 1, … , , the oscillatory movement is defined by
= + ( ); = ; =
The development limited to the order 1 of the movement equation

system presented above leads to the oscillation equations for the set.
This thus involves developing each of their terms to the first order.
= ( , , )
= ( , , )+ ( , , )
where ( , , ) is the second member of the Lagrange

equation of the motion of the solid around the equilibrium
position . We thus obtain the equation of the small movements (or

oscillations) of :
= −( + ) +
The oscillatory motion of the set of n solids is described by the

system:
(1 ) +( + ) − =0
(2 ) − +( + ) − =0
……………………………………………………….
( ) − + + − =0
…………………………………………………….....
( −1 ) − +( + ) −
=0
( ) − +( + ) =0
2.2.4. Vibration eigen modes of a set of n solids
To give to this example a readily usable result, we introduce

simplification into the statement by setting
= =⋯= =⋯= = =
= =⋯= =⋯= = =
with, moreover, = .
The previous differential system is thus written as
(1 ) +2 − =0
(2 ) − +2 − =0
……………………………………….
( ) − +2 − =0
……………………………………….
( −1 ) − +2 − =0
( ) − +2 =0
To highlight the vibration eigen modes of the set of n solids, a

sinusoidal-type solution of the following form must be sought:
ℰ = ( + )
=− ( + ); ℰ =− ( + )=−
By carrying over a solution of this form into the system above, we

introduce ( + 2) unknowns: , … , , , with the equations
(1) (2 − ) − =0
(2) − + (2 − ) − =0
…………………………………………..
( ) − + (2 − ) − =0
……………………………………………
( − 1) − + (2 − ) − =0
( ) − + (2 − ) =0
By setting = ; =2− , we obtain
(1) − =0
(2) − + − =0
…………………….
( ) − + − =0
……………………..
( − 1) − + − =0
( ) − + =0
Solving the latter system gives, successively
= , = + ⇒ = −1= −1
By also setting
( )= , where ( ) is a polynomial of degree p-1, we obtain

the finite sequence, seemingly recurrent, of common term:
( )= ( )− ( ) with ( ) = 0 and ( )=1
is an amplitude ratio and =2− >0 ⇒ < 2.
Moreover, = is an amplitude ratio, and consequently it is

positive
⇒ 0< <2 ⇒ 0< <2
We then set = 2 cos with 0 < < 1 and sin ≠ 0.
We deduce therefrom
2 cos sin sin 2
= = 2 cos = =
sin sin
= − 1 = 4 cos − 1 = 4(1 − sin )−1

3sin − 4 sin sin 3
= 3 − 4 sin = =
sin sin
= − = − 2 = 8cos − 4 cos
4sin cos cos 2
= 4 cos (2 cos − 1) =
sin
2sin 2 cos 2 sin 4
= =
sin sin
We then have at our disposal the first elements of a seemingly
recurrent sequence with
( )= = 1; ( )=
( )= ; ( )=
Let us now assume that this sequence is true up to the index p – 1,

that is to say,
sin( − 1)
( )=
sin
Continuing this reasoning by recurrence according to the law
( )= ( )− ( )
we have
sin(p − 1) sin( − 2)
= 2 cos −
sin sin
2 cos (sin − cos sin )

= …
sin
(cos 2 sin − cos sin 2 )
− …
sin
sin (2 cos − cos 2 )
=
sin
sin
⇒ =
sin
From the equation (n – 1) above in the same section, we obtain
sin
=
sin
With the recurrence sequence stopping at the equation (n) above,

this involves developing the latter using the same approach as above
sin sin( − 1)
( ) ⇒ − = 2 cos −
sin sin
= sin( + 1) = 0
We obtain from the latter relation
( + 1) = ⇒ =
+1
with = 1, … , ; ≠2; ≠ 0 and ( + 1)
For each value of the coefficient k, there is a corresponding value

= .
Hence, a series of solutions of the form
( ) ( ) sin
= cos[ ( + )] = cos[ ( + )]
sin
( )
which introduces 3n unknowns: , , where = 1, … .
We start by determining by first writing
= = 2− = 2(1 − cos =2 sin

2
For each value of k, we have:
=2 sin for = 1, …
( )
( )
The remaining 2n unknowns, and , for = 1, … , are
determined from the 2n initial conditions of the problem.
(0), ′(0) for = 1, …
2.2.5. Influence of the initial conditions of the problem
The first step is to illustrate the influence of the initial conditions

of a problem and, for the demonstration to be clear, it is sometimes
necessary to simplify the problem being treated. For this reason, the
spring will be removed, and as a result, the (n + 1)th equation of
the different systems presented in the previous sections will be
corrected. We have, successively,
( ∗) = − + (2 + )
( ∗)
− =− (2 + )
( ∗)
− + =0
""" − + =0
( ∗) − +( − ) =0
""" − ( − 1) =0
⇒ ( − 1) − =0
The latter expression can also be written as
sin sin( − 1)
(2 cos − 1) − =0
sin sin
⇒ (2 cos − 1) sin − sin( − 1) = 0
that is, after simplifying
(1 − cos ) sin = sin cos
2sin sin = 2 sin cos cos

2 2 2
54
E 5
As sin ≠ 0 because sin ≠ 0, we

w obtain
1
sin siin = cos cos ⇒ cos( + ) = 0
2 2 2
1 (2
2 − 1)
+ = (2 − 1) ⇒ =
2 2 2 +1
heence, =2 sin .
The valuess of the pulsaation reprresent the eiggen pulsationn modes

off the set off n solids. We thus note a som mewhat predictable
phhenomenon: depending on the extremity of the solid , these
eigen pulsationn modes are different:
– with the spring
s =2 sin
s ;
( )
– without the
t spring =2 sin .
( )
2..3. Studying the vibra us set by passing

ations of a continuou
to
o the limit
As mentioned above, if we wish to arrive at a sufficienttly clear

prresentation of a relativvely complex problem m, it is som metimes
neecessary to examine
e the problem wiithin the conntext of a reelatively
simmple case. For
F this reasoon, we will make
m the folloowing hypotthesis in
orrder to address this part of
o the presenttation.
= ; = , …, = ,…
F
Figure 2.4. Co
onfiguration fo
or passing to the
t limit
We will now study the question of passing to the limit between a

discrete set and a continuous set based on the oscillation equation for
each solid. For the solid
2 2 2
− 0 + +2 0 − 0 =0
where = ( , ) expresses the displacement of the solid

around the equilibrium position = .
( + 1)
( )− ( )= , − ,
+1 +1
""" = + , − ,
+1
Yet, = → 0 when → ∞, that is, we increase the number of

solids in order to have the discrete set tending toward a continuous set.
Likewise, if we consider
( − 1)
( )− ( )= , − ,
+1 +1
""" = − , − ,
+1
when →0
( + , )= ( , )+ ( , )+ ( , )[1 + ( , , )]
2
where ( , , ) → 0 when →0
( − , )= ( , )− ( , )+ ( , )[1 + (− , , )]
where (− , , ) → 0 when → 0.
Consequently,
( )− ( )+ ( )− ( )…
= , 2+ , , + − , ,
2
2 2
2 2+ , , + − , ,
⇒ = 0 2 2
,
( + 1) 2
In order for the discrete set of n solids to be assimilated to a

continuous set, we write
= ; =
for each mass-spring assembly within an interval of length .
Note that designates the density of the solid, E the Young’s

modulus, equivalent to a pressure from a dimensional point of view,
of the continuous set assumed to be homogeneous, toward which the
discrete set of n solids tends. The problem dealt with here relates to
the study of the tension/compression of a homogeneous solid.
Verifying the dimensional consistency of the notations, we have
[ ]= , [ ]= , =
[ ]= , [ ]= = ,⇒ [ ]=
( )
Moreover, = = ⇒ [ ]=
When the set is considered continuous (when → ∞), the

differential equation established above for the solid expresses the
oscillations of a negligibly thick slice of the continuous set.
=
with = = = [ ] ⇒ = ∼
velocity.
The longitudinal vibrations (under tension and under compression)

of the continuous solid are then given by the differential equation
− =0
where expresses the vibration propagation velocity in the continuous

solid, this amount being dependent on the nature of the material of the
solid.
The general solution of an equation of this form
= − + +
is obtained by integrating the differential equation using a variable

separation method, setting
= ( ) ( ) ⇒ = ( ) ( ), = ( )
As = ⇒ =
Note that the first member of this equation is an exclusive function

of time t, and the second is exclusively a function of the variable z. In
order for the equality expressed by this equation to make sense, the
two members, as they are written, need to be independent from the
corresponding variable, the first from z and the second from t. They
can therefore only be equal to the same one constant, K.
= = ⇒ − = 0, =− =0
In order for these two differential equations to express an

oscillatory motion, we must have K < 0, that is to say, = − . We
then obtain
"+ =0 ⇒ = cos + sin
+ =0 ⇒ = cos + sin
hence the general solution expressing the oscillatory motion of the

continuous set, toward which the discrete set of n solids tends, is at the
limit:
= ( cos + sin )( cos + sin )
where:
=2 sin with =
( )
+1
=2 sin
2( + 1)
For the transition from the discrete set to the continuous set, we
have → ∞, consequently
→0 ⇒ sin →
2( + 1) 2( + 1) 2( + 1)
⇒ ∼ when →∞
Thus, using a passage to the limit of a discrete set of n solids, the

vibration eigen modes of a continuous solid working in
tension/compression are given by
=
With the general solution of the vibratory system given by
( , )= − + +
which involves writing the above expression as
( , )= ( cos + sin )( cos + sin )
by developing the different products of sinusoidal functions:
−
( , )= cos + …
2
+
+ sin + …
2
+
+ cos − …
2
−
− sin − …
2
Put in this form, we do find the propagation equation of a wave, for

which we now need to determine the constants from the initial
conditions and boundary conditions of the studied motion.
2.3.1. Taking the boundary conditions into account at any

instant
– For = 0, that is, at point O

(0, ) = 0, ∀ ⇒ ( ) (0) = 0 ⇒ (0) = 0
yet
( )= cos + sin ⇒ (0) = =0
– For = , that is, at point O’

( , ) = 0, ∀ ⇒ ( ) ( )=0 ⇒ ( )=0
⇒ Dsin =0
As ≠ 0, because we would have ( )=0 ∀ , we therefore

have
sin =0 ⇒ = =
– At instant t = 0, we set
( , 0) = ( ) and ( , 0) = ( )
as
( , )= sin cos + sin
so we can identify
( )= sin
( )= sin
To enable calculation of coefficients and , the following

integral first needs to be calculated:
1
= 2sin sin
2
Let us start be recalling that: 2 sin sin = cos( − ) − cos( +

)
1
= cos(k − n) −cos(k + n)
2
– In the case where ≠
1 sin(k − n) sin(k + n)
I= − =0
2 (k − n) (k + n)
– In the case where =
1 (k + n) 1 sin(k + n)
= 1− = − =
2 L 2 (k + n) 2
Therefore, we can write, in general terms,
= sin sin =
2
where is a Kronecker symbol.
We now determine the two terms
( ) sin = sin sin
= =
2 2
( ) sin = sin sin
""" = =
2 2
hence the two constants
2
= ( )sin
2
= ( )sin
After determining the different constants, the general form of the

solution is therefore given as
2
( , )= sin cos ( )sin …
2
+ sin ( )sin
2.4. Exercises
2.4.1. Exercise 1: movement equations – equilibrium
( ⋮ ⃗, ⃗, ⃗) is attached to the Earth, and ⃗ is an upward vertical

unit vector. Let us study the plane mechanical system, as shown in
Figure 2.5.
O
(R)
Figure 2.5. Configuration of the solid (S)
The system is composed of:

– a homogeneous disk of center G, radius R and mass M, welded to
two homogeneous rods, of negligible mass, AH and KB. Points A, H,
G, K and B are aligned;
– two axes, A and B, each of mass m, for which we can overlook

the cross-section dimensions. These axes move without friction in
fixed grooves of respective directions, ⃗ and ⃗;
– of a spring, R, with one extremity fixed to O and the other to the
mobile axis, A. Its free length is 2 ℓ0, its stiffness ( ) and its mass is
negligible.
We use (S) to refer to the solid consisting of the set of the disk, two
rods and two axes
We set: ⃗ = 2ℓ ⃗, ⃗=ℓ ⃗
( ⃗, ⃗) = ( ⃗, ⃗) =
: twice-derivable function of time with derivatives: ′ and ′′.
In addition to the efforts of gravity, the solid (S) is subject to:

– the action of the spring, R:
{ → } = [− ( − 2ℓ ) ⃗ ⋮ ⃗ ];
– the link efforts

{ → } =[ ⃗ ⋮ ⃗ ]
{ → } =[ ⃗ ⋮ ⃗ ].
Part one: kinematics

1) Express the vector ⃗ as a function of ℓ and . Determine the
reduction elements at A of the kinematic torsor . Determine the
instant center of rotation .
2) Determine the trajectory of in (T): Γ ( ), as well as the

velocity and acceleration vectors of in (T): v⃗ ( ) and ⃗J ( ); we
will express these vectors on the basis ( ⃗, ⃗, ⃗).
3) Determine the trajectory of in (S): Γ ( ) as well as the

velocity and acceleration vectors of in (S): v⃗ ( ) and
⃗J ( )J⃗ ( ); we will express these vectors on the basis ( ⃗, ⃗, ⃗).
4) Compare the vectors v⃗ ( ) and v⃗ ( ), and then the vectors

⃗J ( ) and ⃗J ( ).
Part two: dynamics

We will express all the vectors requested on the basis ( ⃗, ⃗, ⃗).
1) a) Determine the vector ⃗ , then the vectors v⃗ ( ) and ⃗J ( ).
b) Determine the reduction elements at G of the dynamic torsor

{ }.
2) Determine the reduction elements at G of { ̅ → } : torsor

grouping together all of the outside efforts exerted on the solid (S).
3) Write the scalar equations deduced from the fundamental

principle of dynamics applied to (S).
4) Deduce the movement equation from the previous equations.
5) a) Calculate the kinetic energy of (S).
b) Calculate the power developed by all of the outside efforts

acting on (S).
c) Write the energy–power equation and show that we find the

equation of motion obtained in Question 4.
Part three: dynamics of two solids

A solid ( ) of mass is fixed by a pivot link of axis ⃗ to the
disk (S) at D, with ⃗ = ⃗ ( > 0). The frame ( ⋮ ⃗, ⃗, ⃗) is
attached to the solid ( ). Its center of gravity is given by:
⃗= ⃗ ( > 0)
0 0
Its inertia matrix is [ ]( ⃗, ⃗, ⃗) = 0 − .
0 −
We set: ( ) = ( ⃗, ⃗) measured on ⃗.
1) As a function of , , , , , ′′, using their components on

the basis ( ⃗, ⃗, ⃗), calculate the reduction elements at D of the
dynamic torsor:
2) The link efforts between the disk and the solid ( ) are
represented by:
{ → } ={ ⃗+ ⃗ ⋮ ⃗+ ⃗}
a) Write the fundamental principle applied to the solid ( ) in the

basis ( ⃗, ⃗, ⃗).
b) Write the fundamental principle applied to the solid ( ) in
the basis ( ⃗, ⃗, ⃗).
c) Determine a relation linking , ℓ, ℓ , , , , , and
when = and = is an equilibrium position.
2.4.1.1. Solution to exercise 1: movement equations –

equilibrium
Figure 2.6. Position of the instant center of rotation, Q0

Part one: kinematics

1) From the triangle OAB, right-angled at O, we obtain: ⃗=
2ℓ sin ⃗
The reduction elements at A of the kinematic (or velocity

distributing) torsor of the movement of (S) with respect to the frame,
attached to the Earth (T), are (for definitions and applications, see
Volume 1 of this series, entitled: “Location, Kinematics and
Kinetics”):
={ ⃗ ⋮ ⃗ }
With point A belonging to the solid (S), we have the following

rotation rate vector:
⃗ = ⃗
and the following velocity vector of A:
⃗ = 2ℓγ'cos ⃗
Here, we have a sliding torsor of axis ( ⋮ ⃗) where is defined

by:
⃗ ∧⃗ ⃗ ∧ 2ℓ cos ⃗
⃗= = = 2ℓ cos ⃗
(‖ ⃗ ‖)
The instant center of rotation of the movement studied here is

the fourth vertex of the rectangle .
2) The trajectory Γ ( ) of in (T) is defined by:
⃗= ⃗+ ⃗ = 2ℓ(sin ⃗ + cos ⃗) = 2ℓ ⃗ −
2
The “basis” of the movement here is a circle of center O and radius
2ℓ.
By deriving with respect to time, we obtain:

– the velocity of in (T):
v⃗ ( ) = 2ℓ (cos ⃗ − sin ⃗) = 2ℓ ⃗(− )
– the acceleration of in the frame (T):

⃗J ( ) = 2ℓ ⃗(− ) − 2ℓ ′ ⃗ −
The latter expression highlights the tangential acceleration and the

normal acceleration.
3) The trajectory Γ ( ) of in (S) is defined by:
⃗ = 2ℓ cos (sin ⃗ + cos ⃗)
= ℓ cos 2 ⃗ + ℓ(1 + cos 2 ) ⃗
= ⃗ + ℓ(sin 2 ⃗ + cos2 ⃗)
or by ⃗ = ℓ(sin 2 ⃗ + cos2 ⃗)
The “mobile part” of this motion is a circle of center G and radius ℓ.
By deriving with respect to t, the components in (S) of the situation

bipoint of in (S), we obtain:
– the velocity of in the frame (S):
v⃗ ( ) = 2ℓ (cos 2 ⃗ − sin2 ⃗)
– the acceleration of in the frame (S):

⃗J ( ) = 2ℓ (cos 2 ⃗ − sin2 ⃗)
− 4ℓ (sin 2 ⃗ + cos2 ⃗)
4) By comparing the two velocity vectors, we obtain:
v⃗ ( ) = v⃗ ( ) = 2ℓ ⃗(− )
– This confirms that ℳ⃗ = 0⃗ as well as the moment at any

point on the ⃗ axis.
– ⃗J ( ). ⃗(− ) = ⃗J ( ). ⃗(− ) : has the same tangential
acceleration in (S) and in (T);
– ⃗J ( ). ⃗ − = 2J⃗ ( ). ⃗ − : different normal

accelerations because the curve radii of the trajectories are different
(here, double of one another).
Figure 2.7. Representation of the basis and mobile part of the motion of (S)
Part two: dynamics
1) a) ⃗ = ℓ(sin ⃗ + cos ⃗) = ℓ ⃗ −
By deriving, with respect to t, the components in (T) of the

situation bipoint of G, we obtain:
– the velocity of G in the frame (T):
v⃗ ( ) = ℓ (cos ⃗ − sin ⃗) = ℓ ⃗(− )
– the acceleration of G in the frame (T):

⃗J ( ) = ℓ (cos ⃗ − sin ⃗) − ℓ (sin ⃗ + cos ⃗)
b) The sum (resultant) of the dynamic torsor is:
⃗{ }=( + 2 )J⃗ ( )
With the plane ( ⋮ ⃗, ⃗) being a symmetry plane for

distribution of the masses of (S), we have:
⋮ ⃗, ⃗ =0; ⋮ ⃗, ⃗ = 0; = ⃗ = +2 ℓ
Thus, the dynamic moment of (S) in its movement with respect to

(T) is:
ℳ⃗ { }= ⃗ + ⃗ ⋀ ( ⃗ )=( +2 ℓ ) ⃗
2
Hence the dynamic torsor:
{ } = {( + 2 )ℓ[( cos − sin ) ⃗ − ( sin

+ cos ) ⃗ ]/ ⃗}
2) The outside efforts exerted on (S) are:
{ ̅→ }={ → }+{ → }+{ → }+{ → }
For the resultant, we have:
⃗{ ̅ → } =Y ⃗ + ⃗- (M+2m)g ⃗ − 2 (ℓ sin − ℓ ) ⃗
Let us calculate the moments at G of these torsors:
ℳ ⃗{ → }= ⃗∧ ⃗ = ℓ sin ⃗
ℳ ⃗{ → }= ⃗∧ ⃗ = −ℓ cos ⃗
ℳ ⃗{ → } = 0⃗
ℳ ⃗{ → } = ⃗ ∧ [−2 (ℓ sin − ℓ )] ⃗
= −2 ℓ(ℓ sin − ℓ ) cos ⃗
hence:
{ ̅ → } = {[ − 2 (ℓ sin − ℓ)] ⃗
+ [ − (M + 2m)g] ⃗
/[ℓ Ysin − ℓ cos − 2 ℓ(ℓ sin − ℓ )cos ] ⃗}
3) By considering the frame attached to the Earth (T) as a Galilean

frame, we have:
{ }={ ̅→ }
The equation of the sum (resultant) projected onto ⃗ and onto ⃗

gives:
( + 2 )ℓ( cos − sin ) = − 2 (ℓ sin − ℓ ) [1]
−( + 2 )ℓ( sin + cos ) = −( +2 ) [2]
The equation of the moment at G projected onto ⃗ gives:
= ℓ sin − ℓ cos − 2 ℓ(ℓ sin − ℓ ) cos [3]
4) To write an equation independent of the link efforts (movement

equation), we create the following combination of equations derived
from the fundamental principle of dynamics:
(3)+(1) ℓ cos − (2) ℓ sin
We obtain:
= −( + 2 )ℓ + ( + 2 ) ℓ sin
− 4 ℓ(ℓ sin − ℓ ) cos
i.e.
[C + (M + 2m)ℓ ]γ = (M + 2m)gℓ sin γ − 4kℓ(ℓ sin γ −

ℓ ) cos γ [4]
5) The kinetic energy of (S) is defined by:
( ) = ( + 2 ) ⃗( ) + ⃗ . (⃗ )
i.e.
1
( ) = [( + 2 )ℓ + ]
2
The powers developed in (T) by the efforts are given as
{ → }= { → }=0
{ → }=0
{ → } = −( + 2 ) ⃗. ⃗ = ( +2 ) ℓ sin
{ → } = [−2 (ℓ sin − ℓ )] ⃗. ⃗
= −4 ℓ (ℓ sin − ℓ ) cos
Hence the energy–power equation:
( )
= { ̅→ }
i.e.
[ +( + 2 )ℓ ]
= [( + 2 ) ℓ sin − 4 ℓ(ℓ sin − ℓ ) cos ]
By simplifying with , we find the movement equation of (S): [4].
Part three: dynamics of two solids
O
Figure 2.8. Configuration of the two solids
1) The center of gravity of ( ) is situated in (T) and given as
⃗= ⃗+ ⃗+ ⃗
= ℓ⃗ − + ⃗( ) + ⃗( )
2
By deriving ⃗ with respect to t, we obtain:

– the velocity of in (T):
⃗ = ℓ ⃗(− ) + ⃗ + + ⃗ +
2 2
– the acceleration of in (T):
⃗ = ℓ ⃗(− ) + ⃗ + + ⃗ +
2 2
−ℓ ⃗ − − ⃗( ) − ⃗( )
2
Hence the components in the basis (2) of the dynamic resultant:
⃗ = ⃗ [− − cos( − ) + sin( − )
+ℓ cos( + ) − ℓ sin( + )]
+⃗ [ + sin( − ) + cos( − )
−ℓ sin( + ) − ℓ cos ( + )]
Let us set:
⃗ = [ ⃗ + ⃗ ]
The dynamic moment at D of ( ) in its movement in (T) is

defined by:
ℳ ⃗{ }= ⃗∧ ⃗ + ⃗ + ⃗ ∧ (⃗ )
with ⃗ = ⃗
Let us calculate each term separately using its components in the

basis (2) = ( ⃗, ⃗, ⃗):
⃗∧ ⃗ = ⃗∧( ⃗+ )⃗ = ⃗
= [ + sin( − )
+ cos( − ) + ℓ sin( + )
− ℓ cos ( + )] ⃗
⃗ = (− ⃗+ ⃗)
⃗ ∧ ( ⃗ )= ⃗
hence:
ℳ ⃗{ }= ⃗− ⃗
+ ( + )
+ [ cos( − ) − ℓ sin( + )
+ sin( − ) − ℓ cos ( + )] ⃗
2) a) The fundamental principle of dynamics applied to the solid

(S) is written as
{ }={ → }+{ → }+{ → }+{ → }−{ → }
With a view to projecting the resultant onto the basis (x ⃗,y ⃗,z ⃗)
and the moment at G of this torsor equality onto the basis (2), let us
calculate:
ℳ ⃗{ → } = ℳ⃗ + ⃗∧ ⃗= ⃗ + ⃗ + ⃗ ∧( ⃗ + ⃗ )
= ⃗ + ⃗ + [ ( − )+ ( − )] ⃗
Hence, six scalar consequences are given as
[1′] ( + 2 )ℓ( cos − sin ) = − 2 (ℓ sin −

ℓ )− cos + sin
[2′] −( + 2 )ℓ( sin + cos ) = −( +

2 ) − sin − cos
[3′] 0 = 0
[4′] 0 = L
[5′] 0 = M
[6′] = ℓ sin − ℓ cos − 2 ℓ(ℓ sin − ℓ ) cos −

[ sin( − ) + cos( − )]
b) The fundamental principle applied to the solid ( ) is written

as
={ → }+{ → }
With a view to projecting onto the basis ( ⃗, ⃗, ⃗), the resultant

and the moment at D of this torsor equality, let us calculate:
ℳ ⃗{ → } = ℳ ⃗{ → }+ ⃗⋀ ⃗{ → }
= 0⃗ − ⃗⋀ ⃗=− ⃗
Hence, six scalar consequences are given as
[1′′] [− − cos( − ) + sin( − ) +

ℓ cos( + ) − ℓ sin( + )] = −
[2′′] [ + sin( − ) + cos( − ) −

ℓ sin( + − ℓ cos ( + ) = −
) ]
[3′′] 0 = 0
[4′′] =
[5′′] − =
[6′′] ( + ) + [ cos( − ) − ℓ sin( +

)+ sin( − ) − ℓ cos ( + )] = −
c) At a solid’s equilibrium position, its dynamic torsor is null.

Therefore, the first members of equations [1′] to [6′] and from [1′′] to
[6′′] are zero. At equilibrium, the values of the situation parameters of
each solid must cancel out the second members.
Equations [3′] and [3′′] are always satisfied.
Equations [4′], [5′], [4′′] and [5′′] indicate that L = M = O.
For γ = and = , equations [1′′] and [2′′] provide:
=− and =0 (values of the contact efforts at

equilibrium).
Equation [6′′] is satisfied for =
Equations [1′] and [2′] provide the values of the link efforts at
equilibrium:
ℓ√2
X=2 −ℓ , =( +2 + )
2
Lastly, equation [6′] is satisfied if:
ℓ√2
4 −ℓ = +2 + (1 − )
2 ℓ
which constitutes the relation between the (constant) data of the

problem so that this equilibrium position is achievable.
2.4.2. Exercise 2: movement around an equilibrium position
The device represented in Figure 2.9 within its symmetry plane is

composed of two solids and three springs.
76
E 5
Fiigure 2.9. Sysstem of two so

olids and three
e springs
The solid ( ) is a reectangular, homogeneou

h us parallelepiiped, of
mass
m , onee face of whhich is in co ontact with the
t horizontaal plane
( ′ ⋮ ⃗, ⃗) attaached to the Galilean fraame < > = ( ⋮ ⃗, ⃗, ⃗). ( )
is in straight-liine translatioon along ( ′ ⋮ ⃗) with resspect to < >.
We attach to ( ) the frame < 1 > = ( ⋮ ⃗, ⃗, ⃗), beeing the

ceenter of masss of ( ). Thhe position of
o ( ) with respect to < > is
deefined by:
⃗ = ( )⃗
Two identiical tension/ccompression

n springs, ( ) and ( ),, of free
lenngth ℓ and stiffness
s k, are
a fixed at B and C to ( ) and at A and O,
reespectively, to < > succh that:
⃗ = ⃗= ⃗ (b
b being a positive constannt)
{ → } = (ℓ + − ) ⃗ ⋮ 0⃗ = { → }
The solid ( ) is compposed of a cylindrical baar of axis ( ⋮ ⃗),

welded to a sppherical mass of center . ( ) has mass andd center
off mass defined by ⃗ = ⃗ (a a being a possitive constannt), with
thhe matrix reppresentative of the inerttial operator at on thhe basis
( ⃗, ⃗, ⃗) attaached to ( ):
)
0 0
= 0 0
0 0
( ) is in perfect pivot link of axis ( ⋮ ⃗) with respect to ( ).
We set:
( ) = ( ⃗, ⃗) = ( ⃗, ⃗).
A torsion spring ( ), of stiffness Γ > 0 has one extremity fixed to

( ) and the other to ( ). We set:
{ → } = 0⃗ ⋮ −Γα ⃗ = −{ → }
The gravity is defined by: ⃗ = − ⃗ ( > 0)
We associate:
– the link of the plane ( ⋮ ⃗ , ⃗) with ( ) the torsor:
{ → } ={ ⃗ ⋮ ⃗+ ⃗}
– the link of ( ) with ( ) the torsor:

{ → } ={ ⃗+ ⃗+ ⃗ ⋮ ⃗+ ⃗}
Questions and Solutions for exercise 2
Question 1:
Using their components on the basis ( ⃗, ⃗, ⃗), determine the
reduction elements at of the torsors:
a) Kinematics: , ,
b) Dynamics: ,
1) a) The solid ( ) being in translation with the frame ( /

⃗, ⃗, ⃗), assumed to be Galilean, its position is situated by: ⃗ = y ⃗ .
The solid ( ) is in pivot link of axis ⃗ with respect to ( ).
We set: ⃗ = a ⃗ = a ⃗( )
We have the kinematic torsors:
= ⃗ /⃗ = ⃗/ ′⃗
= ⃗ /⃗ −⃗ = { ′ ⃗/ ⃗ }
By composing the movements, we obtain:
= + ={ ′ ⃗/ ′ ⃗ }
b) The sum (or resultant) of the dynamic torsor of the solid (S1) is:
⃗ = ⃗= "⃗
and its dynamic moment at is:
ℳ⃗ = ( )( ⃗)+ ⃗^ ( )( ⃗) = 0⃗
as ⃗ = 0⃗
The resultant of the dynamic torsor of the solid (S2) is:
⃗ = ⃗= − ⃗( ) + " ⃗( + ) + ′′ ⃗
= - ( cos + " )⃗ + ( "− sin +

′′cos ) ⃗
For the dynamic moment of ( ) at , we have:
ℳ⃗ = ⃗∧ ⃗+ ( )( " ⃗) + ′ ⃗ ∧ ( )( ′ ⃗)
= (m y′′ cos + C ′′) ⃗
Question 2:
Write, in tensorial form, the fundamental principle of dynamics
applied to (S1), and then to (S2):
={ → }+{ → }+{ → }+{ → }+

{ → }+{ → }
={ → }+{ → }-{ → }
with { → } = − ⃗/0⃗ and { → } = − ⃗/0⃗
hence ℳ ⃗{ → }= ⃗ ∧ (- ⃗) = sin ⃗
Question 3:
Write the scalar consequences on ( ⃗, ⃗, ⃗) of the fundamental
principle applied to (S1), and then to (S2).
Six scalar consequences of the equation of the resultant and the

moment at are, for (S1):
[1] 0 = X+ - ;
[2] "= +2k(ℓ+b-y) ;
[3] 0 = ;
[4] 0 = ;
[5] 0 = + ;
[6] 0 = N+Γ ;
and for (S2):

²
[7] - − " =- − g;
[8] - ²
+ " + "=- ;
[9] 0=- ;
[10] 0 = - ;
[11] 0 = - ;
[12] ′′ '' = -Γ + .
Question 4:
Deduce therefrom the two movement equations (equations not
involving an unknown link).
As equation [12] does not contain any unknown link, it is therefore

a movement equation. We obtain the second equation by adding,
member by member, equations [2] and [8]:
[13] ( + ) "- aα'² sinα+ aα'' = 2 (ℓ + − )

[12] + Cα'' = −Γ +
Question 5:
– Verify that = 0 is an equilibrium position and determine :
the value of y at equilibrium.
– For what value of Γ does another equilibrium position
∈ 0, also exist?
The possible equilibrium positions of the set of two solids are

obtained by setting = and = in the movement equations.
[13]eq: 0 = 2k(+b - )
[12]eq: 0 = −Γ + because ′= ′′ = ′′ = 0
From the first equation, we deduce that = ℓ + b.
= 0 satisfies the equation [12]eq and therefore corresponds to an

equilibrium position of the set of two solids.
In general, the equation [12]eq is verified if:
[14] = sin
Vibration Mec
chanics of Syste
ems of Solids 81
Figure 2.1
10. Plot of func
ctions
Fromm the plot of functionns = sin , = , = , w

we
deducee that:
– = is thee straight-linee tangential to the originn on the plot of
;
– = plot intersectss the plot of
o (besiddes the origin)
if k = < 1.
We therefore haave a secondd equilibrium

m position iff Γ < ga. Its
value ∈ ]0, [ is
i a solution of equation [14].
[
Questiion 6:
Wriite the moveement equatiions linearizzed around thhe equilibriuum
positionn: y = , = 0
By writing:
w y = + (t) and = (t), we connsider andd
as infinnitely small first-order
f vaalues.
At the
t first orderr, cos = 1 and
a sin = .
Therefore, the linearized movement equations are:
[13]lin: ( + ) ′′+ 2k + ′′ = 0
[12]lin: ′′+ ′′+ ( − g) =0
The presence of the term ′′ in the first equation and the term
′′ in the second equation indicates that there is a first-order
coupling between the variations of the movement parameters of the
two solids.
Question 7:
²
In the case where: = = , = = , G = 2ma², Γ =
²
3ma²
a) Show that the movement equations are reduced to the system:

''+d1 ω²0 ε1 +d2 ε2 " = 0
²
''+d3 ε2 ′′ + =0
where , , and are numerical coefficients to be determined.
By replacing these particular values in the linearized movement

equations, we obtain:
1
''+ω²0 ε1 + ε2 " = 0
2
′′ + 2 " + 2 ² = 0
We do find the differential system indicated with:
d1 = 1; d2 = ½; d3 = 2; d4 = 2
b) We wish to find solutions of the form:

= ( + )
= ( + )
Determine the eigen pulsations of the system.

By replacing these expressions of and in the linearized

movement equations, we obtain the system:
² ²
− − =0
2
² ² ²
− +2 − =0
The ratios extracted from these two equations are written as
( ² ²) ²
[15] = ²
= ²)
(
The last two fractions provide the values of (eigen pulsations),

which satisfy the movement equations:
− =
− = ⟹ ⇒
− =−
² ²
=
² ²
=2
Thus, the eigen pulsations of the system are: = and

= √2 .
c) Determine the general solution ( ) and ( ) of the movement

equations.
From equations [15], we deduce as a function of for each
(i = 1 or 2). corresponds to = and corresponds to = − .
Hence, the general solution:
[16] (t) = cos + + √2 +
(t) = cos + - √2 +
Question 8:
If the initial conditions are:
(0) = , (0) = 0
(0) = 0, (0) = 0
that is to say, if we break the equilibrium by giving a displacement

to the solid (S1) and releasing the set, give the solutions (t) and ( )
representing the small motions about the equilibrium position = 0,
y = . What is the trajectory of the center, G2, of the sphere, and what
is its movement with respect to (g) and with respect to (S1)?
Let us replace the initial conditions in the system [16] to determine

the arbitrary constants: , , and resulting from integration
of the system of equations of motion. We obtain:
= +
0= −
0=− − √2
⇒ sin =sin =0
0=− + √2
Without restricting the generality, one can take = = 0 to

determine and from the first two equations:
= +
⇒ = =
0= − 2
Hence: ( )= (cos + √2 )
( )= (cos − √2 )
The center G2 of the sphere is situated in (g) by:
⃗= ⃗+ ⃗=( + )⃗ + ⃗ (a positive constant)
At the first order:
[ ⃗]= [cos ⃗ + ⃗] = ⃗ + ⃗ with = Thus:
⃗= (ℓ + + ) ⃗ + a( ⃗ + ⃗)= a ⃗+ (ℓ + b+ cos t) ⃗
The center of the sphere has a straight-line trajectory (constant

coordinate a on the x-axis) in (g), described with a simple vibratory
movement of amplitude .
In the frame ( / ⃗, ⃗, ⃗) attached to (S1), we locate G2 by:
⃗= a ⃗ = a (cos ⃗+sin ⃗)
At the first order, we have:
⃗= a( ⃗ + ⃗) = a ⃗ + (cos t - cos√2 t) ⃗
G2 thus has a straight-line trajectory with respect to (S1), described

with a vibratory movement governed by the law k(cos −
cos√2 t).
2.4.3. Exercise 3: dynamics of an RTT robot (one rotation +

two translations)
  
( 
)
In the Galilean frame (g)= Og x g , yg , zg , with zg upward
vertical, attached to the Earth, we study the movement of a set (D) of
three solids, (S1), (S2) and (S3). The linking and guide parts, which are
not essential for understanding the movements, are not described and
will be considered to be of negligible mass.
  
(
– The solid (S1), to which we attach the frame O1 x1 , y1 , z1 , is )

( )
in pivot link of axis O1 z1 with respect to (g). We set: O1 = Og
      ⎯⎯
→ 
γ 1 = ( x g , x1 ) = ( y g , y1 ) measured on z1 = z g , O1G 1 = d1 z1 ,
d1: positive constant.
m1: the mass of (S1); G1: its center of mass.
I O1 : designates the inertial operator at O1 of (S1), represented in

  
the basis ( x1 , y1 , z1 ) by the matrix:
 A1 0 0 
 IO1 ( S1 )  =  0 B1 0 
 
 0 0 C1 
  
(
– The solid (S2), to which we attach the frame O 2 x 2 , y 2 , z 2 , is )

in sliding link of axis ( O 2 z 2 ) with respect to solid (S1). We set:
⎯⎯
→       
O1O 2 = z 2 z 2 ; x 2 = x1 , y 2 = y1 , z 2 = z1
⎯⎯ → 
O 2 G 2 = b 2 y 2 , b2: positive constant.
I O 2 : designates the inertial operator at O2 of (S2), represented in

  
the basis ( x1 , y1 , z1 ) by the matrix:
 A 0 0
 2 
I
 O2 ( )
S2  =  0

B2 −D2 

 0 −D2 C2 

  
(
– The solid (S3), to which we attach the frame O 3 x 3 , y3 , z 3 , is )
 
(
in sliding link of axis O 3 z 3 = O 2 y 2 ) ( ) with respect to solid (S2).
We set:
⎯⎯ →
      
O 2 O 3 = z 3z 3 x 3 = z 2 , y 3 = x 2 , z 3 = y 2
⎯⎯→  
O 3G 3 = a 3 x 3 + d 3 z 3 , a3 and d3: positive constants.
I O 3 : designates the inertial operator at O3 of (S3), represented in

  
the basis ( x 3 , y3 , z3 ) by the matrix:
 A 0 −E 3 
 3 
I
 O3 ( ) 
S3  = 

0 B3 0 

 −E 3 0 C3 

γ 1 , z 2 and z 3 : twice-derivable functions of time.
Given efforts:
{π → 1} , {π → 2} and {π → 3} due to the action of gravity

 
with local intensity g = − g z g
    
{Δ }
g →1 O
1
{ 
}
= 0 N1z1 , {Δ1→ 2 }
O2
{ }{
= Z2 z 2 0 , Δ2→ 3 }
O3
{
= Z3 z 3 0 }
Link efforts: we have here three perfect links
{L } ; {L } ; {L } represent the link efforts

g →1 1→ 2 2→ 3
88
E 5
Fig
gure 2.11. Ro
obot (one rotattion + two tran
nslations)
Prreliminary question:
q
Write the basis
b change matrices: [ g,1
g ] and [ 2,3] .
Q
Questions:
  
1) Using thheir componnents in the basis
b (x2 , y2 , z2 ) , determ
mine the
reeduction elem
ments of thee following kinematic torsors, at thhe point
inndicated:
g 
a)   at O1 and att O3
1 
g 
b)   at O2 and at G2
2
g 
c)   at O3 and at G3
3
2) Calculate the Galilean kinetic energy:

g
a) of the solid (S1): T1 ;
g
b) of the solid (S2): T2 ;
g
c) of the solid (S3): T3 .
3) Calculate the power developed by each given effort.
4) Write, for the set (D), the Lagrange equations relating to the
parameters: γ 1 , z 2 , z3 which provide movement equations in this
study, where the links are perfect.
5) In the particular case where Z2 is independent of t, calculate z2(t)

and describe the movement of the solid (S2).
6) In the particular case where Z2 = −k ( z 2 −  0 ) with k and ℓ0

positive constants, calculate z2(t) and describe the motion of the solid
(S2) and the solid (S3) that it carries.
Solution to exercise 3: dynamics of a robot
Preliminary question:
The basis change matrices between two bases of this exercise are:
− 0 0 1 0
[ , 1]= 0; [2,3] = 0 0 1
0 0 1 1 0 0
Solutions:
1) a) The kinematic torsor of the solid ( ) in pivot link with the
frame has as its reduction elements
at : = ′ ⃗ 0⃗
1
This torsor being a slider of axis ( / ⃗ )=( / ⃗ )=( / ⃗ ).
We have:
ℳ⃗ = ℳ⃗ = ℳ⃗ = 0⃗
1 1 1
and ℳ⃗ = ℳ⃗ + ω⃗ ∧ ⃗= ⃗ ∧ ⃗=- ⃗
1 1
b) The solid ( ) being in sliding link with the solid ( ) of axis
( / ⃗ ) = ( / ⃗ ), its kinematic torsor is:
1 1
= = 0⃗ ⃗
2 2
By composition of movements, we have:
1
= + ={ ′⃗ | ⃗ }
2 2 1
By again using the formula of the moment of a torsor, we obtain:
ℳ⃗ = ′ z⃗ + γ′ ⃗ ∧ ⃗ = - γ′ ⃗ + ⃗
2
This vector represents the velocity of the point in its movement
in (g): ⃗ ( ).
c) The solid ( ) being in sliding link with the solid ( ) of axis

( | ⃗ ) = ( | ⃗ ) its kinematic torsor is: 2 = 0⃗ ′ ⃗ .
3
Through composition of movements, we have:
2
= + ={ ′⃗ | ⃗+ ′⃗ }
3 3 2
We deduce from this:
ℳ⃗ = ′ ⃗+ ⃗+ ′ ⃗∧ ⃗
3
=- ′ ′ ⃗+ ⃗+ ′ ⃗
ℳ⃗ = ℳ⃗ + ⃗∧ ⃗
3 3
= ℳ⃗ + ′ ⃗ ∧( ⃗ + ⃗ )
3
= - (z d ) ′ x⃗ + ⃗+ ′ ⃗= ⃗ ( )
This vector represents the velocity of the point G3 in its movement

in (g).
2) From the general formula for kinetic energy
= [⃗ ( )] + ⃗ .( ⃗ ∧ ⃗ ( )) +
⃗ . ( )( ⃗ )
we can calculate:
a) = ⃗ . ( )( ⃗ ) = the kinetic energy of

( );
b) = [⃗ ( )] + ⃗ .( ⃗ ∧ ⃗ ( )) +
⃗ . ( )( ⃗ )
= + γ′ ⃗ . ( ⃗ ∧ ⃗)+
= + the kinetic energy of ( );

c) = [⃗ ( )] + (⃗ ∧ ⃗ ). ⃗ ( )
+ ⃗ . ( ) ⃗ = ( + + )
+ ⃗ ∧( ⃗+ ⃗ )]. ⃗ ⃗ ( )+
= ( + + )+ (− ′ ⃗ ).(- ′ ⃗ +
1
′ ⃗+ ⃗) +
2
= [( +2 ) + + ]+ the kinetic
energy of ( ).
To obtain the Lagrange equations, we need to know the total

kinetic energy of the system.
(D) = 1∪2∪3:
T= + + = [ ( +2 )+ + + ]+
( + ) +
3) The three links being, by definition, perfect, they do not develop

power.
The powers developed by the efforts of gravity are obtained from:
{ → }= { → }=- ⃗ . ⃗ ⃗ ( ), i = 1, 2 or 3
Thus: { → 1} = 0 because G1 belongs to the axis of rotation

( |⃗ )
{ → 2} = − ⃗ .. − ⃗ + ⃗ =− z′
{ → 3} = − ⃗ . .[-( ) ′ ⃗ + ′ ⃗ + ⃗]=
− z′
The powers developed by the given efforts are:
{Δ → 1} = {Δ → 1} = ⃗ . ⃗ = ′
1
{Δ → 2} = 1 ⃗ 0⃗ =
2
{Δ → 3} = 2 ⃗ 0⃗ =
3
4) We have three Lagrange equations for the mechanical system

studied, one per parameter: , and .
For , we can write
( ): − =Π { → }
The power coefficient relative to is:
Π { → } = 1, { → } = ⃗ . ℳ⃗ {∆ → 1} =
We obtain the same result by taking the coefficient of ′ in the

total power expression.
To obtain the Lagrangian relative to , let us calculate each term

separately:
with = 0 and = ′ [ ( +2 )+ + + ]
− = [ ( +2 )+ + + ]+2 ′
( ′ + ′ )
hence:
( )∶ [ ( +2 )+ + + ]+2 ′
′ ( + )=
The Lagrange equation relative to is written as
( ): − = Π {2 ∪ 3 → 2 ∪ 3}
with
Π {2 ∪ 3 → 2 ∪ 3} = 2, {2 ∪ 3 → 2 ∪ 3} =
⃗ . ⃗ {2 ∪ 3 → 2 ∪ 3}= - ( + )g +
= 0,
=( + ) ′ , =( + ) ′′
hence:
( ):( + ) ′′ = −( + ) +
The Lagrange equation relative to is written as
( )∶ − = Π {3 → 3}
with Π {3 → 3} = 3, {3 → 3} = ⃗ . ⃗{3 → 3} =
= ( + )
= z ′, = z ′′
hence:
( )∶ z ′′ − ( + )=
We see that there is coupling of the variations of and as

functions of t.
5) In the case where is independent of t

we have: ′′ = − + independent of t
then ′ = (− + ) + ′
= − + + ′ +
+ 2
We deduce therefrom that the solid ( ) is falling relative to (S1) at

constant acceleration, along ⃗.
and ′ are the initial movement conditions.
6) In the case where = − ( − ℓ ) (for example under the

action of a traction-compression spring of stiffness k and natural
ℓ
length (ℓ ), ( ) becomes ′′ + =− + , which
admits
– the solution without the second member:
= +
with ω the natural pulsation ω = ;
– the particular solution =ℓ − g=ℓ − ;
Hence, the variations of z2 as a function of time:
(t) = + +ℓ −
We deduce from this that the solid ( ) and the solid ( ) that it
carries, oscillate along ⃗ with the pulsation = around
the equilibrium position defined by = ℓ − .
A and B are constants to be determined from the initial conditions.

3
Vibrations with N Degrees of Freedom
3.1. Introduction
The study of the vibrations of a rigid solid, a set of several solids or

a deformable solid is based on knowledge of the variations as a
function of time – or of the frequency – of several parameters.
The variations in these parameters with respect to time tend to be

coupled. The movement equations of the mechanical set considered
then constitute a second-order differential equations system.
Linear modeling of a system with 2 degrees of freedom (DOF)can

rapidly give a differential system of the same form as that constituting
the movement equations of a system with two mass-spring-dampers:
f 1 (t ) f 2 (t )
k1 k2 k3
m1 m2

x
c1 c2 c3
x1 x2
Figure 3.1. Schematic diagram of a complete system with 2 DOF

Let mi (kg) be the mass of the solid (Si) (i = 1 or 2) located at

distance xi (t) (m) from its equilibrium position. Let us use xi′ to note
the velocity and xi′′ to note the acceleration of (Si). The two solids are
subject to the efforts of gravity acting in the direction perpendicular to

x , to the efforts due to the springs of stiffness kj (j = 1, 2 or 3) (N/m),
to the efforts of the dampers of constant cj (Ns/m) and to the outside
efforts fi (t) (N).
The movement equations of the two masses in translation (assured


by a perfect sliding link) along x are
m1 x1′′=− k1 x1 − c1 x1′ − k 2 (x1 − x2 ) −c2 ( x1′ − x2′ ) + f1 (t )
m2 x2′′=k 2 ( x1 − x2 ) +c2 (x1′ − x2′ ) − k3 x2 − c3 x2′ + f 2 (t )
The variations of x1 and x2 with respect to time are coupled

elastically (spring of stiffness k 2 ) and by damping (viscous here, of
constant c 2 ). In the event that one of the masses is not linked to the
frame, the term concerning the spring and the corresponding damper
simply needs to be eliminated from the movement equations. If none
of the outside efforts applied to the system, such as the perfect sliding
links, has a component in the direction of translation, we have
vibrations around an overall motion (rigid body motion) and the
parameters x1 and x2 are measured from the initial position of each
solid.
This differential equations system can be written in matrix form as
m1 0   x1′′ c1 + c2 − c2   x1′ 

+
 0 m   x′′   − c
 2 2  2 c2 + c3   x2′ 
k + k − k2   x1   f1 
+ 1 2 =
 − k2 k2 + k3   x2   f 2 
Vibrations with N Degrees of Freedom 99
Thus, the movement equations of a mass-spring-damper system,

with N DOF, are represented by the following system:
[ M] { x′′} + [C] { x′} + [ K] { x} = { f }

where [ M] is the mass matrix, squared (N × N) and diagonal, of
positive elements m i ; [ C] is the damping matrix, square (N × N) and
symmetric, of elements cij = c ji ; [ K] is the positive-defined stiffness
matrix, square (N × N) and symmetric, of elements k ij = k ji ; {x } is
the column matrix (N × 1) representing the displacement vector and
{ f } is the column matrix (N × 1) representing the excitation force
vector.
The case where the mass matrix is not diagonal but simply
symmetric will be treated as an example, as the procedure for solving
the movement equations is the same. A system with 2 DOF will serve
as an example when the detailing of the process with explanation of
the results is sought.
The solution to this linear differential equations system is obtained

in two stages.
3.2. Homogeneous system – free vibrations (f1 = f2 = 0)
3.2.1. Without damping (cij = 0)
In the absence of excitation and damping, the movement equations

in matrix form are written as
 m1 0   x1′′  k11 − k12   x1  0

0 + =
 m 2   x ′2′   − k12 k 22   x 2  0
where we have set k12 = k2, k11 = k1 + k2 and k22 = k2 + k3.

Thus, the vibrations of the “conservative” system to be studied are

generally governed by the system:
(Σ ) [ M]{ x′′} + [ K]{ x} = { 0}
3.2.2. Solving the system ( Σ )
Let us seek the harmonic functions: { x} = { X} e iω t
and consequently: { x ′′} = − ω 2 { X} e iω t ,
solutions of the system (Σ ) . The vectors {X} and the scalars ω,

which are independent of t, must satisfy the relation:
([ K ] − ω 2
[ M ]) {X} eiω t = { 0} , ∀t
This is a generalized eigen value problem.
The system Σ ( ) admits solutions, other than the basic solution

{X }= {0 }, if:
det| K − M|=0
For a system with N DOF, this relation is an equation of degree N

in ω 2 :
a 2 N ω 2 N + a 2 N − 2 ω 2 N − 2 + ... +a 0 = 0
It therefore admits N roots in ω 2 , which are real and positive

because of the properties of the mass and stiffness matrices. We
deduce N positive roots: ω1 , ω2 , ... , ωr , ... , ωN .
These are the natural pulsations of the free, undamped system, as

we will see in the system’s temporal response.
For each of these ωr , we determine a vector { Xr } from the
relation: ([ K ] − ω 2
r [ M ]) {X r } = { 0}
Corresponding to a particular mode, of pulsation ωr , there is a
form vector { Xr } , which translates the mode shape (normal mode).
The set of pulsations ω r and form vectors { Xr } (r = 1, 2, ..., N)
constitutes the modal model of the structure.
3.2.2.1. Solution of (Σ2): 2-DOF system

Thus, for a 2-DOF system, we have:
m1 0   x1′′ k1 + k 2 − k 2   x1  0

+ =
(Σ 2 )  0 m2   x 2′′   − k 2 k 2 + k 3   x 2  0
The eigen pulsations are obtained from the relation:
k1 + k 2 − m1ω 2 − k2
dét =0
− k2 k 2 + k 3 − m 2ω 2
By developing the determinant, we obtain the equation
(k 1 )( )
+ k 2 − m1ω 2 k 2 + k 3 − m 2ω 2 − k 22 = 0
i.e.
m1 m 2ω 4 − [m1 (k 2 + k 3 ) + m 2 (k1 + k 2 )] ω 2 + k1 k 2 + k 2 k 3 + k1 k 3 = 0
Here, we have a biquadratic equation in ω 2 with a positive

discriminant, positive root sum (the opposite sign to that of the
coefficient of ω 2 ) and positive root product (the same sign as
the coefficient of the term without ω ). We deduce therefrom that the
roots in ω 2 are real and positive, except in the case of an overall

motion (k11 = k 22 = k12 = k 2 ) , where one of the roots is zero.
We can conclude that this equation always admits two real positive
roots: ω1 and ω2, two eigen pulsations of the conservative system.
 X 1r 
The vector {X r}=   corresponding to ω r , r ∈ {1, 2}, of
 X 2r 
arbitrary magnitude, is a solution vector of the system ( Σ 2 ) obtained
from one or the other of these two dependent equations:
 k1 + k 2 − m1ω r2 − k2   X 1r   0 
  = 
 − k2 k 2 + k 3 − m2ω r2   X 2 r  0
Hence the solution of ( Σ 2 ) :
 x1  − iω1t  X 11  −iω 2 t  X 12 
(iω1t
 x  = B1e + C1e ) iω 2 t
(
 X  + B2 e + C 2 e X  )
 2  21   22 
which can also be written using trigonometric (real) functions:
 x1   X 11   X 12 
 x  = (D1 cos ω1t + E1 sin ω1t )  X  + (D2 cos ω 2 t + E 2 sin ω 2 t )  X 
 2  21   22 
or alternatively: x1 = X 11 A1 sin (ω1t + ϕ1 ) + X 12 A2 sin (ω 2 t + ϕ 2 )
x 2 = X 21 A1 sin (ω1t + ϕ1 ) + X 22 A2 sin (ω 2 t + ϕ 2 )
where A1 , A2 , ϕ1 and ϕ 2 are constants to be determined from the

initial conditions x1 ( 0 ) , x 2 ( 0 ) , x1′ ( 0 ) , x′2 ( 0 ) . We can see here
that the arbitrary choice of form vector magnitude disappears by
introducing the initial conditions. These displacement expressions
give the eigen movement of the two solids in translation, initiated by

these initial conditions.
3.2.2.2. Modal model

For an N-DOF system, we will use the following property.
3.2.2.2.1. Orthogonality property of the modal model

The modal model consists of the N natural pulsations ωr obtained
from: det| K − M | = 0 and the N corresponding form vectors
{X r } obtained from:
([ K ] − ω 2
r [ M ]) {X r } = { 0}
The form vectors of two different modes, r and s, corresponding to
ωr and ωs, are not orthogonal and their scalar product is non-zero:
{Xs }T {X r } ≠ 0 .
On the other hand, by multiplying, on the left-hand side, the
definition relation concerning the rth mode, with the transposed
T
matrix of a form vector {X s } that is different (r ≠ s) , we obtain a
zero scalar product:
{Xs }T ([K] −ω r2 [M]){X r } = 0

Likewise, for the s-mode form vector, we have:
([K ] −ω [M ]){X } = {0}

2
s s
which, transposed and multiplied by { Xr } , on the right-hand side,

gives:
{Xs }T ([K]T −ωs2[M]T ){X r } = 0

Given that [K] and [M] are symmetric matrices: [ K ] = [ K ] and

T
[M ] T
[ ]
= M , we obtain by subtracting, member by member, the
relations above:
(ω 2
s )
−ω r2 {X s } [M ]{X r } = 0
T
As the modes are different ωs ≠ ωr , we have:
{X s }T [M]{X r } = 0
then we also have:
{X s }T [K]{X r } = 0
Furthermore, by multiplying the left-hand of the same relation with

{X r }T we have:
{X r }T [K ]{X r } = ω r2{X r }T [M ]{X r }
and using k r to note the quadratic form: k r = {X r }T [K ]{X r } > 0
and μr for the quadratic form: μ r = {X r }

T
[ M ] {X r } > 0 , this
relation becomes:
kr k r : generalized stiffness
2
r = of the rth mode.
r r : generalized mass
{ X r } [ K ] {X r }
T
The ratio ω = 2
is the Rayleigh quotient.
{X r } [ M ] {X r }
T r
Taking into account the orthogonality properties of the modes, we

obtain:
 0
 = μ
– [ Q] [ M ] [ Q] =   [ ] the generalized mass matrix,
T
μr
 0 
 0
– [ Q] [ K ] [ Q] =
T     the generalized stiffness
 kr  = Kg 
 0 
matrix, with [Q] the matrix, the rth column of which is composed of
the vector {X r } :
[Q] = {X1} , ... , {X r } , ... , {X N } .
 0  0
 μr  and  kr  are diagonal matrices of
   
 0   0 
elements, respectively, μr and kr at the rth row and rth column.
−1
We deduce that the product [ μ ]  K g  gives:
−1
 0  0  0
    =   = [Ω]
 μr  kr ωr
2
   
 0    0    0 
the spectral matrix.

Thus, the space model of the system described by [M] and [K] is
 0
 
converted into a modal model described by [ Ω] =  ωr2  and
0 

[Q] the basis change matrix leading to the modal basis.
NOTE.– Because the form vectors {X r } are defined with an arbitrary

magnitude, the matrices [ μ ] and  K g  are obtained to the nearest
scale factor and are therefore not unique (they are said to be
“generalized”).
However, their product [μ ] −1[K g ] gives a single matrix:

 0  0
 ωr2   μr 
  , because by inverting   the scale factor is
 0   0 
 0
 
canceled out by that of  kr  when their product is calculated.
 0 
3.2.2.2.2. Eigen movement

We can determine the variations with respect to time of the
physical parameters, xs, describing the motion of the structure, as a
linear combination of N coordinates (said to be generalized), the
variations of which are uncoupled. The method known as modal
synthesis or modal superposition consists of:
– set: {x} = [ Q ] {y}
in the matrix writing of (Σ) movement equations of the free system:
[M ]{ x ′′} + [K ]{ x} = {0}
– multiply the left-hand side by [ Q ] .
T
We thus obtain:
[ Q ] [ M ][ Q ]{y ′′ }+ [ Q ] [ K ][ Q ]{ y }= { 0 }
T T
which corresponds to:
[μ ]{ y′′} +
[K ]{ y} = {0}
g
By multiplying, on the left-hand side, each member of this

−1
equation by [ μ ] , we obtain the relationship matrix:
 0
(σ ) {y ′′ } +   y = 0
ω { } { }
2
 r
 0 
(σ ) is a differential system with yr uncoupled unknowns (the

generalized coordinates), of which the rth equation:
y ′r′ +ω r2 y r = 0
admits as solution:
y r = A r cosω r t + B r sinω r t = C r sin ( ω r t + ϕ r ) .
The constants A and B or C and are to be determined from

the initial conditions.
We deduce therefrom the expression with respect to time of the N

physical unknowns, xs, of the system Σ : ( )
N N
xs (t ) =  X sr y r (t ) =  X sr Cr sin (ω r t + ϕ r )
r =1 r =1 (r = 1, 2, ..., N)
where:
1
⋮ = is the rth column of [Q]
We can see that the variations in the structure’s displacements are

combinations of the N eigen modes.
3.2.3. Damped free system
3.2.3.1. Modal damping
The modeling [ M] { x′′} + [ K ] { x} = { 0} does not represent the

real systems perfectly.
The addition of a viscous damping (resistant force proportional to

the derivative of the parameters [velocity]) significantly improves the
model.
The simplest method is to use modal damping. This involves a

term representing an energy dissipation, which is of the form:
2δ n ω n y n′ (t )
to be added to the equations of the system (σ ) (concerning the y n

parameters in the modal basis). The term δ n is dimensionless and is
known as the modal damping factor. Its value, between 0 and 1, is
given by experience: based on the measurements of the real system’s
response, we estimate the value of δ n .
Generally, δ n has a value of some % (0 < n ≤ 0.05), unless a

hydraulic damper is present in the system studied, or the structure
contains viscoelastic materials.
Thus, the n e equation of the system (σ ) becomes:
y n′′ (t ) + 2δ n ω n yn′ (t ) + ωn2 yn (t ) = 0

n = 1,2, …, N
and its solution, as for 1 DOF, takes the form:
y n (t ) = An e −δ n ω nt sin (Ω n t + ϕ n )
2
where Ωn = ωn 1 − δ n
and An , ϕ n are arbitrary constants to be determined from the initial

conditions.
y n ( 0 ) = y n0 y n′ (0) = y n′ 0
( ) (
 y n′ + δ nω n y n 2 + y n Ω n 2 
An =  0
) 1/ 2
0
2
0

 Ωn 
y′n 0 + δn ωn y n 0
cos φn = 1/ 2
,
(
 y′ + δ ω y
) ( ) 
2 2
+ yn0 Ωn
 n 0 n n n0

y n0 Ωn
sin φn = 1/ 2
(
 y′ + δ ω y
) ( ) 
2 2
+ yn0 Ωn
 n 0 n n n0

The variations of the physical coordinates x1 , x2 , ... , x N are then

determined from the relation {x} = [ Q ] {y}.
3.2.3.2. Proportional damping

The projection of the movement equations in the modal basis can
be conducted directly onto the unexcited system:
[ ]{ x′′} + [C ]{ x′} + [K ]{ x} = {0}

(S) M
if the damping matrix [C] is a linear combination of the mass and

stiffness matrices:
[C] = β [ M] + γ [ K]
where β and γ are real constants. This form of damping is known as
proportional damping.
In the case of the 2-DOF system (S):
m1 0   x1′′  c11 −c12   x1′   k11 −k12   x1  0

0 + + =
 m2   x2′′  −c12 c22   x2′  −k12 k 22   x2  0
the condition for the symmetric matrix [C] to be decomposed into a

diagonal matrix and a symmetric matrix is:
m k c
det m k c =0
0 k c
Whereby (S) becomes:
[ M ]{x ′′} + ( β [ M ] + γ [ K ]) {x′} + [ K ]{ x} = { 0}

By setting {x} = [Q ]{y} and multiplying, on the left-hand side, by
[Q ]T , we have the system
( Σ ) : [μ ]{y ′′}+ (β [μ ] + γ [K ]){y ′}+ [K ]{ y} = {0}
g g
−1
and by multiplying, on the left-hand side, by [ μ ] , we obtain the
system
(σ ) : [I ]{y ′′}+ (β [I ] + γ [Ω]){y ′}+ [Ω ]{ y} = {0}

composed of uncoupled linear differential equations concerning the

modal parameters y n of the form
( )
y ′n′ + β + γ ω n2 y n′ +ω n2 y n = 0
n = 1, 2, …, N
If we set: β + γ ω n2 = 2 δ n ω n
β γ ωn
i.e. δn = + < 1 for a lightly damped motion,
2ωn 2
we present within the system (σ ) a damping factor for each mode,

depending on the mode’s eigen pulsation.
Thus, the modal equations of the system (σ ) take the same form
yn′′ + 2 δ n ω n yn′ + ω n2 yn = 0 as those obtained by adding a modal
damping to the uncoupled equations of the conservative system.
They give the same solution
y n (t ) = An e −δ n ωnt sin(Ω n t + ϕ n )
which leads to the expression of the physical coordinates with respect

to time:
{x} = [Q ]{y}
Hence the displacement of the ith coordinate:
xi =  X in yn =  X in An e −δ nωnt sin(Ω n t + ϕ n )
n n
which is a linear combination of all the eigen modes.

3.3. Response on the time domain of an excited system
Let us consider a system with proportional viscous damping
[C] = β [ M] + γ [ K]
of which each mass, mi , is excited by a force, fi (i = 1,2, … , N).
The movements are governed by the differential system
(S): [ M]{ x′′} + [ C] { x′} + [ K ] { x} = { f }
By setting { x} = [Q]{ y} and multiplying, on the left-hand side,
by [Q ] , we have the system

T
( Σ ) : [ μ ]{ y′′} + ( β [ μ ] + γ  K g  ) { y′} +  K g  { y} = [Q ]T { f }
and by multiplying, on the left-hand side, by [ μ ]−1 , we obtain the

system
(σ ) : [ I ]{ y′′} + ( β [ I ] + γ [Ω ]) { y′} + [Ω ] { y} = [ μ ]−1 [Q ]T { f } = {F }

composed of uncoupled linear differential equations concerning the
modal parameters y n of the form:
(
yn′′ + β + γ ω n2 ) yn′ + ωn2 yn = Fn n = 1, 2, …, N
where the elements Fn of the matrix of outside forces { F } are linear

combinations of all the forces fi applied to the masses. In the case of
a 2-DOF system subject to two sinusoidal excitations (f1 and f2) in
phase, we can adjust their amplitudes so as to cancel out one of the
modal forces (F1 or F2) and thus excite a single mode (modal
appropriation). This technique is very popular in experimental

mechanics.
The nth uncoupled equation of (σ ) is written as

y ′n′ + 2δ n ω n y n′ + ω n2 y n = Fn
by setting: β + γ ω n2 = 2 δ n ω n to present the modal damping.
The solution of each movement equation is that of 1 DOF excited

by the force Fn :
e − δ nω n t t
yn = d n e −δ nωn t
sin (Ω n t + φn ) +  F (u )e
n
−δ ωn nu
sin Ω n (t − u )du
Ωn 0
with Ω n = ω n 1 − δ n . This expression includes the eigen movement

2
due to the initial conditions (constants dn and φn ) .
Finally, the physical parameters are obtained from the relation

{ x}= [Q] { y}
This explicitly gives, for 2 DOF:
 x = X 11 y1 + X 12 y 2
⎯→  1
⎯
 x 2 = X 21 y1 + X 22 y 2
3.4. Exercises
3.4.1. Exercise 1: eigen modes of a system with 2 DOF
The system mentioned below is composed of a frame, a cart (1)

and a pendulum (2). We wish to study the small movements about the
stable equilibrium position of the set, where the pendulum remains
vertical with its center of mass, G2, below the center of the pivot link,
O2, and where the springs have their natural length.
Figure 3.2. Configuration of the system of two solids
The cart (1), with center of mass G1 and mass m1, is in translation

along B1 x .
  
The pendulum (2), to which we attach the frame (O 2 x 2 , y 2 , z ) , is
⎯⎯→ 
of mass m2, center of mass G2, defined by G 2 O 2 =  y 2 , and moment

of inertia m2 r 2 with respect to G 2 z , where r is the radius of gyration.

The pendulum is in pivot link of axis O 2 z with respect to the frame.
The cart is linked to the frame by a tension/compression spring

B1B2 of stiffness k1 and at A2 to the rigid rod of the pendulum by a
 ⎯⎯→  
traction/compression spring A1A2 of stiffness k2  A 2 O 2 = a y 2  .
 
The position of the set is situated with respect to the Galilean

  
( )
frame, to which the reference frame B1 x, y, z is attached, by: x1
(the displacement of the cart (1) from its equilibrium position), and
   
α = (x , x 2 ) = (y , y 2 ) (the angle between the rod of the pendulum (2)
and the vertical).
At equilibrium, x1 = 0 and α = 0. During the movements, α (in

radians) remains small and at the first order the displacement, x2, of
the point A2 is horizontal and equal to aα.
 
The gravity field is defined by: g = −g y. The links are perfect.
Data: g = 10 m/s2 k1 = 8,400 N/m k2 = 24,000 N/m

=1m a = 0.5 m r = 0.1 m
m1 = 100 kg m2 = 10 kg
1) Write the equations of the small movements of the set.

2) a) Deduce the differential system with variables x1 and x2 = aα.
b) Give the literal expression, then the numerical values of the
mass [ M ] and stiffness [ K ] matrices (we will overlook r2 with
respect to 2).
c) Calculate the eigen pulsations of the system and compare

them to those of the cart alone and the pendulum alone.
d) Determine the corresponding form vectors.
Solution to Exercise 1

1) The dynamic resultant for the cart (1) in translation along x is
written as m 1 x 1′′

The efforts applied to the cart along x are:
– the force of spring B1B2: −k1x1
– the force of spring A1A2: − k 2 (x 1 − aα ) at the first order.

Thus, its movement equation is: m1x1′′ = − k1x1 − k 2 (x 1 − aα )

The moment of inertia along the O2z axis of the pendulum (2) is
given as
(
IO 2 Z (2) = IG 2 Z (2) + m22 = m2 r 2 + 2 )

Then its dynamic moment about the O2 z axis is written as
( )
m 2 r 2 +  2 α ′′
The efforts applied to the pendulum are:


• the torque along O2 z due to gravity:
  
−  y 2 ∧ (− m 2 g y ) = − m 2 g sin α z

i.e. at the first order: − m 2 gα z

• the torque along O 2 z due to the spring A1A2:
  
− ay 2 ∧ [− k 2 (a α − x 1 )x ] = − k 2 a (a α − x 1 ) cos α z

i.e. at the first order: − k 2 a ( aα − x1 ) z

Thus, the moment equation along O 2 z is written as
( )
m 2 r 2 +  2 α ′′ = − m 2 gα − k 2 a (aα − x 1 )
2) a) By setting x 2 = aα , dividing the moment equation by a and

overlooking r2 with respect to 2, we obtain the differential system of
the coupled variations of x1 and x2:
 m1x1′′ + (k1 + k 2 ) x1 − k 2 x 2 = 0
 2
m   m g 
x ′′ − k 2 x1 +  22 + k 2  x 2 = 0
 2 a 2 2  a 
b) Hence the mass matrix:
 m1 0 
 100 0  (kg)
[ M] =  2  =  

0 m2 2
  0 40 
a 
and the stiffness matrix:
 k1 + k 2 −k 2 
[ K ] =  m 2 g  =  32, 400 −24,000 (N/m)
−k 2 + k 2   −24,000 24, 400 
 a 2 
c) The eigen pulsations of the system are the roots of:

2
K −ω M = 0
32, 400 − 100ω 2 −24,000

i.e.
−24,000 24, 400 − 40ω 2
( )
= 4,000 ω 4 − 934ω 2 + 53,640 = 0
Hence: ω1 = 7.84 rad/s and ω2 = 29.5 rad/s.
If the pendulum was locked in its vertical position (α = 0 ) , the

eigen pulsation of the cart would be:
k1 + k 2
Ω1 = = 18 rad/s, under the action of two springs.
m1
If the cart was locked in the equilibrium position of the set

( 1 = 0) , the pulsation of the pendulum would be (we can overlook r2
x
with respect to 2):
ℓ
= (
= 24.57 rad/s, under the action of gravity and
ℓ )
of the spring A1A2.
d) From K − ωi2 M  {Xi } = {0} , i = 1 or 2, we determine the two

 
forms of vectors that define the end positions of the two solids:
( 32, 400 − 100ω ) q i

2
1i − 24,000 q 2i = 0 x1 =
1
1.09
1
and x =
−2.28
We deduce that at the first mode, at pulsation 7.84 rad/s, if the

displacement amplitude of the cart is 1 cm along x (this depends on
the initial conditions), the point A2 of the pendulum moves 1.09 cm,
that is, the pendulum deviates by 2 × 0.0109 rad (1.25°) to the right of
the vertical.
At the second mode, at pulsation 29.5 rad/s, if the cart moves 1 cm

 
along x , the point A2 moves 2.28 cm along − x and the angle formed
by the pendulum with the vertical is –0.0456 rad (–2.61°) measured on

z.
3.4.2. Exercise 2: free and forced oscillations of a

conservative 2-DOF system
2 DOF SYSTEM
x1 x2
k10 k12 k20 

x
m1 m2
f1 (t ) f 2 (t )
Figure 3.3. Schematic diagram of a system with 2 DOF

1) Note (S ) : [M ]{ x ′′}+ [K ]{ x } = { f (t )} the differential system

governing the variations of x1 and x2 with respect to t.
Using this matrix form, write the movement equations of the


masses m1 and m2, in translation along x (perfect links), as shown in
Figure 3.3.
2) We will begin by studying the free system (eigen movement).

Data are as follows:
m 1 = 1 (kg ), k10 = 3 (N/m)
m 2 = 4 (kg ), k12 = 4 (N/m)
k 20 = 12 (N/m)
a) Explain the matrices [M] and [K].

b) Calculate the eigen pulsations and the form vectors of the
system. Give the matrix [Q ] of basis change to the modal basis.
c) Calculate the generalized mass and stiffness matrices and find
the eigen pulsations using these two matrices.
d) Give the relations between the physical coordinates x1 and x2
and the uncoupled modal coordinates y1 and y2. Write the equations
(σ) governing the variations of y1 and y2 with respect to t and give
their solution.
e) Determine x1(t) and x2(t) with the initial conditions:
x (0) = 0, x (0) = 0.025(m), x (0) = x (0) = 0
3) Let us consider the above system excited by the forces:
f1 (t ) = 5sin t (N ), f 2 (t ) = 0
and with zero initial conditions (at rest at t = 0)
Determine the system’s response in the time domain: x 1 (t ), x 2 (t ).


1) The equation of the dynamic resultant along x for each of the
solids of mass m1 and m2 gives the movement equations:
m 1x 1′′ = − k 10 x 1 − k 12 (x 1 − x 2 ) + f1 (t )
m 2 x ′2′ = − k 12 (x 2 − x 1 ) − k 20 x 2 + f 2 (t )
which are written in matrix form:
 m1 0   x1′′   k10 + k12 −k12   x1   f1 ( t ) 

(S)   + = 
k12 + k 20   x 2   f 2 ( t ) 
 0 m 2   x ′′2   − k12
2) a) With the given numerical values, the mass and stiffness

matrices are:
1 0   7 −4 
[M] =   and [ K ] =  
0 4   −4 16 
b) The natural pulsations of the (conservative) system are

obtained from:
det −ω 2 [ M ] + [ K ] = 0
This explicitly gives a zero determinant
7 −ω2 −4
=0
−4 16 − 4ω 2
for the values of ω 2 roots of the equation
(
4ω 4 − 44ω 2 + 96 = 4 ω 2 − 3 ω 2 − 8 = 0)( )
2
ω1 2 = 3 and ω 2 = 8
The eigen pulsations of the system studied are therefore ω1 = 3

and ω2 = 8 rad/s.
 X1j 
The form vectors X j =  { }  are determined from:
X 2 j 
[− ω 1
2
[M ] + [K ]]
X 11  0
=  X 11 = X 21 { X 1 } =  
1
 X 21  0 1
[− ω 2
2
[M ] + [K ]]
X 12  0 
=  − 4 X 22 = X 21 { X 2 } = 
− 4

 X 22  0  1 
Thus, the matrix [Q ] of basis change to the modal basis is:

1 − 4
[Q] =  
1 1 
c) The generalized mass and stiffness matrices are:
  1 1  1 0 1 − 4 5 0 
[μ ] = [Q ]T [M ][Q ] =    0 4 1 1  = 0 20 
  − 4 1     
[K ] = [Q] [K ][Q]
g
T
  1 1   7 − 4 1 − 4 15 0 
=    = 
 − 4 1  − 4 16  1 1   0 160
3 0 
We verify that [μ ] K g = 
−1
[ ] 
0 8 
d) The relation between physical and modal coordinates
x1 = y1 − 4 y 2
{x} = [Q ]{y} gives
x2 = y1 + y 2
This change of variables in (S) leads to two uncoupled equations:
(σ ) y1′′ + 3 y1 = 0 , the solution of which is: (

y1 = A1 sin 3t + ϕ1 )
y2′′ + 8 y2 = 0, the solution of which is: y 2 = A2 sin ( 8t + ϕ 2 )
e) Returning to the physical coordinates, we obtain the
displacements of the two masses with respect to time:
x1 = A1 sin ( )
3t + ϕ1 − 4 A2 sin( 8t + ϕ ) 2
x2 = A sin (
1 3t + ϕ ) + A sin ( 8t + ϕ )
1 2 2
Using the initial conditions provided, we determine the arbitrary

constants:
0= sin − 4 sin
sin =4 sin = 0.02
0.025 = sin + sin
0 = 3 A1 cos ϕ1 − 4 8 A2 cos ϕ 2 
 A1 cos ϕ1 = A2 cos ϕ 2 = 0
0 = 3 A1 cos ϕ1 + 8 A2 cos ϕ 2 
hence: x = 0.02 (cos √3 t − cos √8t)

x = 0.02 (cos √3 t + cos √8t)
X1
0 ,0 5
0 ,0 4
0 ,0 3
0 ,0 2
0 ,0 1
- 0 ,0 1
0 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 1 5 1 6 1 7 1 8 1 9 2 0 t
- 0 ,0 2
- 0 ,0 3
- 0 ,0 4
- 0 ,0 5
Figure 3.4. Variations of x1 with respect to time for the first 20 sec
x2
0.03
0.02
0.01
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
-0.01
-0.02
-0.03
Figure 3.5. Variations of x2 with respect to time
3) The system governing the variations of the modal coordinates of

the excited conservative system is:
(σ ) : {y ′′} + [Ω ]{y} = [μ ] −1 [Q ]T { f (t )} = { F (t )}
5 sin t   F1 
Thus, the modal forces are: [μ ]
−1
[Q ]T  = 
 0   F2 
1 
5 0  1 1 5 sin t   sin t   F1 
 1 − 4 1  0  = − sin t  =  F 
0         2
 20 
and the system (σ ) is formed of two uncoupled equations:
 y1′′ + 3 y1 = sin t

 y 2′′ + 8 y 2 = − sin t
Note that exciting a single mass results in two modes being

excited.
Given that the system is excited from rest (equilibrium of the set),
the solution of each equation of (σ ) is of the form:
1 t
yn =  Fn ( u ) sin ωn ( t − u ) du (see Volume 3 of this set)
ωn 0
1 t
i.e. explicitly: y1 =  sin u sin 3 ( t − u ) du …
3 0
…=
2
1
3 0
{cos[(
t
) ] [( )
3 + 1 u − 3 t − cos 1 − 3 u + 3 t du … ]}
…= [ ]
1  sin ( 3 + 1) u − 3 t sin (1 − 3 ) u + 3 t 
 − 
[ ] t
…
2 3  3 +1 1− 3  0
…
1  sin t sin t sin 3 t sin 3 t  …
=  + + + 
2 3  3 +1 3 −1 3 +1 1− 3 
1 1
y1 = sin t − sin 3 t
2 2 3
1 t
y2 =  − sin u sin 8 ( t − u ) du …
8 0
1  sin t sin t sin 8 t sin 8 t 

… =−  + + + 
2 8  8 +1 8 −1 8 +1 1− 8 
1 1
y2 = − sin t + sin 8 t
7 7 8
Thus, the response in the time domain of the excited system is:
15 1 2
x1 = y1 − 4 y 2 = sin t − sin 3 t − sin 8 t
14 2 3 7
5 1 1
x2 = y1 + y 2 = sin t − sin 3 t + sin 8 t
14 2 3 7 8
3.4.3. Exercise 3: calculation/test correlation
A structure admits 2 DOF. A finite element calculation has given

the following matrices:
 4 0
mass [M] =   (in kg)
0 9
12 −6 
stiffness [K] =   (in N/m)
 −6 27 
1) Deduce its natural pulsations ω1 and ω2 (ω1 < ω2).
2) Tests on the real structure have shown that the amplitude

resonances are obtained at the pulsations:
42 14
Ω1 = rad/s and Ω 2 = rad/s
5 2
We wish to model the structure with a proportional viscous

damping.
a) Deduce the modal damping factors δ1 and δ2 of two modes.
b) For the structure considered, deduce the damping matrix [C]
corresponding to this type of damping.
1) The eigen pulsations are obtained from: K − ω 2 M = 0 , with

the conservative system data provided by the finite element
calculation.
12 − 4ω 2 −6
Thus, in this case we have: =0
−6 27 − 9ω 2
(
i.e. 12 − 4ω 2 )( 27 − 9ω ) − 36 = 0
2
 
(3 − ω ) ( )( 2 − ω ) = 0
2
2
or 36  − 1 = 36 4 − ω 2 2
 
Hence: ω1 = 2 (rad/s) and ω 2 = 2 (rad/s)
2) The purpose of the tests is to deduce the dampings from the real
structure.
a) At the resonance of each mode, we have: Ω n = ωn 1 − 2δ n2 ,

n = 1 or 2.
2 1  Ω n2 
Then from: δ n = 1 − 
2  ωn2 
we deduce two damping factors:
 42 
1   2 √2
δ 12 = 1 − 25  = 1 = = 0.282
2 2  25 5
 
 
 14 
1  1 1
δ 2 = 1 − 4  =
2
2 = = 0.25
2 4  16 4
 
 
b) For a proportional damping, the damping matrix is a linear
combination of the mass and stiffness matrices:
[C ] = β [M ] + γ [K ]
For a system with 2 DOF, the coefficients β and γ can be
determined from the relation:
β + γωn2 = 2δ nωn
2 4
which, for the first mode, is written as β + 2γ = 2 x x 2=
5 5
and for the second mode:
1
β + 4γ = 2x x2 = 1
4
From these two equations, we deduce:
1 3
γ = and β =
10 5
Thus,
3 1 1  36 −6 
[C] = [ M ] + [ K ] = (Ns/m)
5 10 10  −6 81 
is the damping matrix completing the mathematical model of the

structure studied.
3.4.4. Exercise 4: damped system with a single excited

mode
A structure admitting 2 DOF x1 and x2 is modeled, as shown in

Figure 3.6, with the following characteristics:
f1 (t ) f 2 (t )
k1 k2

m1 m2 x
c1 c2
x1 x2
Figure 3.6. Schematic diagram of a damped system with 2 DOF
m1 = 9 kg, k1 = 24 N/m, c1 = 8 N/m/s,

m 2 = 1 kg, k 2 = 3 N/m, c 2 = 1 N/m/s,
1) Write the movement equations of the set. The links that ensure

the translation along x are perfect.
2) Let us consider the free, undamped system.
a) Determine the mass matrix [ M ] and the stiffness matrix [ K ] .
b) Determine the eigen pulsations ω1 and ω2, as well as the

matrix of basis change to the modal basis [Q].
c) By projecting onto the modal basis, give the differential
system with uncoupled modal variables. Solve these equations.
d) Deduce the variations of x1 and x2 with respect to t.
3) Note that we have a proportional damping.

a) Express the damping matrix as a linear combination of the
mass and stiffness matrices.
b) Give the spectral matrix [ Ω] and calculate the damping

factors δ 1 and δ 2 .
c) Write the differential system with uncoupled modal

variables.
d) Calculate the eigen pulsations of the damped structure and
give the solution of the above system.
Deduce the expression of the physical coordinates x1 and x2 with

respect to t.
4) For f1 = 3sin t (N) and f2 = sin t (N), write the structure’s

response to this excitation, with the structure at rest prior to
application of the excitatory forces.
1) The equation of the dynamic resultant for each solid, projected

along the x axis, gives the movement equations of the coupled
system:
m 1 x 1′′ = − k 1 x 1 − k 2 (x 1 − x 2 ) − c1 x 1′ − c 2 (x 1′ − x ′2 ) + f1 (t )
m 2 x ′2′ = k 2 (x 1 − x 2 ) + c 2 (x 1′ − x ′2 ) + f 2 (t )
which can be written in the matrix form as
m 0   x1′′  c1 + c 2 − c 2   x 1′   k1 + k 2 − k 2   x 1   f1 
(S) :  1 + + =
0 m 2   x ′2′   − c 2 c 2   x ′2   − k 2 k 2   x 2  f 2 
2) a) The undamped, unexcited system:
(S1 ) [M ]{ x ′′}+ [K ]{ x } = {0}

9 0 
has the mass matrix [M] =   and the stiffness matrix
0 1 
− 3
[K] = 
27

− 3 3  .
b) The eigen pulsations of the undamped system ω1 and ω2 are

the roots of:
[K ] − ω 2 [I ] =0
27 − 9ω 2 −3
= (
= 9 3−ω2 ) 2
( )(
−9 = 9 4 −ω2 2 −ω2 )
−3 3−ω2
i.e. ω12 = 2 and ω22 = 4, hence ω1 = 2 and ω2 = 2 ( ω n in rad/s).
 X 1n 
The (eigen) vectors of form { X n}=   , which compose the
 X 2n 
basis change matrix to modal basis, are determined from the first
( )
equation (or from the second) of: [K ] − ω 2 n [M ] { X n } = { 0}
i.e. for ω12 = 2 :
 9 − 3  X 11  0 1
− 3 1   X  = 0 → 3X11 = X 21 → { X1} = 3
   21   
and for ω12 = 4 :
− 9 − 3  X 12  0
 − 3 − 1  X  =  0 
   22   
1
→ −3 X 12 = X 22 → { X 2} =  
 −3
Thus, the matrix of basis change to the modal basis is:
[Q] = 
1 1
.
3 − 3
1 1   y1   y1 + y 2 
c) By setting { x} = [Q ]{ y} = 
  =   , in
3 − 3  y 2  3 y1 − 3 y 2 
the system (S1) and by multiplying by [Q ] on the left-hand side, we
T
obtain the system (Σ1 ) with modal variables y1 and y2:
(Σ1 ) : [μ ]{ y ′′} + [K g ]{ y} = {0}

where [ μ ] is the generalized mass matrix, obtained by:
1 3  9 0 1 1  18 0 
[μ ] = [Q ]T [M ][Q ] =    = 
1 −3 0 1  3 −3  0 18
Then, by multiplying the left-hand side of this relation by
1 1 0  ,
[ μ ]−1 = we obtain the system (σ1 ) : { y′′} +
18 0 1 
[Ω]{ y} = { 0} , which is composed of the equations:
y1′′ + 2 y1 = 0
y 2′′ + 4 y 2 = 0
the solutions of which are:
y1 = A1 sin ( 2 t + ϕ1 )
y 2 = A2 sin (2t + ϕ 2 )
d) From { x} = [Q ]{y}, we deduce:

x1 = y1 + y 2 = A1 sin ( )
2 t + ϕ1 + A2 sin (2t + ϕ 2 )
[
x 2 = 3 y1 − 3 y 2 = 3 A1 sin ( )
2 t + ϕ1 − A2 sin (2t + ϕ 2 ) ]
3) The damped system (S 2 ) : [M ]{x ′′}+ [C ]{x ′}+ [K ]{x } = {0} has the
 9 − 1
damping matrix: [C] =  
− 1 1 
a) By noting [ C] = β [ M ] + γ [ K ] , we have: = 0 and = ,

that is to say, the value of the damping matrix is 1/3 of the stiffness
matrix (proportional damping).
1
b) From the relation β + γ ωn2 = 2δ n ωn = ωn2 with
3
 ω1 = 2
[Ω] = 
2 0 1
 , we deduce:  δ1 = , which is
 2
 0 4 2δ1ω1 =
3 2
 3
 ω2 = 2 1
the damping factor of the first mode, and  δ2 = ,
 4 3
2δ 2ω2 = 3
which is the damping factor of the second mode.
c) By carrying these results over into (S2) it becomes:
1 ′ 
[ M ]{x ′′} + [ K ]  {x } + {x}  = {0} .
3 
With the same transformation from physical coordinates to modal

coordinates as for the undamped system, we obtain the system
(σ 2 ) : {y′′} + [Ω] 1 {y′} + {y} = {0}

3 
the uncoupled equations of which are:
 ′′ 2 ′
 y1 + 3 y1 + 2 y1 = 0
 4
 y2′′ + y2′ + 4 y2 = 0
 3
of the form:
y n′′ + 2δ nω n y ′n + ω n2 y n = 0
d) The eigen pulsations of the damped structure are:
√17
Ω = 1− = = 1.37 rad/s
3
4√2
Ω = 1− = = 1.88 rad/s
3
Thus, the solution of (σ 2 ) is:
−
t
 17 
y1 = A1e 3 sin  t + ϕ1 
 3 
−
2t
4 2 
y 2 = A2 e 3
sin  t + ϕ 2 
 3 
and the solution of (S2) is:
−
t
 17  −
2t
4 2 
x1 = y1 + y 2 = A1e 3 sin  t + ϕ1  + A2 e 3 sin  t + ϕ 
3  3 2 
   
−
t
 17  −
2t
4 2 
x2 = 3 y1 − 3 y 2 = 3 A1e 3
sin  t + ϕ1  − 3 A2 e 3 sin  t + ϕ 2 

 3   3 
The constants A 1 , ϕ 1 , A 2 , ϕ 2 will be determined from the given

initial conditions.
4) By conducting the same step to change to modal basis as for the

free system, the excited system
(S): [M ]{ x ′′}+ [C ]{ x ′}+ [K ]{ x } = { f (t )}
transforms to the system (σ ) : { y′′} + [Ω] { y′}+ { y} = { F (t )}

1
3 
with { F } = [μ ] −1 [Q ] T { f (t )}
1 1 0 1 3  3 sin t 
=
18 0 1  1 − 3  sin t 
1 sin t 
= 
3  0 
The uncoupled equations of (σ ) are:
 ′′ 2 ′ 1
 y1 + 3 y1 + 2 y1 = 3 sin t
 4
 y ′2′ + y ′2′ + 4 y 2 = 0
 3
Note that only the modal coordinate, y1, is excited.
Thus, as the system is excited from rest, the zero initial conditions:
x 1 (0 ) = x 1′ (0 ) = x 2 (0 ) = x ′2 (0 ) = 0
lead to: y1 (0 ) = y1′ (0 ) = y 2 (0 ) = y 2′ (0 ) = 0 and y 2 (t ) = 0 , ∀t ,

which means that the system oscillates according to the first eigen
mode of pulsation 17 / 3 rad/s, superposed onto the excited motion.
Let us look for the particular solution of the first equation of (σ )

in the form:
y1 p = B1 sin t + B2 cos t
By carrying over into the first equation of (σ ) , we have ∀t :
2
− B1 sin t − B2 cos t + (B1 cos t − B2 sin t ) + 2(B1 sin t + B2 cos t ) = 1 sin t
3 3
and by identifying the coefficients
2 1
(of sin t): − B1 − B2 + 2 B1 =
3 3
(of cos t): − B 2 + 2 B1 + 2 B 2 = 0

3
3 2
we deduce: B1 = and B2 = −
13 13
Hence the expression of:
−  17
t
 3 2
y1 = A1e sin 
3
t + ϕ1  + sin t − cos t
 3  13 13
Taking into account the initial conditions:
2
y1 (0 ) = 0 = A1 sin ϕ1 −
13
1 17 3
y1′ (0) = 0 = − A1 sin ϕ1 + A1 cos ϕ1 +
3 3 13
2 7
we obtain: 𝐴1 sin 𝜑1 = 13 and 𝐴1 cos 𝜑1 = −
13 17
Thus, the variations of the modal coordinate, y1, are:
1 −𝑡 17 7 17 3 2
𝑦1 = 𝑒 3 2 cos 𝑡− sin 𝑡 + sin 𝑡 − cos 𝑡
13 3 17 3 13 13
𝑖𝑛 𝑡𝑟𝑎𝑛𝑠𝑖𝑒𝑛𝑡 𝑠𝑡𝑎𝑡𝑒
steady state
Finally: x1 = y1 and x2 = 3y1
The term containing the negative exponential decreases rapidly

(after several oscillations) over time.
It is only significant for a transient period.
It, therefore, becomes negligible with respect to the terms sint and
cost (time period 6.28 sec) that dominate the displacements of the
system of two solids (steady state).
x2
Transient Steady state
Period: 6.28 (s)
Figure 3.7. Variations of x2 with respect to time

3.4.5. Exercise 5: system excited by the base
1) Write the movement equations of the system below, excited in

displacement by a simple harmonic function:
k10 k12 k23
m1 m2 x (t ) = x 0 sin ω t
c10 c12 c23

x 1 (t ) x 2 (t )
Figure 3.8. Schematic diagram of a system with 2 DOF excited by the base
 x1′′   x1′   x1 
2) We set: [ M ]   + [ C]   + [ K ]   = { f ( t )}
 x ′′2   x ′2  x2 
Explain the matrices [M ] , [C] , [K ] and { f (t )} .
3) Numerical application:
2
m1 = 2 (kg) c10 = 2 (Ns/m) k10 = 4 (N/m) x0 = (m)
7
m2 = 8 (kg) c12 = 1 (Ns/m) k12 = 4 (N/m) ω=7 (rad/s)
c23 = 11 (Ns/m) k23 = 28 (N/m)
a) Explain the above matrices numerically.

b) Determine the natural pulsations of the system and the
associated eigen vectors.
c) After noting that we have a proportional damping, determine
the damping factors and the eigen pulsations of two modes of the
damped system.
d) Write the equations governing the variations of the uncoupled
modal variables.
e) Write the modal excitation function in the form:
F = A sin (ω t + θ ) .
f) Deduce the values of A and θ (in radians).

g) Determine the expression with respect to time of the modal
variables in the system’s response to excitation (the eigen movement
having disappeared).
h) Deduce the expressions of x1(t) and x2(t) in the steady state.
Plot their variations with respect to time for t ∈ [1 , 6] , assuming that
the steady state is established.
1) The movement equations of the solids of mass m1 and m2,
respectively, are:
m 1x 1′′ = − k 10 x 1 − k 12 (x 1 − x 2 ) − c10 x 1′ − c12 (x 1′ − x ′2 )
m 2 x ′2′ = − k 12 (x 2 − x 1 ) − k 23 (x 2 − x ) − c12 (x ′2 − x 1′ ) − c 23 (x ′2 − x ′)
2) Noting this differential system in the form:
[M ]{ x ′′}+ [C ]{ x ′}+ [K ]{ x } = {f (t )}
we have:
m 0  c + c12 −c12 
[ M ] =  01 m 2 
, [C] =  10−c c12 + c23 
,
  12
k10 + k12 − k12 

[K] = 
 − k12 k12 + k 23 
 0 
and { f ( t )} =   , which represents the
 k 23 x 0 sin ω t + c23 x 0ω cos ω t 
system’s excitation force.
3) a) With the numerical values, these matrices become:

 3 − 1 − 4
[M ] = 
2 0
[K] = 
8
 , [C] =  , ,
 0 8 − 1 12   − 4 32 
 0
{f (t )} = 
0
= 
 2 (4 sin 7t + 11cos 7 t ) f 2 
b) The eigen pulsations are the roots of:
[K ] − ω 2 [M ] =0
8 − 2ω 2 −4
( ) ( )( )
2
i.e. = 16 4 − ω 2 − 16 = 16 5 − ω 2 3 − ω 2 = 0 ,
2
−4 32 − 8ω
hence:
1 = √3 = 1.732
(in rad/s)
2 = √5 = 2.236
The eigen vectors are determined from one or other equation of the
relations:
{ X1} = 
2
[K − 3M ]{ X1} = {0} ⎯
⎯→ X 11 = 2 X 21 ⎯
⎯→ 
1
2
[K − 5M ]{ X 2 } = {0} ⎯
⎯→ X 12 = −2 X 22 ⎯→ { X 2 } =  
⎯
− 1
Thus, the basis change matrix to modal basis is:
[Q ] = 
22

1 − 1
c) Here, the damping matrix is the following linear combination

of the mass and stiffness matrices: [ C] = β [ M ] +
1 1
γ [K ] = [M ] + [K ]
2 4
By setting 2δ nωn = β + γωn2 , we obtain the damping factors of the

two modes:
1 1 5 5
2 3δ1 = + x3 = , hence 1 = = 0.36
2 4 4 8√ 3
1 1 7 7
2 5δ 2 = + x5 = , hence 2 = = 0.39
2 4 4 8√5
We can now determine the eigen pulsations of the damped system.
25 √167
= 1− = √3 1 − = = 1.615 rad/s
64 × 3 8
49 √271
= 1− = √5 1 − = = 2.057 rad/s
64 × 5 8
d) The modal-variable equations are of the form:
y n′′ + 2δ n ω n y ′n + ω n2 y n = Fn
The generalized mass matrix has the following value:
2 1   2 0   2 2  16 0 
[μ ] = [Q ]T [M ][Q ] = 2 =
−1  0 8  1 −1  0 16 

Thus, the modal efforts are given as
 F1  1 1 0   2 1   0  f 2  1 
 F  = [μ ] [Q ] { f (t )} = 16 0 1 2 − 1  f  = 16 − 1
−1 T
 2    2   
with
f 2 = 2 (4 sin 7t + 11 cos 7t )
hence, explicitly:
 ′′ 5 ′ f2
 y1 + 4 y1 + 3 y1 = 16
(σ )  7 f
 y2′′ + y′2 + 5 y2 = − 2
 4 16
f2 2
e) The modal excitation force F = = ( 4sin 7t + 11cos7t )
16 16
can also be written as
F = A sin (7t + θ ) = A(cos θ sin 7t + sin θ cos 7t )
with:
4 4
cosθ = =
2
4 + 11 2 137
11
and sin θ = , θ = 70° or 1.22 rad
137
274
and A =
16
√
then = sin(7 + 1.22)
f) The response to excitation A sin (ω t + θ ) in the steady state is

written as
y n = Bn sin(ω t + θ + ϕ n ) , n = 1 or 2
where:
A / ω n2
Bn = ,
Dn
ω2 ω
1− − 2δ n
ω n2 ωn
cos ϕ n = and sinϕ n =
Dn Dn
For the system (σ ) , with second members F and – F, we have the

denominators:
7 5 7
D = 1− + 2 = 1− + 2
3 8√3 √3
46.82
=
3
7 7 7
D = 1− + 2 = 1− + 2
5 8√5 √5
45.67
=
5
and the response:
√
= sin(7 + + ) = 0.0221sin (7 + + ) (m)
.
√
and =
.
sin(7 + + ) = −0.0226sin (7 + + ) (m)
with φ = −2.95 rad (–169.26°) and φ = −2.87 rad

(–164.45°)
because:
7
1−
= 3 = −46 = −0.982
D 46.82
and
sin ϕ 1 < 0
= = = −0.963 and sin ϕ 2 < 0

.
g) From { x} = [Q ]{ y}, we deduce the displacements of the

masses m1 and m2 in the steady state:
= 2( + ) = 2 0.0221 sin(7 + + )−
0.0226 sin(7 + + ) (m)
= − = 0.0221 sin(7 + + ) + 0.0226 sin(7 + +

) (m)
with:
+ = −1.73 rad and + = −1.65 rad
we have:
1 = 0.0014 cos 7 − 0.0034 sin 7
2 = −0.0443 cos 7 − 0.0053 sin 7

h)
xi(t)
0,05
0,04 x2(t)
0,03
0,02 x1(t)
0,01
0 t
-0,01 1 1,5 2 2,5 3 3,5 4 4,5 5 5,5 6
-0,02
-0,03
-0,04
-0,05
Figure 3.9. Displacements of the two solids excited by the base

4
Modal Analysis of
N Degrees of Freedom
4.1. Introduction
The vibratory behavior of a deformable structure or a set of solids

linked by elastic links depends on the mechanical characteristics of the
system and the amplitude and frequency of excitation.
The forced vibrations of a structure under a simple harmonic

excitation take place at the excitation frequency. Should this
frequency coincide with that of an eigen mode, the structure is in
resonance. The effect of adding damping to its modeling is that it
reduces the amplitudes and creates a phase shift in the structure’s
response. In the steady state, the role of the excitation frequency is
complex but fundamental. In this chapter, we look to obtain the
response of an excited system with respect to the excitation frequency.
4.1.1. Normal modes
As we saw in the previous chapter, a free, undamped structure in

motion admitting N degrees of freedom has N natural frequencies and
there is a natural vibration state corresponding to each of them, the
displacements of which (translations or rotations) constitute a normal
mode. The normal modes of vibration depend solely on the mass and
stiffness values and their distributions over the structure. A structure
can vibrate at a normal mode if subject to suitable initial conditions.

But when the structure is subject to arbitrary initial conditions,

such as in the case of an impulse, the free vibrations that follow are
composed of all the normal modes.
Let us start by considering an undamped system, because in the

case of proportional damping, the normal modes of the conservative
system also define the modal base of the (free or excited) damped
system.
4.2. Response in the frequency domain of a conservative

structure subjected to a harmonic excitation
Consider a structure, the movement equations of which form the

system:
[M ]{x′′(t )} + [K ]{x(t )} = { f (t )}
where [M ] is the mass matrix and [K ] is the stiffness matrix. In the
linear modeling of a structure admitting N degrees of freedom
(N ≥ 2 ) , these two matrices are square, symmetric and positive
defined, of constant and real elements. { x} is the column matrix
representing the displacements (N physical coordinates) of the
structure.
Let {f ( t ) }= {F }e ω i t
be the matrix representing the excitation
force vector.
The response of the system will be of the form:
{x ( t ) }= {X }e ω
i t
where {F} and {X} are matrices (N × 1) (representing vectors) with

elements in ℂ, independent of t.
These expressions carried over into the movement equations
[M ]{ x ′′}+ [K ]{ x } = {f (t )}
Modal Analysis of N Degrees of Freedom 147
enable the time parameter to be eliminated
([ K ] − ω [ M ] ){X }= {F }
2
The frequency response of the structure is given by
{X } = ( [ K ] − ω [ M ] ) { F }
2 −1
which can be written as
{ X} = [α (ω )]{ F}
with [α ( ω ) ] = ([ K ]− ω [ M ] )
2 −1
being the receptance matrix (N × N), representing the structure’s

response model in displacement, with the general term:
( ) is the coef icient of in the expression of
of =
Fk
α (ω ) Xj
Figure 4.1. Schematic diagram of excitation and response of the system
Put more simply, α jk represents the response of the structure’s jth

coordinate, in the case where it is excited by the force Fk only,
applied to its kth coordinate:
X = F if F = 0, ∀m ≠ k, therefore
m ∈ l, … , k − 1, k + 1, … , N
In order to explain the elements of the receptance matrix, let us

invert the last matrix relation, then multiply each member of the
left-hand side of the equality by [P ] and the right-hand side by [P ] .
T
We obtain
[P]T ([K ] − ω 2 [M ])[P] = [P]T [α (ω )] −1[P]

where [P] = [Q] μ [ ] −1 / 2
is the matrix of the normalized form vectors
(with respect to the modal mass).
 0
 
[ ]
With μ −1 / 2 =
1
 the rth column of [P ] is
 μr 
 0 
 P1r   q 1r 
 .   . 
  1   1
 . =  . = {X r }
 .  μ r  .  μ r
P  q 
 Nr   Nr 
[P]T [M ][P] = ([μ ]−1 / 2 ) [Q]T [M ][Q][μ ] −1 / 2 = [I]

T
Given that
 0
and that [P] [K ][P] = [μ ]
T −1 / 2
[K ][μ ]
g
−1 / 2
= 

ω 2
r


 0 
 0
we have  
 = [ P ] [α ( ω )] [P ]
T −1
 ω r2 −ω 2
0  

By multiplying this relation by ([P] )

T −1
on the left-hand side, and
by [P] on the right-hand side, then inverting, we obtain:
−1
−1
 0

[α ( ω ) ] = [ P ]  ω r2 − ω 2  [P ] T
0
  
NOTE.–
The general element αjr of the product [P][ ] is: α = ∑ P
but = ( : the Kronecker symbol)
1
If [Ω ] is a diagonal matrix of elements Ω r = ,
ω 2
r −ω 2
then = Ω .
Thus, the general element
of [P] [Ω] [P] is: = P ΩP =
and the product [ P ][ Ω ][ P ] T

leads to a symmetric matrix.
This reciprocity phenomenon for the receptance matrix [α ] can be

applied to several of the structures’ characteristics.
 Xj  X 
Here: α jk =   = α kj =  k 
 F 
 Fk  Fm = 0, ∀m≠k  j  Fm = 0, ∀m ≠ j
This indicates that we have the same response from the structure
by measuring X j and by exciting at the coordinate k, as by
measuring X k with the same excitation applied at the coordinate j.
N Pjr Pkr N A jk
Thus: α jk (ω ) =  = r
r =1 ω −ω
2
r
2
r =1 ω −ω2
2
r
where r A jk is the modal constant, which, for the rth mode, links the j
and k coordinates.
α jk ( ω ) is the sum of the rational fractions:
A jk A jk A jk
α jk (ω ) = 1
+ 2
+ ... + N
ω −ω
2
1
2
ω −ω
2
2
2
ω −ω2
2
N
whose numerators r A jk are independent of ω (also known as the

residuals) and for which the denominators are canceled out at the
poles, which are the structure’s eigen pulsations.
Thus, from { X} = [α (ω )]{ F} we obtain the displacements of the

structure with respect to the excitatory forces:
N N N p jr p kr
X j =  α jk Fk =  ω Fk
k =1 k =1 r =1
2
r −ω2
X j = α j1F1 + α j2 F2 + . . . + α jN FN
N p jr p1r N p jr p Nr
Xj =  F1 + . . . +  FN
r =1 ω −ω
2
r
2
r =1 ωr2 − ω 2
We can see that by exciting a structure at a single point (i.e. if only

one of the Fk values is non-null), we excite all of the eigen modes of
the structure and they contribute to obtaining the variations as a
function of the excitation frequency of a coordinate, Xj.
4.3. Response of a structure with proportional viscous

damping to a harmonic excitation
Let us imagine the particular case where the damping matrix is

expressed as a linear combination of the structure’s mass and stiffness
matrices.
[C] = β [M ] + γ [K ]
The general condition in order to have a proportional damping is:
([M ] −1
[K ])([M ]−1[C]) = ([M ]−1[C])([M ]−1[K ])
As a result, the similar matrix of [C] in the modal basis will be a
diagonal matrix.
The movement equations: [M ]{x"}+ [C ]{x ′}+ [K ]{x } = {f (t )}
for a harmonic response in displacement {X}eiω t to the harmonic

excitation {F}eiω t become:
(− ω [M] + iω [C] + [K]){X} = {F}

2
( ( ) )
i.e. (1 + iγ ω ) [K] + iβ ω − ω 2 [M] {X} = {F}
The structure’s receptance matrix is:
[α (ω )] = [(1 + iγ ω )[K ] + (iβ ω − ω 2 )[M ]]−1

Let [Q] be the matrix constituted by the form vectors {X r }of the
undamped system, then we have:
 0  0
[Q] [C][Q] = β  μ r  + γ  k r  = [Cg ]
T   
 0   0 
the generalized damping matrix and:
 0
  = [C ]
[P] [C][P] = β [I ] + γ  ω r
T 2
 n
 0 
the normalized damping matrix with respect to the modal mass,

because [P] = [Q][μ ]
−1 / 2
is the matrix of the normalized form vectors
with respect to the modal mass.
Thus:
[α ]−1 = [K ] + iω [C] − ω 2 [M ]
becomes:
 0
[P]T [α ]−1 [P] = (iβω − ω )[I] + (1 + iγω )  ω r
2  2 

 0 
 0
=  ω r − ω + i(β + γω r )ω
2 2 2 
 
 0 
which is the diagonal matrix of elements:
(1 + iγ ω ) ωr2 + (iβω − ω 2 ) = ωr2 − ω 2 + 2iδ rωrω

by setting: + =2 where is the damping factor of the
rth mode.
Thus, the receptance matrix:
 0
 1  T
[α ] = [P]   [P]
 ω r − ω + 2iδ rω rω
2 2

 0 
is the symmetric matrix, which has the following general element:

N Pjr Pkr
α jk (ω ) = 
r =1 ω −ω + 2iδ rω rω
2
r
2
which has complex poles (roots of denominators of fractions):
ω = iδ r ω r ± ω r 1 − δ r2
that we can also write in the form: iω = −δ rω r  iω r 1 − δ r

2
In the case of a proportional damping, the damped system has the

same form vectors as the system without damping and the components
of these vectors are real.
The modulus of the (complex) rth fraction of α jk is:
Pjr Pkr 1
= Pjr Pkr
ω − ω + 2iδ rω rω (ω − ω 2 ) + (2δ rω rω )
2 2 2
2 2
r
r
and it reaches its maximum value for ω = ω r 1 − 2δ r = ω R : the

2
amplitude resonance pulsation of the rth mode.
4.4. Frequency response of a structure with proportional

hysteretic damping
The hysteretic damping of a structure, in the movement

equation [M ]{x"}+ ([K ] + i[D ]){x} = {f (t )} , is represented by the
matrix [D], which, in the case of a proportional damping, is of the
form:
[D ] = β [M ] + γ [K ]
By also using here the normalized form vectors of the undamped
system, composing the matrix [P ], we have:
 0

[P] [D][P] = β [I] + γ  ω r
T 2 

 0 
Thus,
[P]T ([K ] − ω 2 [M ] + i[D])[P] =

 0
(iβ − ω )[I] + (1 + iγ ) 
2
ω r2 

 0 
 0
=  ω r − ω + i (β + γω r )
2 2 2 
 
 0 
 0
= ω r2 − ω 2 + iηrω r2 
 
 0 
β
by setting ηr = γ + : damping loss factor.
ω r2
Thus, the receptance matrix [α (ω )] has the following general

element:
N Pjr Pkr
α jk (ω ) = 
r =1 ω − ω 2 + iηrω r2
2
r
Hence, the displacements of the structure:

N N N Pjr Pkr
X j =  α jk Fk =  ω Fk
k =1 k =1 r =1
2
r − ω 2 + iη r ωr2
4.5. Exercises
4.5.1. Exercise 1: receptance matrix of a conservative

structure
A structure with 2 DOF is modeled by the system:
x1 x2
k1 k2 k3 
x
m1 m2
f1 ( t ) f2 ( t )
Figure 4.2. Schematic diagram of a conservative system with 2 DOF

The links are perfect and the displacements x1 and x2 along x are
measured from equilibrium. The spatial characteristics of the system
are:
m1 = 1 kg, m2 = 1 kg, k1 = k3 = 0.4 MN/m, k2 = 0.8 MN/m
I) Consider the free motions with the initial conditions:
x1 (0) = x10 , x 2 (0) = x 20 , x1′ (0) = 0, x′2 (0) = 0
1) Write the movement equations.

2) We wish to find solutions in the form: { x} = { X}eiω t
a) Determine the eigen pulsations of the system.
b) Determine the form vectors {X1} and {X2}. Give the

modal matrix [Q].
c) Determine the normalized mass and stiffness matrices.

d) Solve the modal model.

e) Deduce therefrom the variations of the physical coordinates
as a function of time, in the case where the initial conditions are:
e.1) x10 = 1 cm e.2) x10 = 1 cm e.3) x10 = 1 cm
x 20 = 1 cm x 20 = −1 cm x 20 = 0
II) Consider the forced vibrations of the system with f =

F e (k = 1 or 2) and the harmonic response of the system:
x j = X je iω t (j = 1 or 2).
1) Determine the receptance matrix [α (ω )] using:

k1, k2, k3, m1, m2 and ω
2) Give the numerical values of the elements α jk of this matrix.

Decompose the α jk values into simple elements in ω 2 .
3) Express the X j values with the help of the Fk and ω values.
4) What do X1 and X2 become when ω is close to:
a) 2 x103 rad / s
b) √0.4x10 rad/s
I) 1) The free movement equations is:
m1x1′′ + (k1 + k 2 )x1 − k 2 x 2 = 0

m 2 x ′2′ + (k 2 + k 3 )x 2 − k 2 x1 = 0
or in the matrix form:
 m1 0   x1′′   k1 + k 2 − k 2   x 1  0
+ =
0
 m 2   x ′2′   − k 2 k 2 + k 3   x 2  0
[M]{ x′′} + [K]{ x} = { 0 }

 1 0 1.2 −0.8
with: [M ] =   (kg) and [ K ] = (MN/m)
0 1 −0.8 1.2
2) By carrying over: { x} = {X}eiω t , we deduce:

a) the eigen pulsations of the structure are the roots of:
k1 + k 2 − m1ω 2 − k2
=0
− k2 k 2 + k 3 − m 2ω 2
i.e.: m1m2ω 4 − [(m1 + m2 ) k 2 + m1k3 + m2 k1 ]ω 2 + k1k 2 + k1k 3 + k 2 k3 = 0
With the numerical values ω4 – 2.4 × 106ω2 + 0.8 × 1012 = 0 we

obtain the squares of the eigen pulsations ω12 = 4x105 (rad / s) 2
and ω22 = 2x106 (rad / s)
2
{ X r } = 
q1r 
b) The form vectors  , r = 1 or 2, are deduced
q 2 r 
from the first relation of:
( [K] − ω [M] ){X } = {0},

2
r r
( )
i.e.: k1 + k 2 − ω r2 m1 q1r = k 2 q 2 r
1
Hence, for r = 1: q11 = q 21 and { X1} =  
1
1
and for r = 2: − q12 = q 22 and { X 2 } =  
 − 1
1 1 
Thus, [Q] =   is the basis change matrix to the modal basis.
1 − 1
c) The generalized mass matrix is:
[Q] T [M ][Q] = [μ ] =   (kg) and the generalized stiffness matrix:

2 0
 0 2
[Q] T [K ][Q] = K g = 
0,8 0 6
[ ]  10 (N/m)
 0 4
 x1   y1   x1 = y1 + y 2
d) By setting:  x  = [Q]  y  , that is to say, 
 2  2 x 2 = y1 − y 2
the modal-variable equations are:
y1′′ + ω12 y1 = 0
y′2′ + ω 22 y 2 = 0
with:
= √0.4 × 10 = 632.4 rad/s
= √2 × 10 = 1,414 rad/s
They admit as solutions:
y1 = A1 sin(ω1t + ϕ1 )
y 2 = A 2 sin(ω 2 t + ϕ 2 )
e) The displacements of the masses m1 and m2 with respect

to time are:
x1 = A1 sin(ω1t + ϕ1 ) + A2 sin(ω2 t + ϕ 2 )
x 2 = A1 sin(ω1t + ϕ1 ) − A2 sin(ω2 t + ϕ 2 )
With the initial conditions: x1′ (0) = x′2 (0) = 0 , we have

A1 cosϕ1 = A 2 cosϕ 2 = 0 , then:
e.1) from x1 (0) = A1 sinϕ1 + A2 sinϕ 2 = 1

A1 sinϕ1 = 1
x 2 (0) = A1 sinϕ1 − A2 sinϕ 2 = 1 A2 sinϕ 2 = 0
we obtain: x1 (t ) = x 2 (t ) = cosω1t (cm)
with ω1 = 632.4 rad/s.
The system oscillates in phase, according to the first (symmetric)

mode.
k1 k2 k3
m1 m2
Figure 4.3. Symmetric oscillations of the system with 2 DOF
e.2) With: x1 (0) = A1 sinϕ1 + A2 sinϕ 2 = 1 A1 sinϕ1 = 0

x 2 (0) = A1 sinϕ1 − A 2 sinϕ 2 = −1 A2 sinϕ 2 = 1
we have: x1 (t ) = −x 2 (t ) = cosω 2 t (cm)
with ω2 = 1.414 rad /s.

The system oscillates in phase opposition, according to the second

(antisymmetric) mode.
k1 k2 k3
m1 m2
Figure 4.4. Antisymmetric oscillations of the system with 2 DOF
e.3) With: x1 (0) = A1 sinϕ1 + A2 sinϕ 2 = 1 A1 sinϕ1 = 1 / 2
x 2 (0) = A1 sinϕ1 − A 2 sinϕ 2 = 0 A 2 sinϕ 2 = 1 / 2
we have: x1 (t ) =
1
(cosω1t + cosω2 t ) = cos ω1 + ω 2 t  cos ω 2 − ω1 t 
2  2   2 
(in cm)
and x 2 (t ) =
1
(cos ω1t − cos ω 2 t ) = sin ω1 + ω 2 t  sin ω 2 − ω1 t  (in cm)
2  2   2 
The system is “beating”: the movements of the two masses have

the same pseudo-periods
2π 2π
T1 = = = 6 ms
Ω1 ω1 + ω 2
2
2π 2π
and T2 = = = 16 ms
Ω2 ω 2 − ω1
2
T2
but the mass m2 is offset by with respect to the mass m1.
4
II) The forced vibrations of the system are governed by the

equations
m1x1′′ + (k1 + k 2 )x1 − k 2 x 2 = f1 (t )

m 2 x ′2′ + (k 2 + k 3 )x 2 − k 2 x1 = f 2 (t )
1) With x j = X je iω t and f k = Fkeiω t , j, k ∈ {1, 2 }
( )
we have: k1 + k 2 − ω 2 m1 X1 − k 2 X 2 = F1
( )
− k 2 X1 + k 2 + k3 − ω 2 m2 X 2 = F2
or, in matrix form:
 k1 + k 2 − ω 2 m1 − k2   X1   F1 
   =
2 F 
 − k2 k 2 + k 3 − ω m 2  X 2   2
For this example, let us conduct the calculations analytically. By

[ ]
inverting the matrix K − ω 2 M , we obtain the receptance matrix:
X F
= [ ( )]
X F
1 k +k − m k F
=
det[K − M] k k +k − m F
Thus:
k2 + k3 − ω 2m2 X1
α11 = = if F2 = 0
F1
K − ω 2M
k1 + k 2 − ω 2 m1 X1
α 22 = = if F2 = 0
F1
K − ω 2M
k2
α12 = = α 21
K − ω 2M
X X
= if F = 0, or = if F = 0
F F
2) With the numerical values, we obtain the determinant:
|K − M| = 1.2 x 10 − − 0.8 x 10
−0.8 x 10 1.2 x 10 −
= − 2.4 x 10 + 0.8 x 10
and the receptance matrix elements:
1.2 x 10 −
=
− 2.4 x 10 + 0.8 x 10
−0.5 −0.5
= + =
− 0.4 x 10 − 2 x 10
(for this example)
0.8 x 10
=
− 2.4 x 10 + 0.8 x 10
−0.5 −0.5
= + =
− 0.4 x 10 − 2 x 10
(always).
NOTE.– We could obtain these results directly from the relation:
Pj1Pk1 Pj2 Pk 2
α jk = 2 2
+
ω −ω
1 ω 22 − ω 2
−1/2 1 1 1 
where Pjr is the element of the matrix [ P ] = [ Q][ μ ] =   of
2 1 −1
change to the normalized modal basis.
α11 = α 22
0.0002
0.00015
0.0001
0.00005
ω (rad / s )
0
0 200 400 600 800 1000 1200 1400
-0.00005
-0.0001
Figure 4.5. Variation of elements and of the receptance matrix

α12 = α 21
0.0002
0.00015
0.0001
0.00005
0
0 200 400 600 800 1000 1200 1400
ω (rad / s )
-0.00005
-0.0001
Figure 4.6. Variation of elements and of the receptance matrix
The graphical representation of an element of the receptance

matrix with respect to the pulsation of the excitation ω enables us to
deduce the values of the eigen pulsations of the system (peaks).
3) The amplitudes of the displacements with respect to the

amplitudes of the forces applied are:
0.5(F + F ) 0.5(F − F )
X = F + F =− −
− 0.4 x 10 − 2 x 10
0.5(F + F ) 0.5(F − F )
X = F + F =− +
− 0.4 x 10 − 2 x 10
. ( )
4) When ω is close to √0.4 x 10 rad/s: X ≅ X ≅
.
as the second fraction becomes negligible in relation to the first. It

can be said that only the first mode is excited: the two masses oscillate
in phase (like the free system with the first initial conditions case).
This is also indicated by the equal components of the first form
vector.
When ω is close to 2x103 rad / s : X1 ≅ − X 2 ≅ − 02,5(F1 − F62 )

ω − 2 x10
At this pulsation, the second mode is excited: the two masses

oscillate in phase opposition (like the free system in the second initial
conditions case). Its form vector does have opposite components.
4.5.2. Exercise 2: receptance matrix of a structure with

proportional viscous damping
Consider the system with 2 DOF presented in Figure 4.7.
f1 (t) f2 (t)
k10 k12
m1 m2
c10 c12 c20

x1 (t ) x 2 (t )
Figure 4.7. Schematic diagram of a damped system with 2 DOF
1) Write the movement equations of the system situated by x1 and

x2 from its equilibrium position (the links are perfect).
2) Let us continue with the following data:
m1 = 2 (kg) c10 = 1.6 (Ns/ m) k10 = 12 ( N / m)
m2 = 5 ( kg) c12 = 1 (Ns / m) k12 = 10 ( N / m)
c20 = 1 (Ns / m)
Write the mass [M], stiffness [K] and damping [C] matrix of the
system.
Express [C] as a linear combination of [M] and [K].
3) Determine the eigen pulsations and the form vectors of the

undamped system.
Write the matrix [Q] composed by the form vectors.
 
4) Determine the generalized mass [μ ] and stiffness  K g 
 
matrices. Calculate [μ ]
-1
[K ].
g
5) Determine the (viscous) damping factor δ r of each mode.
6) Detail the elements α jk (ω ) of the receptance matrix [α (ω )],

and the system’s response to the excitation forces f1(t) and f2(t).
During an experimental test, how can we measure each term of the
receptance matrix?
1) The movement equations are written in the matrix form:
 m1 0   x1′′  c10 + c12 − c12   x1′  k10 + k12 − k12   x1   f1 (t )

 0 m   x ′′  +  − c +
c12 + c 20  x ′2   − k12
=
k12  x 2  f 2 (t )
 2  2  12
2) Numerically, we have:
 22 − 10
[M ] = 
2 0 2.6 −1
 , [C] = , [K ] =  
 0 5 −1 2  − 10 10 
2,6 − 1 1 2 0 1  22 − 10
Note that [C] =  =  +   = β [M ] + γ [K ]
 − 1 2  5 0 5 10  − 10 10 
That is,
[C ] = β [M ] + γ [K ], we are looking here at proportional damping

with
1 1
= and =
5 10
3) The eigen pulsations of the undamped structure are the positive
roots of:
22 − 2ω 2
[K] − ω 2 [M ] =
− 10
(
= 10 ω 4 − 13ω 2 + 12 )
− 10 10 − 5ω 2
i.e.: ω1 = 1 rad / s and = √12 = 3.46 rad/s
{ X r } = 
q1r 
The corresponding form vectors  are determined
q 2 r 
from:
{X1} = 
r =1 1
 ⎯⎯→
⎯ 2q11 = q 21 
2
[K − ω M ]{X } = {0}
2
r r


{X 2 } = 
 r =2 5
 ⎯⎯⎯→ q12 = −5q 22 
 − 1
1 5 
Thus, the basis change matrix to the modal basis is [Q] =  
2 − 1
22 0 
4) [μ ] = [Q] T [M ][Q] =  
 0 55
0 
[K ] = [Q] [K ][Q] = 220
g
T
660

By calculating [μ ] -1 [K g ] for this example, we do find:
1 
 22 0  22 0  1 0  ω 2 0 
= = 1
 1   0 660 0 12  0 ω 2 
0   2
 55 
as we saw in the previous chapter, regardless of the magnitude chosen

for the form vectors { X r } .
5) As we have a proportional damping with: = and = , the

damping factors of the two modes are determined from the relation:
1 = 0.15 (15%)
= + = +
2 2 10 20 √
= 0.202(20.2%)
6) The matrix of the normalized form vectors (with respect to the

modal mass) is:
[P] = [Q][μ ] −1 / 2
 1   1 5 
0  
1 5   22
= = 22 55  =  p11 p12 

2 − 1  0 1   2 1   p 21 p 22 
−
 55   22 55 
The receptance matrix elements are then determined from the

relation:
2 p jr p kr
α jk (ω ) = 
r =1 ω − ω 2 + 2iδ r ω ω r
2
r
which is written for the structure studied:
p11 p11 p12 p12

α11 = +
ω − ω + i0,3ω1ω ω 2 − ω 2 + i 7 ω ω
2
1
2
2 2
5 12
1 5
= 22 + 11
1 − ω 2 + 0,3iω 12 − ω 2 + 1,4iω
p11 p 21 p12 p 22
α12 = +
ω − ω + i0,3ω1ω ω 2 − ω 2 + i 7 ω ω
2
1
2
2 2
5 12
1 1
−
= 11 + 11
1 − ω 2 + 0,3iω 12 − ω 2 + 1,4iω = α 21
= +
− + 0.3 7
− +
5√12
2 1
= 11 + 55
1− + 0.3i 12 − + 1.4i
As the receptance matrix connects the excitation to the response of
the structure:
 X1  α11 α12   F1 
 X  = α  
 2   12 α 22   F2 
we can obtain its components:

– α11 by measuring the response of the coordinate 1 and only
 X1 
exciting the mass 1 α11 = ;
 F1 
– α 22 from measuring the response of the coordinate 2 and only

 X 2 
exciting the mass 2 α 22 = ;
 F2 
– α12 = α 21 from the response of the coordinate 1, by exciting the

mass 2, or from the response of coordinate 2, by exciting the mass 1
 X X 
α12 = 1 = α 21 = 2  .
 F2 F1 
α11 0.2
0.15
0.1
0.05
0
0 1 2 3 4 5 6 7 8
Figure 4.8. Bode plot of the modulus of the frequency response function α11
Plotting the modulus of a frequency response function highlights

the pulsations of the damped system (vertexes) and indicates that the
second mode is more damped than the first. The half-power method
makes it possible to deduce the values of the two damping factors
from this curve (see Ewins (1995)).
ℑm
0 ℜe
-0.05 -0.025 0 0.025 0.05 0.075 0.1 0.125
-0.02
-0.04
-0.06
nd
2 mode
-0.08
1st mode
-0.1
ω R2
-0.12
-0.14
ωR1
-0.16
Figure 4.9. Nyquist diagram of α11 for ω ∈ [ 0, 4 ] rad/s

4.5.3. Exercise 3: case of a non-diagonal mass matrix
Linear modeling of a system with two degrees of freedomhas

2 1  5 − 4
provided the mass [M ] =   and stiffness [K ] =   matrices
1 2 − 4 5 
in SI units.
1) Calculate the eigen pulsations of the system ω1 and ω2 (> ω1).
2) Determine the corresponding form vectors: {X1} and {X2}.
3) Subsequently, to improve the model we add a proportional

1 1
damping of factors δ 1 = and δ 2 = in the first and second
3 3
modes, respectively.
Calculate the damping matrix [C] as a linear combination of
[M] and [K ] : [C] = β [M] + γ [K] .

4) Calculate the generalized mass, stiffness and damping matrices.
5) Give the matrix [P ] of basis change to the normalized modal

basis with respect to the mass.
6) In this section, we subject the structure to a harmonic excitation

and focus on the system’s frequency response.
Determine the receptance matrix elements [α (ω )] .
7) In the last section, we focus on the time domain response of the

damped structure to the excitation:
{ f (t )} = 
1
 sin 3t (in N)
 − 1
a) Write the differential equations governing the variations of

the (uncoupled) modal variables.
b) Give the solution of this system in steady state.

c) Deduce therefrom the variations with respect to time of the
physical variables representing the two degrees of freedom in the
steady state.
8) Compare the two forms of the structure’s response.
5 − 2ω 2 − 4 −ω2
1) [K] − ω [M ]
2
= (
= 5 − 2ω 2 ) − (4 + ω ) = (9 − ω )(1 − 3ω ) = 0
2 2 2 2 2
− 4 −ω2 5 − 2ω 2
1
Hence the natural pulsations: ω1 = = 0.577 (rad/s) and
3
= 3 (rad/s)
2) The corresponding form vectors are:
{ X1} =  
1
( [K] − ω [M] ){X } = {0}
2
1 1 ⎯
⎯→ q11 = q 21 ⎯⎯→
1
( [K] − ω [M] ){ X } = {0} ⎯⎯→

2
2 2 q12 = − q 22 ⎯⎯→ { X 2 } = 
1

 − 1
3) From β + γωr2 = 2δ rωr , we deduce the linear system with

unknowns β and γ:
for r = 1: β + γω12 = 2δ1ω1
for r = 2: β + γω22 = 2δ 2ω2 ,
the solution of which is:
2ω1ω 2 (δ 1ω 2 − δ 2ω1 ) 2(δ 2ω 2 − δ 1ω1 )

β= and γ =
ω 22 − ω12
ω 22 − ω12
With the numerical values for this case, we obtain
8 2
= =
13 13
Hence the damping matrix:
8 2 1 2  5 − 4 2 0
[C] = β [M ] + γ [K ] = + =
13 1 2 13  − 4 5  0 2
1 1 
4) With [Q ] =   , we have:
1 − 1
[μ ] = [Q] T [M ][Q] = 
6 0

0 2 
[K ] = [Q] [K ][Q] = 02

g
T 0
18

0
[C ] = [Q] [C][Q] = 40
g
T
4

We verify that we also have:
[C ] = β [μ ] + γ [K ]
g g
the same relation with the generalized matrices.
 1 1 
 6 2 
5) [P] = [Q][μ ] −1 / 2 =
 1 1 
−
 6 2 
6) The response of the structure in displacement is given by:

{ X} = [α ]{ F}
with [α ] the receptance matrix of elements:
2 Pjr Pkr
α jk = 
r =1 ω − ω 2 + i2δ rω rω
2
r
P11 P11 P12 P12 1/ 6 1/ 2

α11 = + = +
ω12 − ω 2 + i2δ 1ω1ω ω 22 − ω 2 + i 2δ 2ω 2ω 1 2 2 9 − ω 2
+ i 2ω
−ω + i ω
3 3
P11 P21 P12 P22 1/ 2 − 1/ 2

α12 = + = + = α 21
ω12 − ω 2 + i2δ 1ω1ω ω 22 − ω 2 + i2δ 2ω 2ω 1 − 3ω 2 + i2ω 9 − ω 2 + i2ω
P21 P21 P22 P22 1/ 2 1/ 2

α 22 = + 2 = +
ω − ω + i2δ 1ω1ω ω 2 − ω + i2δ 2ω 2ω 1 − 3ω + i2ω 9 − ω 2 + i2ω
2
1
2 2 2
7) To obtain the structure’s response in the time domain, let us set

the following in the movement equations:
{ x} = [Q]{ y}.
a) By multiplying by [Q] T on the left-hand side, we obtain the
uncoupled equations with modal variables y1 and y2:
[μ ]{ y′′}+ [ Cg ]{ y′}+ [ K g ]{ y} = [Q] T {f }

because the matrices [ ], and [ ]
are diagonal
6y1′′ + 4y1′ + 2y1 = 0
2y′2′ + 4y′2 + 18y 2 = 2 sin 3t
Note that only the second mode is excited, and the excitation has
the same pulsation as the structure’s second natural pulsation.
b) In the steady state (the eigen movement having

disappeared), we have: y1 = 0 and:
f0 1
y2 = sin (Ω t + ϕ ) ,
k  2 1/ 2
2 2
 Ω   Ω 
1 − 2  +  2δ 2  
 ω2   ω 2  
with Ω = ω 2 = 3 (rad/s)
π
and ϕ = − : the excitation and the response are in quadrature
2
π
(phase-lag response of ), we deduce:
2
1 1 π 1
y2 = sin  3t −  = − cos 3t
1
9 2 x x1  2 6
3
c) The physical coordinates are deduced from:
 x1   y1 
 x  = [Q]  y  → x1 = − x 2 = y 2
 2  2
1
x1 = − cos 3t
6
1
x 2 = cos 3t
6
8) To compare the two forms of the response, let us set F1 = 1 and

F2 = –1 in the harmonic response of the structure (6°). We obtain:
X1 = α11F1 + α12F2 X 2 = α 21F1 + α 22 F2

= α11 − α12 = α 21 − α 22
1 −1
= 2
=
9 − ω + i2ω 9 − ω 2 + i2ω
At the excitation of pulsation ω = 3 (rad/s), we have the following

response:
1 1 1
– at the first coordinate: X1 = = − i of modulus X1 =
6i 6 6
π
(displacement magnitude) and of phase ϕ1 = − (phase lag with
2
respect to excitation);
−1 1 1
– at the second coordinate: X 2 = = i of modulus X 2 =
6i 6 6
3π
(displacement magnitude) and of phase ϕ 2 = − (phase lag with
2
respect to excitation).
We find that only the second mode (of pulsation ω2 = 3 rad/s) is
excited and we obtain the same magnitudes and phase lags for the two
degrees of freedom as in the response in the time domain (7°).
References
Borel, M., Vénizélos, G. (2016). Movement Equations 1 – Location,

Kinematics and Kinetics. ISTE Ltd, London, and Wiley, New York.
Borel, M., Vénizélos, G. (2017a). Movement Equations 2 – Mathematical
and Methodological Supplements. ISTE Ltd, London, and Wiley, New
York.
Borel, M., Vénizélos, G. (2017b). Movement Equations 3 – Dynamics and
Fundamental Principle. ISTE Ltd, London, and Wiley, New York.
Borel, M., Vénizélos, G. (2018). Movement Equations 4 – Equilibriums and
Small Movements. ISTE Ltd, London, and Wiley, New York.
Ewins, D.J. (1995). Modal Testing: Theory and Practice. John Wiley &
Sons, New York.
Inman, D. (1994). Engineering Vibration. Prentice Hall, Pearson, London.
Thomson, W.T. (1993). Theory of Vibration with Applications, 4th ed.
Prentice Hall, London.
Weaver Jr., W., Timoshenko, S.P., Young, D.H. (1990). Vibrations Problems
in Engineering, 5th ed. John Wiley & Sons, New York.

Index
B, C, D inertia matrix, 36
initial conditions, 54
boundary conditions, 59
conservative system, 100
K, L, M
damper, 97
damping factor, 152 kinetic energy, 11
degree of freedom, 36 Lagrange equations, 40
link, 14
E, F mass matrix, 99
modal
effort
constant, 150
generator, 3
synthesis, 106
receiver, 4
eigen
N, O, P
mode, 48
pulsation, 101 normal mode, 145
energy-power equation, 42 oscillation, 42
equilibrium, 40 poles, 150
form vector, 101 power, 21
friction coefficient, 20 proportional damping, 109
fundamental principle of
dynamics, 39 R, S, V
receptance matrix, 147
G, H, I
residuals, 150
generalized coordinates, 106 resonance, 145
harmonic excitation, 151 rolling, 28
hysteretic damping, 153 set of solids, 1

sliding velocity, 20
solidifying parameters, 9
spatial model, 105
steady state, 145
stiffness, 97
velocity distributor, 37
vibration, 97
Other titles from
in
Mechanical Engineering and Solid Mechanics
2019
EL HAMI ABDELKHALAK, POUGNET PHILIPPE
Embedded Mechatronic Systems 1– Second Edition Revised and Updated
BAYLE Franck
Reliability of Maintained Systems Subjected to Wear Failure Mechanisms:
Theory and Applications
(Reliability of Multiphysical Systems Set – Volume 8)
LE VAN Anh, BOUZIDI Rabah
Lagrangian Mechanics: An Advanced Analytical Approach
MICHELITSCH Thomas, PÉREZ RIASCOS Alejandro, COLLET Bernard,
NOWAKOWSKI Andrzej, NICOLLEAU Franck
Fractional Dynamics on Networks and Lattices
SALENÇON Jean
Virtual Work Approach to Mechanical Modeling
Viscoelastic Modeling for Structural Analysis
2018
BOREL Michel, VÉNIZÉLOS Georges
Movement Equations 4: Equilibriums and Small Movements
(Non-deformable Solid Mechanics Set – Volume 4)
FROSSARD Etienne
Granular Geomaterials Dissipative Mechanics: Theory and Applications in
Civil Engineering
RADI Bouchaib, EL HAMI Abdelkhalak
Advanced Numerical Methods with Matlab® 1: Function Approximation
and System Resolution
(Mathematical and Mechanical Engineering SET – Volume 6)
Advanced Numerical Methods with Matlab® 2: Resolution of Nonlinear,
Differential and Partial Differential Equations
(Mathematical and Mechanical Engineering SET – Volume 7)
2017
Movement Equations 2: Mathematical and Methodological Supplements
Movement Equations 3: Dynamics and Fundamental Principle
BOUVET Christophe
Mechanics of Aeronautical Solids, Materials and Structures
Mechanics of Aeronautical Composite Materials
BRANCHERIE Delphine, FEISSEL Pierre, BOUVIER Salima,
IBRAHIMBEGOVIC Adnan
From Microstructure Investigations to Multiscale Modeling:
Bridging the Gap
CHEBEL-MORELLO Brigitte, NICOD Jean-Marc, VARNIER Christophe
From Prognostics and Health Systems Management to Predictive
Maintenance 2: Knowledge, Traceability and Decision
EL HAMI Abdelkhalak, RADI Bouchaib
Dynamics of Large Structures and Inverse Problems
(Mathematical and Mechanical Engineering Set – Volume 5)
Fluid-Structure Interactions and Uncertainties: Ansys and Fluent Tools
KHARMANDA Ghias, EL HAMI Abdelkhalak
Biomechanics: Optimization, Uncertainties and Reliability
LEDOUX Michel, EL HAMI Abdelkhalak
Compressible Flow Propulsion and Digital Approaches in Fluid Mechanics
Fluid Mechanics: Analytical Methods
MORI Yvon
Mechanical Vibrations: Applications to Equipment
2016
Movement Equations 1: Location, Kinematics and Kinetics
BOYARD Nicolas
Heat Transfer in Polymer Composite Materials
CARDON Alain, ITMI Mhamed
New Autonomous Systems
DAHOO Pierre Richard, POUGNET Philippe, EL HAMI Abdelkhalak
Nanometer-scale Defect Detection Using Polarized Light
DE SAXCÉ Géry, VALLÉE Claude
Galilean Mechanics and Thermodynamics of Continua
DORMIEUX Luc, KONDO Djimédo
Micromechanics of Fracture and Damage
(Micromechanics Set – Volume 1)
Stochastic Dynamics of Structures
GOURIVEAU Rafael, MEDJAHER Kamal, ZERHOUNI Noureddine
From Prognostics and Health Systems Management to Predictive
Maintenance 1: Monitoring and Prognostics
KHARMANDA Ghias, EL HAMI Abdelkhalak
Reliability in Biomechanics
(Reliability of Multiphysical Systems Set –Volume 3)
MOLIMARD Jérôme
Experimental Mechanics of Solids and Structures
RADI Bouchaib, EL HAMI Abdelkhalak
Material Forming Processes: Simulation, Drawing, Hydroforming and
Additive Manufacturing
2015
KARLIČIĆ Danilo, MURMU Tony, ADHIKARI Sondipon, MCCARTHY Michael
Non-local Structural Mechanics
SAB Karam, LEBÉE Arthur
Homogenization of Heterogeneous Thin and Thick Plates
2014
ATANACKOVIC M. Teodor, PILIPOVIC Stevan, STANKOVIC Bogoljub,
ZORICA Dusan
Fractional Calculus with Applications in Mechanics: Vibrations and
Diffusion Processes
Fractional Calculus with Applications in Mechanics: Wave Propagation,
Impact and Variational Principles
CIBLAC Thierry, MOREL Jean-Claude
Sustainable Masonry: Stability and Behavior of Structures
ILANKO Sinniah, MONTERRUBIO Luis E., MOCHIDA Yusuke
The Rayleigh−Ritz Method for Structural Analysis
LALANNE Christian
Mechanical Vibration and Shock Analysis – 5-volume series – 3rd edition
Sinusoidal Vibration – Volume 1
Mechanical Shock – Volume 2
Random Vibration – Volume 3
Fatigue Damage – Volume 4
Specification Development – Volume 5
LEMAIRE Maurice
Uncertainty and Mechanics
2013
ADHIKARI Sondipon
Structural Dynamic Analysis with Generalized Damping Models: Analysis
ADHIKARI Sondipon
Structural Dynamic Analysis with Generalized Damping Models:
Identification
BAILLY Patrice
Materials and Structures under Shock and Impact
BASTIEN Jérôme, BERNARDIN Frédéric, LAMARQUE Claude-Henri
Non-smooth Deterministic or Stochastic Discrete Dynamical Systems:
Applications to Models with Friction or Impact
Uncertainty and Optimization in Structural Mechanics
KIRILLOV Oleg N., PELINOVSKY Dmitry E.
Nonlinear Physical Systems: Spectral Analysis, Stability and Bifurcations
LUONGO Angelo, ZULLI Daniele
Mathematical Models of Beams and Cables
SALENÇON Jean
Yield Design
2012
DAVIM J. Paulo
Mechanical Engineering Education
DUPEUX Michel, BRACCINI Muriel
Mechanics of Solid Interfaces
ELISHAKOFF Isaac et al.
Carbon Nanotubes and Nanosensors: Vibration, Buckling
and Ballistic Impact
GRÉDIAC Michel, HILD François
Full-Field Measurements and Identification in Solid Mechanics
GROUS Ammar
Fracture Mechanics – 3-volume series
Analysis of Reliability and Quality Control – Volume 1
Applied Reliability – Volume 2
Applied Quality Control – Volume 3
RECHO Naman
Fracture Mechanics and Crack Growth
2011
KRYSINSKI Tomasz, MALBURET François
Mechanical Instability
SOUSTELLE Michel
An Introduction to Chemical Kinetics
2010
BREITKOPF Piotr, FILOMENO COELHO Rajan
Multidisciplinary Design Optimization in Computational Mechanics
DAVIM J. Paulo
Biotribolgy
PAULTRE Patrick
Dynamics of Structures
SOUSTELLE Michel
Handbook of Heterogenous Kinetics
2009
BERLIOZ Alain, TROMPETTE Philippe
Solid Mechanics using the Finite Element Method
LEMAIRE Maurice
Structural Reliability
2007
GIRARD Alain, ROY Nicolas
Structural Dynamics in Industry
GUINEBRETIÈRE René
X-ray Diffraction by Polycrystalline Materials
KRYSINSKI Tomasz, MALBURET François
Mechanical Vibrations
KUNDU Tribikram
Advanced Ultrasonic Methods for Material and Structure Inspection
SIH George C. et al.
Particle and Continuum Aspects of Mesomechanics

Movement Equations 5 Dynamics of Sets of Solids

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Movement Equations 5 Dynamics of Sets of Solids

Hochgeladen von

Copyright:

Verfügbare Formate

Movement Equations 5

Non-deformable Solid Mechanics Set

Dynamics of a Set of Solids

ISTE Ltd John Wiley & Sons, Inc.

© ISTE Ltd 2019

Library of Congress Control Number: 2019946982

British Library Cataloguing-in-Publication Data

Chapter 1. Set of Solids with Neither

Chapter 2. Vibration Mechanics of Systems of Solids . . . . . 35

Chapter 3. Vibrations with N Degrees of Freedom . . . . . . . 97

Chapter 4. Modal Analysis of N Degrees of Freedom . . . . . . 145

The authors, Georges Venizelos in Paris and Abdelkhalak El Hami

This book, Movement Equations 5, focuses on the dynamics of sets

Chapter 1 presents the direct application of the fundamental

to structural strains like tensile, compression, bending or torsion

Having applied the fundamental principle of dynamics to the

Chapter 3 presents the study of the vibrations of a rigid solid, those

The vibratory behavior of a deformable structure or of a set of

In Chapter 4, we look to obtain the response of an excited system

ψ ,θ ,ϕ Euler angles, respectively precession,

  velocity at instant t of the material

  acceleration at instant t of the

 rotation vector or rotation rate of

 drive velocity of material point M

Coriolis acceleration applied to the

d ( λ )  derivative in relation to time of the

  torsor characterized by its two

  scalar invariant of torsor {T } ,

inertial Coriolis torsor of the solid

partial rotation rate relative to the

component of the variable Qα ′ of

 canonical linear component of the

Eαβ canonical quadratic component of

 part of the resultant of known

 part of the resultant of known

 part of the moment of known

aαβ , bαβ , cαβ , dα ( t ) algebraic components of the

 couple of unknowns that make up

D ( s ) = L[ d ( t ) ] Laplace transform of the forced

F ( s ) = L[ f ( t ) ] Laplace transform of the forced

Set of Solids with Neither

The direct application of the fundamental principle of dynamics to

1.1. Identifying a chain of solids with neither loops nor

In a Galilean frame  g  , we consider a series of n solids (S1), (S2),

(S1) is linked to  g  and (S2) only;

(S2) is linked to (S1) and (S3) only;

Movement Equations 5: Dynamics of a Set of Solids,

(Si) is linked to (Si-1) and (Si +1) only;

(Sn) is linked to (Sn-1) only.

– “downstream of (Si)” refer to the subset of solids (Sj) such that

– “(Sj-1)” refer to the previous solid (Sj);

Each of the solids in the chain is framed and located analogously

1.2. Applying the fundamental principles of mechanics

The fundamental principles governing mechanics take on

Figure 1.1. Set of two disjoint solids

1.2.1. Principle of effort generators

Two effort generators and , considered disjoint, i.e. such that

Figure 1.2. Effort generation diagram

The efforts exerted by on are represented by the torsor

1.2.2. Principle of effort receivers

The effort generator acts simultaneously on the two effort receivers

Figure 1.3. Effort reception diagrams

The efforts exerted by on are represented by the torsor