PDF Stability of Axially Moving Materials Nikolay Banichuk Ebook Full Chapter
PDF Stability of Axially Moving Materials Nikolay Banichuk Ebook Full Chapter
PDF Stability of Axially Moving Materials Nikolay Banichuk Ebook Full Chapter
Nikolay Banichuk
Visit to download the full and correct content document:
https://textbookfull.com/product/stability-of-axially-moving-materials-nikolay-banichuk/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
https://textbookfull.com/product/optical-materials-
microstructuring-surfaces-with-off-electrode-plasma-1st-edition-
nikolay-l-kazanskiy/
https://textbookfull.com/product/elastomer-based-composite-
materials-mechanical-dynamic-and-microwave-properties-and-
engineering-applications-nikolay-dishovsky/
https://textbookfull.com/product/post-soviet-racisms-1st-edition-
nikolay-zakharov/
https://textbookfull.com/product/return-migration-to-afghanistan-
moving-back-or-moving-forward-1st-edition-marieke-van-houte/
Thermal and Time Stability of Amorphous Alloys 1st
Edition Cheretaeva
https://textbookfull.com/product/thermal-and-time-stability-of-
amorphous-alloys-1st-edition-cheretaeva/
https://textbookfull.com/product/stability-and-control-of-
nonlinear-time-varying-systems-shuli-guo/
https://textbookfull.com/product/ecological-networks-and-
territorial-systems-of-ecological-stability-laszlo-miklos/
https://textbookfull.com/product/the-stability-and-shelf-life-of-
food-second-edition-persis-subramaniam/
https://textbookfull.com/product/education-policy-reform-trends-
in-g20-members-1st-edition-nikolay-d-nikandrov-auth/
Solid Mechanics and Its Applications
Nikolay Banichuk
Alexander Barsuk
Juha Jeronen
Tero Tuovinen
Pekka Neittaanmäki
Stability
of Axially
Moving
Materials
Solid Mechanics and Its Applications
Volume 259
Founding Editor
G. M. L. Gladwell, University of Waterloo, Waterloo, ON, Canada
Series Editors
J. R. Barber, Department of Mechanical Engineering, University of Michigan,
Ann Arbor, MI, USA
Anders Klarbring, Mechanical Engineering, Linköping University, Linköping,
Sweden
The fundamental questions arising in mechanics are: Why?, How?, and How much?
The aim of this series is to provide lucid accounts written by authoritative
researchers giving vision and insight in answering these questions on the subject of
mechanics as it relates to solids. The scope of the series covers the entire spectrum
of solid mechanics. Thus it includes the foundation of mechanics; variational
formulations; computational mechanics; statics, kinematics and dynamics of rigid
and elastic bodies; vibrations of solids and structures; dynamical systems and
chaos; the theories of elasticity, plasticity and viscoelasticity; composite materials;
rods, beams, shells and membranes; structural control and stability; soils, rocks and
geomechanics; fracture; tribology; experimental mechanics; biomechanics and
machine design. The median level of presentation is the first year graduate student.
Some texts are monographs defining the current state of the field; others are
accessible to final year undergraduates; but essentially the emphasis is on
readability and clarity.
Springer and Professors Barber and Klarbring welcome book ideas from
authors. Potential authors who wish to submit a book proposal should contact
Dr. Mayra Castro, Senior Editor, Springer Heidelberg, Germany, e-mail:
[email protected]
Indexed by SCOPUS, Ei Compendex, EBSCO Discovery Service, OCLC,
ProQuest Summon, Google Scholar and SpringerLink.
Pekka Neittaanmäki
123
Nikolay Banichuk Alexander Barsuk
Russian Academy of Sciences State University of Moldova
Institute for Problems in Mechanics Chisinau, Moldova
Moscow, Russia
Tero Tuovinen
Juha Jeronen Faculty of Information Technology
Faculty of Information Technology University of Jyväskylä
University of Jyväskylä Jyväskylä, Finland
Jyväskylä, Finland
Pekka Neittaanmäki
Faculty of Information Technology
University of Jyväskylä
Jyväskylä, Finland
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
v
vi Preface
of the fourth-order polynomial equation are presented in detail. This has applica-
tions in the stability analysis of one-dimensional differential equation models whose
characteristic equations reduce into polynomials.
Chapter 3 deals with non-conservative dynamic systems with a finite number of
degrees of freedom. The system is investigated under small perturbations, charac-
terized by a small parameter. The characteristic polynomial and a series expansion
with respect to the small parameter are used for the eigenvalues and eigenvectors to
evaluate the critical stability parameter and to study ideal and destabilizing per-
turbations. Sufficient conditions for stability are obtained and described. Some
examples of ideal perturbations and the structure of the corresponding perturbation
matrices are considered. The stability of the systems subjected to deficient pertur-
bations is also investigated, and the determination of the deficiency index is pre-
sented. Ziegler’s double pendulum is considered as a classical example of a
non-conservative system. We give a detailed exposition, deriving the governing
equations starting from the principle of virtual work, analyze the stability of the
system, and consider some special cases. We present numerical simulations using
new visualization techniques, in which the bifurcation behavior can be seen.
Chapter 4 deals with the general methods of bifurcation analysis, applied to
continuous systems, and some methods of optimization of the critical stability
parameter. We will briefly introduce the different types of stability loss and then
look at conditions under which merging of eigenvalues may occur. We will look at
a problem where applying symmetry arguments allows us to eliminate multiple
(merged) eigenvalues, thus reducing the problem to determining a classical simple
eigenvalue. We then discuss a general technique to look for bifurcations in prob-
lems formulated as implicit functionals. This is useful for a wide class of problems,
including many problems in axially moving materials. At the end of the chapter, we
will consider a variational approach to the stability analysis of an axially moving
panel (a plate undergoing cylindrical deformation).
Chapters 5 through 8 form the main content of the book, concentrating specif-
ically on axially moving materials. We start by introducing the theory of axially
moving materials in a systematic manner, the aim being to give a complete over-
view of the fundamentals. By presenting the theory as a self-contained unit, it is our
hope that this chapter may especially help the student or new researcher just
entering the research field of axially moving materials. Specialists, on the other
hand, can benefit from the discussion on the effects of the axial motion on the
boundary conditions, a topic that has received relatively little attention. This nat-
urally leads to a mixed formulation, which both reduces the continuity requirements
on the solution and clearly shows how the boundary conditions arise, contrasting
the classical treatment of the transverse deformations of axially moving elastic and
viscoelastic materials using fourth- and fifth-order partial differential equations.
Chapter 5 starts by considering the general balance laws of linear and angular
momentum. We derive some general equations for beams and specialize them to the
small-displacement regime. We discuss linear constitutive models for elastic and
viscoelastic materials, and highlight the connection between beams and panels. We
then introduce axial motion in a systematic manner, via a coordinate transformation,
Preface vii
and discuss how this affects the boundary conditions. We consider the dynamic
linear stability analysis of axially moving elastic and viscoelastic materials. We
numerically look at bifurcations in the stability exponents. As a result, we find that
the small-viscosity case behaves radically different from the purely elastic case.
Because no real material in papermaking is purely elastic, this has important
practical implications for the correct qualitative understanding of real physical
systems.
Chapter 6 treats bifurcations of axially moving elastic plates made of isotropic
and orthotropic materials. A static stability analysis is performed to find the critical
axial drive velocity and the corresponding shape in which the system loses stability.
As a special topic of particular interest for process industry applications, we then
look at the axially moving isotropic plate, but with an axial tension distribution that
varies along the width. It is seen that as far as the critical velocity is concerned, the
classical simplification assuming homogeneous tension is acceptable, but the
eigenmode is highly sensitive to even minor variations in the axial tension distri-
bution in the width direction.
Chapter 7 deals with the theoretical analysis of bifurcations of axially moving
strings and beams. Critical velocities of bifurcations of the traveling material are
determined for torsional, longitudinal, and transverse vibration types. The stability
analysis of the axially moving string with and without damping is performed. We
also comment on exact eigensolutions for axially moving beams and panels. The
stability analyses of an axially moving web with elastic supports, and when sub-
jected to a uniform gravitational field, are also considered.
Chapter 8 deals with bifurcations in fluid–structure interaction in the context of
axially moving materials. This chapter includes a brief introduction to fluid
mechanics, after which we look at analytical solutions for two-dimensional
potential flow in fluid–structure interaction with an axially moving panel. A Green’s
function approach is used to analytically derive the fluid reaction pressure in terms
of the panel displacement function. This simplifies the numerical problem to
solving a one-dimensional integrodifferential model. An added-mass approximation
of the derived solution is discussed. Numerical results based on the original exact
solution are presented, for both elastic and viscoelastic axially moving panels
subjected to a potential flow.
The book concludes with Chap. 9, considering optimization problems in ther-
moelasticity. We find the optimal thickness distribution for a beam to resist thermal
buckling and the optimal material distribution for a beam of uniform thickness, with
the same goal. Finally, we derive a guaranteed double-sided estimate that governs
energy dissipation in heat conduction in a locally orthotropic solid body, which
holds regardless of how the local material orientation is distributed in the solid
body. Connecting to the main theme of this book, this has applications in the
analysis of heat conduction in paper materials, which is important when considering
the drying process in papermaking.
This book is addressed to researchers, specialists, and students in the fields
of theoretical and applied mechanics, and of applied and computational mathe-
matics. Considering topics related to manufacturing and processing, the book can
viii Preface
Acknowledgements The research presented in this book was supported by the Academy of
Finland (grant no. 140221, 301391, 297616); RFBR (grant 14-08-00016-a); Project RSF
No. 17 19-01247; RAS program 12, Program of Support of Leading Scientific Schools (grant
2954.2014.1); Jenny and Antti Wihuri Foundation. The authors would like to thank Matthew
Wuetrich for proofreading. For discussions and valuable input, the authors would like to thank
Svetlana Ivanova, Evgeny Makeev, Reijo Kouhia, Tytti Saksa, and Tuomo Ojala.
Contents
ix
x Contents
In this chapter, we present some prototype bifurcation problems that arise in the
mechanics of rigid and deformable structural elements. These problems are typical
for engineering applications and characterize the approaches that can be applied in
the investigation of stability. Some methods of bifurcation theory will be presented
in the context of stability studies of the considered one-dimensional mechanical
problems.
Consider the problem of the stability of an absolutely rigid column of length , which
is elastically clamped at one end, and assume that a compressive force P is applied
to the free end, as shown in Fig. 1.1.
For sufficiently large values of P, when the rigid column loses its vertical trivial
position (i.e., loses stability), the force vector retains its original direction, but its line
of action is displaced in a parallel manner, and the point of application remains at
the free end of the column. We adopt rectangular coordinates x and z, where the end
of the column is at the origin O of this coordinate system, and the x axis coincides
with the direction of the force P, see Fig. 1.1. The axis z lies in the buckling plane
for the beam. Its direction is indicated in Fig. 1.1. We restrict our discussion to small
deflections of the beam. The magnitude of the deflection of the axis of the beam is
measured from the line of action of the force and described by the angle θ, while
and C denote the length of the beam, and the rigidity of the elastic clamping,
respectively. The total potential energy J of the considered mechanical system is the
sum of the elastic energy U and the potential of external force , in other words,
J =U +. (1.1.1)
z 0
x
Taking into account the expressions for the elastic energy of a deformed hinge,
1 2
U= Cθ , (1.1.2)
2
and for the potential of the external force, represented by the product of the force
and the vertical displacement of the point of its application, that is,
1 2
J= Cθ − P (1 − cos θ) . (1.1.4)
2
In the equilibrium position the total potential energy of the considered conservative
mechanical system has a stationary value. In accordance with Lagrange’s theorem on
the stability of an equilibrium position, if, in some position of a conservative system,
the total potential energy has a strict minimum, then this position is the position of
stable equilibrium of the system.
Note here that a position of a system is called a position of equilibrium if the
system, starting from that position with zero velocity, remains in that position. The
position of equilibrium is called stable if, for sufficiently small initial deviations and
sufficiently small initial velocities, the system does not go beyond the limits of an
arbitrary small (preassigned) neighborhood of the position of equilibrium during the
entire period of motion, and similarly has arbitrary small velocities.
1.1 Rigid Column with Elastic Clamping 3
Note also that if the total potential energy J of a conservative system at a position
of equilibrium does not have a minimum there, then the given position of equilibrium
is unstable.
The system under consideration is characterized by the single coordinate θ, and
to analyze the behavior of the system we will use the following expressions for the
derivatives of J :
dJ
= Cθ − P sin θ, (1.1.5)
dθ
d2 J
= C − P cos θ . (1.1.6)
dθ2
To find the equation of equilibrium, we equate the first derivative (1.1.5) to zero. We
have
P sin θ = Cθ . (1.1.7)
There are several equilibrium solutions (equilibrium positions) of the Eq. (1.1.7).
The equilibrium position θ = 0 is stable if
d2 J
>0. (1.1.8)
dθ2
Using (1.1.6), we have the condition
d2 J
= C − P > 0 . (1.1.9)
dθ2
Thus, the total potential energy is at a minimum when P < C and θ = 0 (vertical
position). If P > C and θ = 0, then, as follows from Eq. (1.1.9), the total potential
energy is at a maximum, and consequently the vertical position of the beam is unstable
in this case (see Fig. 1.2).
To study the stability of an equilibrium of the system in non-trivial positions
(θ = 0), it is possible to use the Eq. (1.1.7), the expression (1.1.6), and to require
d2 J θ
=C 1− >0. (1.1.10)
dθ2 tan θ
This inequality is fulfilled when θ < tan θ, in other words for |θ| < π. Thus, the
second derivative is positive for the deflected positions if P > k. Consequently, the
position with θ = 0 is stable for P > C.
In the case
C
θ=0 and P= , (1.1.11)
4 1 Prototype Problems: Bifurcations of Different Kinds
1
stable
θ
−π 0 π
the second derivative of the total potential energy is equal to zero, that is,
d2 J
=0, (1.1.12)
dθ2
and consequently it is necessary to derive and estimate the expressions for higher
derivatives. We have
d3 J d4 J
= P sin θ, = P cos θ. (1.1.13)
dθ3 dθ4
The third derivative is equal to zero, but the fourth derivative is positive,
d4 J
=C >0. (1.1.14)
dθ4
Thus, the nondeflected position (θ = 0), in this case (P = C/), is characterized by
a minimal total potential energy, and consequently this position is stable (see Fig. 1.2,
where P̃ = P/C).
As it was shown previously, the vertical state of the beam is stable up to P ≤ C/.
Therefore, if we load the beam and P̃ ≤ 1, then the beam remains in its original
vertical state. If the force exceeds this value, then the vertical position becomes
unstable, and the beam leaves the original position. Taking into account that, in the
vicinity of the point B, there is another stable deflected position of equilibrium,
the beam passes to the deflected state. The corresponding behavior is shown by the
arrows in Fig. 1.3. If P̃ > 0, then the beam cannot stay in the vertical position, and
it deviates to the right or to the left.
Thus, the first stage in Fig. 1.3 is characterized by the point B, where the ordinate
θ
axis corresponding to the original vertical position crosses the curve P̄ = that
sin θ
corresponds to the deflected position.
1.1 Rigid Column with Elastic Clamping 5
unstable P̃
θ
P̃ = sinθ
P cr
1
stable
π θ
−π − π2 0
2
π
Fig. 1.3 When we load the beam and P̃ ≤ 1, then the beam remains in its original vertical state.
If the force exceeds this value, then the vertical position becomes unstable, and the beam leaves
the original position. Taking into account that in the vicinity of the point B there is another stable
deflected position of equilibrium, the beam passes to the deflected state. The corresponding behavior
is shown by the arrows
Points at which the solution splits into two (or more) branches, such as the point
B, are called bifurcation points or branch points.
If we go through the bifurcation point B, then the original equilibrium state of the
beam loses the stability property, and the points on the line θ = 0 above the point B
correspond to unstable positions.
Bifurcation Phenomenon
If the equation
F(u, λ) = 0 , (1.1.15)
In the discussion given below, following the paper by Banichuk [1], we shall consider
the problems of stability for a compressed rod (column). We shall, in the course of
the investigation, generalize the well-known solution given in the article by Clausen
[2], originally developed for the specific case having a rigid support.
We assume that a cantilevered column is elastically held at one end and a com-
pressive force P is applied to the free end (as shown in Fig. 1.4).
6 1 Prototype Problems: Bifurcations of Different Kinds
For sufficiently large values of P, while the column loses its stability and buckles,
the force vector retains its original direction, but its line of action is displaced in a
parallel manner, and the point of application remains at the free end of the column.
We adopt rectangular coordinates x and z, where the end of the column is at the
origin O of this coordinate system, and the axis x coincides with the direction of the
force P, see Fig. 1.4. The axis z lies in the buckling plane of the beam. Its direction
is indicated in Fig. 1.4.
We restrict our discussion to small deformations of the beam, and we study its
equilibrium within the framework of the linear theory of elasticity. The magnitude
of the deflection of the bent axis of the beam, denoted as w(x), is measured from the
line of action of the force, while denotes the length of the beam. Assuming that the
shape of the cross section of the beam is uniform, we shall derive the basic relations
arising in the problem of maximizing the load causing the loss of stability. We shall
find the optimum (in the above sense) distribution of cross-sectional area S = S(x)
as a function of the distance x measured along the axis of the beam. We have
d2 w
EI + Pw = 0 , E I = C2 S 2 ,
dx 2
dw
w(0) = 0 , c − Pw =0, (1.2.1)
dx x=
S dx = V , P → max ,
0 S
1.2 Elastic Column and Its Optimization 7
where E, c, V and I = I (x) are respectively Young’s modulus, the coefficient of the
rigidity of foundations, the given volume of the beam and moment of inertia of the
cross-sectional area, relative to the axis perpendicular to the bending plane x z and
intersecting the neutral axis of the beam at the point with the coordinate x. The load
P denotes the eigenvalue of the boundary-value problem of Eq. (1.2.1).
A necessary condition for maximizing the critical load of compressed columns
was derived in Lurie [3], Nikolai [4], Chentsov [5], Clausen [2], Keller [6]. It was
obtained that
w2 = S 3 . (1.2.2)
This condition was originally derived for other types of boundary conditions; how-
ever, it is not hard to show that it remains valid in the case considered above. The
necessary condition for maximizing the critical load relates the functions w and S
and, together with the basic relations (1.2.2), leads to a solvable boundary-value
problem. Using (1.2.2) to eliminate the function S from the bending equations and
from the isoperimetric condition, and introducing the nondimensional variables
x C2
x̃ = , w̃ = w , S̃ = S , P̃ = P,
P2 V C
we obtain the following relations:
d2 w
+ w −1/3 = 0 , (1.2.3)
dx 2
dw
w(0) = 0 , − Pw = 0 ,
dx
1
γ V C2
w 2/3 dx = √ , γ= .
0 P c
w = μ1 sin3 θ, θ0 ≤ θ ≤ θ1 ,
where μ1 , μ2 , θ0 and θ1 are constants that are to be determined. To find the values
of these constants, we make use of the boundary conditions at x = 0 and x = 1,
and the conditions x (θ0 ) = 0 and x (θ1 ) = 1. After substituting these values into the
solution (1.2.4), we have
8 1 Prototype Problems: Bifurcations of Different Kinds
1 √ 2/3
sin θ0 = 0 , μ2 = − 3μ1 θ0 , (1.2.5)
2
1 √ 2/3 1
3μ1 θ1 − θ0 − sin 2θ1 = 1 ,
2 2
2/3
√ cos θ1
μ1 = 3 .
P sin3 θ1
2/3
Equation
√ (1.2.5) are satisfied if we substitute θ0 = 0, μ2 = 0, θ1 = θ∗ and μ1 =
3 (cos θ∗ ) / P sin3 θ∗ , where θ∗ denotes a root of the equation
1
θ − sin 2θ
2
ϕ (θ) = 3 cos θ =P (1.2.6)
2 sin3 θ
that satisfies the inequality 0 ≤ θ∗ ≤ π/2. Equation (1.2.6) may be obtained by sub-
2/3
stituting the expression for μ1 into (1.2.5), satisfying (1.2.5) and setting θ0 = 0.
If we add to θ0 = 0 and to θ1 = θ∗ a quantity that is a multiple of π (while the
constants μ1 and μ2 are still computed by the use of (1.2.5)), this can result only in
the change of sign in (1.2.4) for w. Since w is determined by (1.2.3) up to its sign,
this translation along the θ axis does not result in a different solution. It is in principle
possible to find the solution of (1.2.5) with θ0 = 0 and πi < θ1 < π (i + 0.5), where
i = 1, 2, . . . . At the same time, it is easy to see by looking at (1.2.4) that the
function w has additional loops. Since we are computing the eigenvalue P and the
corresponding eigenfunction has a minimum number of loops, it is not necessary
to consider such (more complex) solutions. The solution of the problem posed in
(1.2.5) with
π(i − 0.5) < θ1 < πi, i = 1, 2, . . .
does not exist, since in this case the right-hand side of (1.2.5) is negative, while the
left-hand side is positive.
Let us study some properties of the functions ϕ(θ) that will be needed in our future
discussion. For small values of θ, we represent ϕ(θ) by a Taylor
series,
with accuracy
up to terms of the order θ3 . Then we have ϕ(θ) ≈ (1/5) 1 − θ2 . Consequently, as
θ → 0, the function ϕ (θ) → 1, with ϕ(θ) < 1. Further, let us consider the case
θ → π/2. For the sake of convenience, we introduce the variable ω = π/2 − θ,
which approaches zero as θ → π/2. Substituting the variable ω for ϕ in (1.2.6) and
eliminating θ and then decomposing ϕ in a Taylor series for small values of ω, we
derive ϕ ≈ (3ω(π − 4ω)) /4. Therefore, ϕ approaches zero as θ → π/2. Using an
asymptotic series and some simple calculations, we obtain a graph of the function
ϕ(θ), which is illustrated in Fig. 1.5.
We can see from this graph that as the variable θ varies between zero and π/2, the
function ϕ(θ) monotonically decreases from one to zero. Hence P < 1. In terms of
1.2 Elastic Column and Its Optimization 9
0.5
0
0.25π 0.5π θ
The graph of the function (θ) shown in Fig. 1.6 was drawn with the help of very
simple calculus.
As θ decreases from π/2 to zero, the function increases in a monotonic fashion
from zero to infinity. As θ → 0, this function has an asymptotic representation
√
3 3
(θ) ≈ (1.2.8)
5θ
Consequently, there is the solution for (1.2.7), which satisfies the conditions 0 <
θ1 ≤ π/2 for an arbitrary nonnegative value of the parameter γ.
Thus, for given values of the constants C2 , , V , and c, finding the optimum
distribution of the cross-sectional area S(x) and of the corresponding value of the
load P is reduced to the following sequence of computational steps.
For given values of C2 , , V , and c, we compute the value of the nondimensional
parameter γ. Using this value of γ, we solve Eq. (1.2.7) and compute the value of
θ1 . Next, we calculate the magnitude of the critical load P, which according to Eq.
(1.2.6) is equal to P = ϕ(θ1 ). The computation of the critical shape (of the deformed
column) is carried out by using the formulas
10 1 Prototype Problems: Bifurcations of Different Kinds
0 0.5 π
1 √
θ−sin 2θ 2 P sin2 θ
x= 2 , S= (1.2.9)
1 √ 1
θ1 − sin 2θ1 3γ θ1 − sin 2θ1
2 2
where 0 ≤ θ ≤ θ1 . Equation (1.2.9) follow directly from Eqs. (1.2.2), (1.2.4) and
(1.2.5). The results derived for γ = 0.5i, i = 1, 2, . . . and for ten values of the
quantities θ1 and P are illustrated in Table 1.1.
In Fig. 1.7, curves 1 and 2 illustrate the optimum distribution of S = S(x) for
γ = 0.5 and γ = 5. The dotted curve shows the distribution S(x) for a rigidly held
beam (γ = 0). Using (1.2.9) it is possible to show that as the rigidity of the support c
1.2 Elastic Column and Its Optimization 11
2
S 1
2
1
1
0 0.5 x 1
Fig. 1.7 Curves 1 and 2 illustrate the optimum distribution of S = S(x) for corresponding values
of γ = 0.5 and γ = 5. The dotted curve shows the distribution S(x) for a rigidly held beam (γ = 0)
Let a straight elastic rod of length be positioned along the x axis of a Cartesian
coordinate system x yz. The ends of the rod are rigidly held at the points x = 0
and x = . The rod is torsioned by a couple M applied to its end (see Fig. 1.8). It
is assumed that the rod has constant rigidity in various cross-sectional planes, and
therefore that
E I y = E Iz = a ,
where E is Young’s modulus for the material, and I y and Iz are the moments of
inertia of the transverse cross-sectional area with respect to lines passing through a
point on the neutral axis of bending and parallel to the axes y and z, respectively.
In studying the stability of the rod and the critical values of the applied torque,
we shall follow Euler’s static approach. Let y = y(x) and z = z(x) be functions
describing the shape of the axis of the buckled rod. We write down the corresponding
equations of equilibrium and the boundary conditions:
2
d2 d y d3 z
a 2 =M 3 ,
dx 2 dx dx
12 1 Prototype Problems: Bifurcations of Different Kinds
x
0 z
2
d2 d z d3 y
a 2 =M 3 , (1.3.1)
dx 2 dx dx
dy dz
y(0) = = z(0) = =0,
dx x=0 dx x=0
dy dz
y() = = z() = =0,
dx x= dx x=
We note that Eq. (1.3.1) result only if we assume that the deflections are “small”.
The functions y(x) = z(x) ≡ 0 describe the equation of the unbuckled axis satisfying
the equilibrium equations and the boundary conditions (1.3.1). According to Euler’s
theory, the magnitude of the critical load and the loss of stability are determined by
the smallest eigenvalue and the corresponding eigenfunctions y(x) ≡ 0 and z(x) ≡ 0
for the boundary-value problem (1.3.1).
For the particular case when torsional rigidity is uniformly distributed, that is
a = const, finding the torque that causes a loss of stability is known to reduce to the
computation of the smallest root for the (transcendental) equation
M M
tan = . (1.3.2)
2a 2a
The magnitude of the critical torque is equal to ±8.988 a/. Following the article
Banichuk and Barsuk [7], let us derive an analogue of this formula for the more
general case when the distribution of torsional rigidity is a fairly arbitrary function
1.3 Elastic Rod Under Torsion 13
of the variable x, i.e. a = a(x). To accomplish it, we integrate twice both sides of the
Eq. (1.3.1). Then we multiply the second equation in its integrated form by i (where
i is the imaginary unit) and add these equations side by side. After introduction of
the complex variable
w(x) = y(x) + i z(x) , (1.3.3)
x x x
t i Mϕ(t) ei Mϕ(t) dt
c1 e−i Mϕ(x) e dtdx + c2 e−i Mϕ(x) dtdx = 0 , ϕ(x) ≡ .
0 0 a(t) 0 0 a(t) 0 a(t)
(1.3.7)
When rigidity is uniformly distributed along the length of the rod, that is, if a(x) =
const in (1.3.8), we arrive at the well-known formula tan (M/(2a)) = M/(2a).
We note that the complex Eq. (1.3.8) can be written as a system of two real equations
In this manner a study of the stability of torsioned rods with variable rigidity may be
reduced to a computation of quadratures for a given function a(x) and the solution
of the transcendental Eq. (1.3.8) for M.
We shall consider separately the case of the symmetric distribution of bending
rigidity when a(x) = a( − x). In this case
ϕ(x) = ϕ + (x) ,
2
where (x) is an antisymmetric function, that is, (x) = −( − x). Using these
properties of the functions a(x) and (x) and the rules of integration on symmetric
domains of symmetric or antisymmetric functions, (1.3.8), for example, is trans-
formed into the form
2 cos (M((x))) 2
f (M) = dx [χ1 cos (M(x)) + χ2 sin (M(x))] dx+
0 a(x) 0
(1.3.10)
x−
2 2 2
sin (M(x)) dx [χ3 cos (M(x)) − χ4 sin (M(x))] dx = 0 .
0 a(x) 0
Here
ξ−
x ξ−
x
2 2
χ1 (x) = cos (M(ξ)) dξ , χ2 (x) = sin (M(ξ)) dξ ,
0 a(ξ) 0 a(ξ)
x x
sin (M(ξ)) cos (M(ξ))
χ3 (x) = dξ , χ4 (x) = dξ .
0 a(ξ) 0 a(ξ)
The essential difference between (1.3.8) and (1.3.10) is that (1.3.10) is a single
equation in the unknown M instead of a system of equations.
Now let us consider a torsioned rod simply supported at the points x = 0 and
x = . Finding a static condition for the loss of stability (Greenhill’s problem) leads
to the boundary value problem
d2 y dz d2 z dy
a 2
=M , a 2
= −M , (1.3.11)
dx dx dx dx
We multiply both sides of the second equation in (1.3.11) by i and add them side by
side. We introduce a function
d2 w dw
a 2
= −i M , (1.3.13)
dx dx
w(0) = w() = 0 .
Integrating equation (1.3.13) twice and taking care of the assigned boundary con-
ditions, we obtain a formula for w and a system of two simultaneous linear, homo-
geneous equations that can be solved to determine the constants of integration. The
necessary condition on the existence of nontrivial solutions (determinant of coeffi-
cients for this system vanishing) reduces to a single complex equation in M, which
may be written as a system of the two real equations
cos (Mϕ(x)) dx = 0 , sin (Mϕ(x)) dx = 0 . (1.3.14)
0 0
The smallest common root of these equations defines a critical divergence torque.
Therefore, finding critical torques amounts to finding a simultaneous solution for the
system (1.3.14).
We shall show that for a symmetric distribution of rigidity, the system (1.3.14)
may be replaced by a single equation. Let a(x) be a quite arbitrary distribution of
bending rigidity. We write 1/a(x) as a sum of the two terms
1
= (x) + (x) , (1.3.15)
a(x)
where (x) and (x) are, respectively, symmetric and antisymmetric functions
with respect to the midpoint of the interval [0, ]. The function ϕ(x) can also be
represented as a sum of a symmetric and antisymmetric functions ω(x) and ψ(x),
respectively
ϕ(x) = ω(x) + ψ(x) , (1.3.16)
x
2
ω(x) = (t)dt + (t)dt ,
0 0
x
2
ψ(x) = (t)dt − (t)dt ,
0 0
We obtain
cos (Mψ(x)) cos (Mω(x)) dx = 0 , (1.3.17)
0
cos (Mψ(x)) sin (Mω(x)) dx = 0 .
0
sin Mϕ cos (Mψ(x)) dx = 0 .
2 0
It follows directly that the system (1.3.14) can be reduced in this case to the single
equation
cos Mψ dx = 0 , , ψ(x) = ϕ(x) − ϕ . (1.3.18)
0 2 2
For a constant distribution of rigidity, Eq. (1.3.18) predicts the well-known value
of the critical torque
EI
M = 2π .
The solution of Eq. (1.3.18) can be found very effectively using one of several
analytic or numerical techniques, in particular, Newton’s methods. Thus, equation
(1.3.18) results in a real simplification particularly in solving problems, such as the
optimization of rigidities, in which it is necessary to recompute the critical value of
M several times for various approximate values of a(x).
Let us examine some of the simplest problems of optimizing stability. Let a rod
have similarly shaped cross-sections, and consequently
a(x) = k [S(x)]2
where S(x) is the function defining the distribution of magnitude of the area. We
formulate the following problem in terms of nondimensional variables: we need
to find a function of cross-sectional area S(x) satisfying Smin < S(x), and also a
constant volume condition, and assign a maximum to the critical value of the torque:
M∗ = max M, (1.3.19)
S
1.3 Elastic Rod Under Torsion 17
where M is computed from (1.3.10) and (1.3.18), and Smin is a given nondimensional
quantity.
For the indicated constraints on the volume of the rod, and the allowable values
of its thickness, problem (1.3.19) was solved numerically in Banichuk and Barsuk
[7] using an iterative algorithm. This algorithm, applied to our optimization prob-
lem, consists of finding an approximate value of a(x) using Newton’s method (in
the process of iterating critical values of M) from the current value of a(x), and
then improving the variation of the control function a(x) by the gradient projection
technique. The computation of the critical value of M from (1.3.10) and (1.3.18)
was carried out to within 10−5 accuracy. The computation was terminated when the
magnitude of the gradient of the optimized functional (i.e. the error in satisfying the
optimality conditions) was smaller than 10−3 .
The optimal functions S(x) were found for several values of the parameter Smin .
Figure 1.9 illustrates the function S(x) found as a result of these numerical compu-
tations for a rod simply supported at both ends.
Curves 1, 2, and 3 correspond to the values of the parameter Smin = 0.98, 0.92
and 0.88, respectively. For these values of the parameter, the magnitude of the critical
torque is M = 6.56, 7.12, and 7.80, respectively. The relative gain in comparison with
a rod having a constant rigidity function and a unit volume amounts to 4.45%, 14.9%,
and 24.2%, respectively. Numerical solutions were also carried out for rods with
built-in ends. The expression for the gradient in the stability optimization problem
was carried out with the help of the following representation for the eigenvalue of
the boundary-value problem (1.3.1):
1 d2 w d2 w ∗
0
2
S Re · dx
dx 2 dx 2
M =− (1.3.20)
1 dw d2 w ∗
0 Im · dx
dx dx 2
where w(x) is computed from (1.3.5) and (1.3.6), and w ∗ is the complex conjugate
function of w.
0 0.2 0.4
x
Another random document with
no related content on Scribd:
demandai, à brûle-pourpoint, en quoi consistait la difficulté, et qui, en
définitive, la provoquait.
— C’est le Travail, me dit-il. Vous feriez mieux de ne pas vous en
occuper.
LE TRAVAIL
Mais dès que l’on se mettait à poser des questions l’on se perdait
dans un labyrinthe d’allusions, de réserves, de discours, pour la
plupart faits avec contrainte, comme si les interlocuteurs récitaient
quelque chose appris par cœur. En voici quelques exemples : Un
homme m’accula dans un coin au moyen d’une phrase — une seule
— où tout, à peu près, était lettre majuscule : — Le Sentiment
Général chez notre Peuple est qu’il FAUT QUE LES JAPONAIS
S’EN AILLENT !
— Fort bien, répliquai-je. Comment pensez-vous vous y
prendre ?
— Cela ne nous regarde pas. Le Sentiment Général, etc.
— Sans doute. Le Sentiment est une belle chose, mais que
comptez-vous faire ? Il ne condescendait pas à s’abaisser jusqu’à
fournir des détails, mais ne cessait de répéter le Sentiment, que, me
conformant à ma promesse, je ne manque pas de rapporter ici.
Un autre fut un peu plus explicite. — Le Chinois, dit-il, nous
désirons le garder, mais le Japonais doit partir !
— Mais alors qui prendra sa place ? Ce pays-ci n’est-il pas un
peu trop neuf pour qu’on en chasse les gens à coups de pied ?
— Nous devons développer nos Ressources lentement,
Monsieur, sans perdre de vue les Intérêts de nos Enfants. Nous
devons conserver le Continent pour les Races qui s’assimileront
avec les Nôtres. Nous ne devons pas nous laisser submerger par
des Étrangers.
— Alors amenez votre propre race, et amenez-la vite, osai-je
répondre. Voilà cependant la seule observation qu’il ne faut pas faire
dans certaines régions de l’Ouest. Je perdis de mon prestige
furieusement pendant qu’il expliquait (tout comme les Hollandais
l’avaient fait au Cap, il y a de ça des années) comme quoi la
Colombie Britannique n’était pas du tout aussi riche qu’elle en avait
l’air ; qu’elle était étouffée par des capitalistes et monopoleurs de
toute sorte ; la main-d’œuvre des blancs devait être interrompue,
nourrie et chauffée pendant l’hiver ; que les frais d’existence étaient
énormes ; qu’on touchait à la fin d’une ère de prospérité, que les
années maigres arrivaient enfin, que même si des mesures étaient
devenues nécessaires pour faire venir d’autres blancs elles devaient
être prises avec beaucoup de précaution. Puis il ajouta que les tarifs
des chemins de fer de la Colombie Britannique étaient si élevés
qu’ils empêchaient les émigrants de s’y rendre.
— Mais est-ce que les tarifs n’ont pas été diminués ? demandai-
je.
— Si, — si, je crois, mais on a tellement besoin d’immigrants
qu’ils sont raflés avant d’arriver aussi loin du côté Ouest. Il ne faut
pas oublier non plus que la main-d’œuvre d’ouvriers de métier est
bien différente de la main-d’œuvre agricole. Elle dépend de tant de
choses. Et puis, il faut que le Japonais parte.
— C’est ce que l’on m’a dit. Mais on m’a raconté aussi qu’il y
avait des laiteries et des vergers dans la Colombie Britannique qu’on
a dû abandonner parce qu’on ne trouvait personne ni pour traire ni
pour cueillir les fruits. Pensez-vous que ce soit vrai ?
— Voyons, vous ne voudriez pas qu’un homme, avec tous les
débouchés possibles que notre pays lui prodigue, aille traire des
vaches dans un pré ? Un Chinois suffit pour cela. Il nous faut des
races qui puissent s’assimiler, etc., etc.
— Mais l’Armée du Salut n’a-t-elle pas offert, il y a peu de temps,
d’amener ici trois ou quatre mille Anglais ? Qu’est devenue cette
idée ?
— On n’y a pas, — comment dirai-je ? — donné suite.
— Pourquoi pas ?
— Pour des raisons politiques, je crois. Il ne nous faut pas des
Gens susceptibles d’abaisser le Niveau de la Vie. Voilà pourquoi les
Japonais devront s’en aller.
— Alors pourquoi garder les Chinois ?
— Avec les Chinois on s’entend ; on ne peut pas se passer d’eux.
Mais il nous faut une Émigration d’un Type tel qu’il lui soit possible
de s’assimiler avec Notre Peuple. Je pense être clair.
J’espérais qu’il l’avait été.
Et maintenant écoutez ce qu’ont à dire une mère et une
ménagère.
— Ce joli état de choses nous coûte notre santé et celle de nos
enfants. Avez-vous entendu dire couramment « la Frontière fait
souffrir les femmes et le bétail ? » Nous ne sommes pas sur la
Frontière ici, mais à certains égards c’est pis parce que nous avons
tous les luxes et toutes les apparences — du joli cristal et de
l’argenterie, à étaler sur la table. Il faut les épousseter, les polir, les
arranger une fois que notre travail de ménage est fini. Sûrement que
cela ne vous dit pas grand’chose, à vous, mais essayez-en pendant
un mois ! Nous n’avons pas de domestiques. En ce moment un
Chinois coûte cinquante ou soixante dollars par mois. Nos maris
n’ont pas toujours le moyen de payer ça. Quel âge me donnez-
vous ? Je n’ai pas encore trente ans. Dieu soit loué en tous cas que
j’aie empêché ma sœur de venir dans l’Ouest. Ah ! oui, c’est un beau
pays — pour les hommes !
— Ne pourriez-vous pas faire venir des domestiques
d’Angleterre ?
— Je n’ai pas le moyen de payer le transport d’une fille pour
qu’elle se marie au bout de trois mois. Et puis elle ne voudrait pas
travailler. Une fois qu’elles ont vu le Chinois à l’œuvre elles refusent.
— Et vous, vous ne refusez pas le Japonais ?
— Certainement pas. Personne ne le refuse. Tout ça c’est de la
politique. Les femmes des hommes qui gagnent six à sept dollars
par jour — les ouvriers spécialistes — ont des domestiques chinois
et japonais. Nous ne pouvons pas faire comme eux, nous autres.
Nous avons à penser à l’avenir, à épargner, mais ceux-là dépensent
jusqu’à leur dernier centime. Ils savent, eux, qu’ils n’ont rien à
craindre. Ils sont Le Travail. On s’occupera d’eux quoiqu’il arrive.
Vous pouvez juger, vous-même, si l’État s’occupe de moi.
Un peu plus tard, j’eus l’occasion de traverser une grande et
belle ville entre six et sept heures par une matinée fraîche. Des
Chinois et des Japonais livraient aux maisons silencieuses le lait, le
poisson, les légumes, etc. Pour cette corvée glaciale pas un seul
blanc n’était visible.
Plus tard encore un homme vint me voir sans trop afficher son
nom. Il faisait d’assez importantes affaires et il me fit comprendre
(d’autres avaient parlé à peu près de même) que si je répétais ses
paroles ses affaires en souffriraient. Il parla sans discontinuer
pendant une demi-heure.
— Dois-je donc en conclure, lui dis-je, que ce que vous appelez
« Le Travail » domine absolument cette partie-ci du monde ?
Il fit signe que oui.
— Qu’il est difficile d’amener ici l’ouvrier de métier ?
— Difficile ? Seigneur Dieu ! s’il me faut un ouvrier
supplémentaire pour mon travail (je paie, bien entendu, les gages
fixés par l’Union) il faut que je le fasse venir en cachette ; il faut que
j’aille le rencontrer comme par accident, plus bas sur la ligne, et si
l’Union vient à le savoir, très probablement elle lui intimera l’ordre de
s’en retourner vers l’Est, ou le congédiera et l’enverra de l’autre côté
de la Frontière.
— Même s’il appartient à l’Union ? Pourquoi ?
— On lui dira que les conditions de Travail ne sont pas bonnes
ici. Il sait fort bien ce qu’il faut entendre par là. Il fera demi-tour assez
vite. Je fais pas mal d’affaires, et je ne puis m’exposer à entrer en
lutte ouverte avec les Unions.
— Qu’arriverait-il au cas où vous le feriez ?
— Savez-vous ce qui se passe de l’autre côté de la Frontière ?
On fait sauter les gens à coups de dynamite.
— Mais ici on n’est pas de l’autre côté de la Frontière.
— C’est bigrement trop près pour être agréable. Et puis, aux
témoins, tout autant qu’aux autres, il arrive d’être dynamités. Voyez-
vous, la situation créée par le Travail ne résulte pas de ce que l’on
fait ou ne fait pas de notre côté, tout est géré là-bas. Vous avez pu
vous rendre compte qu’en général on en parle avec précaution.
— Oui, je l’ai remarqué.
— Eh bien, tout cela n’est pas bien. Je ne dis pas que les Unions
d’ici feraient quelque chose contre vous, — et, retenez-le bien, je
suis partisan moi-même des droits du Travail, le Travail n’a pas de
meilleur défenseur que moi-même ; j’ai été, moi aussi, ouvrier, bien
que je sois patron aujourd’hui. N’allez pas croire, n’est-ce pas, que je
sois contre le Travail.
— Pas le moins du monde. Je comprends fort bien. Vous trouvez
seulement que le Travail agit, parfois, — comment dire —
inconsidérément ?
— Voyez ce qui se passe de l’autre côté de la Frontière ! On a dû
vous expliquer que cette petite affaire avec les Japonais à
Vancouver a été manigancée en-dessous, n’est-ce pas ? Il m’est
avis que les nôtres ne l’auraient pas fait tout seuls.
— On me l’a souvent répété. Est-ce jouer tout-à-fait beau jeu que
d’incriminer un autre pays ?
— On voit que vous ne demeurez pas ici, vous. Mais, pour
reprendre, si nous nous défaisons des Japonais actuellement, l’on
viendra nous demander bientôt de nous défaire de quelqu’un
d’autre. Il n’y a aucune limite, Monsieur, aux exigences du Travail,
aucune !
— Je croyais que tout ce qu’ils demandaient c’était de bons
gages pour du bon travail ?
— En Angleterre peut-être, mais ici ils ont l’intention de diriger le
pays, ah, oui alors !
— Et le pays ? comment cela lui plaît-il ?
— Nous ne sommes pas loin d’en avoir assez. Ça n’importe pas
beaucoup dans les moments de plein rendement, les patrons feront
tout plutôt que d’arrêter le travail — mais lorsque les temps seront
mauvais, vous en aurez des nouvelles. Notre pays est un pays riche,
— malgré tout ce qu’on raconte — mais nous sommes arrêtés à tout
propos par la main-d’œuvre. Voyez-vous, il y a des quantités et des
quantités de petites affaires que des amis à moi désireraient lancer.
Des affaires partout, si seulement on les laissait tranquilles — mais
non !
— Ça, c’est dommage. A propos, que pensez-vous de la
question Japonaise ?
— Je ne pense pas. Je sais. Les deux partis politiques jouent le
jeu du vote du Travail, mais comprenez-vous ce que cela veut dire ?
Je m’efforçais de comprendre.
— Et pas un ne dira la vérité, à savoir que si l’Asiatique s’en va,
ce côté-ci du Continent disparaîtra complètement, à moins que nous
n’obtenions une immigration blanche libre. Pourtant tout parti qui
proposerait l’immigration blanche sur une large échelle serait
blackboulé aux prochaines élections. Je ne vous dis là que ce que
pensent les Politiciens. Mon avis personnel est que si un homme
osait résister au Travail — pas que j’en veuille le moins du monde au
Travail — et parlait raison, bien des gens se rangeraient à son avis
— sans bruit, bien entendu. Je crois qu’il obtiendrait, au bout d’un
certain temps, même l’immigration blanche. Il serait blackboulé aux
premières élections, ça va sans dire, mais en fin de compte… Nous
en avons assez du Travail. Je tenais à vous dire la vérité.
— Merci bien. Et vous ne pensez pas qu’une tentative pour
introduire l’immigration blanche réussisse ?
— Non, si elle ne convenait pas au Travail. Essayez, si vous
voulez, mais vous verrez ce qui arrivera.
En tenant compte de cette indication j’ai fait une expérience dans
une autre ville. Il y avait là trois hommes haut placés, riches, tous les
trois vivement préoccupés du développement de leur terre, tous trois
affirmaient que ce qui manquait à la terre c’était l’immigration
blanche. Et tous les quatre nous en avons parlé, dans tous les sens,
de toutes les manières possibles et imaginables. La seule chose sur
laquelle leur avis ne variait pas c’était, qu’au cas où l’on importerait
de quelque façon que ce fût des blancs en Colombie Britannique, le
recrutement pouvant être confié à des particuliers ou à d’autres, les
mesures nécessaires devaient être prises secrètement, sans quoi
les affaires des intéressés en souffriraient.
A cet endroit, j’abandonnai la conversation touchant la Grande
Question qui Agite Toute Notre Communauté ; je vous laisse, à vous
plus spécialement, Australiens et Habitants du Cap, le soin d’en tirer
vos conclusions.
Extérieurement, la Colombie Britannique paraît être la région la
plus riche et la plus séduisante de tout le Continent. En plus de ses
ressources propres elle a de bonnes chances d’acquérir un
immense commerce avec l’Asie, qu’elle désire ardemment. Sa terre,
en bien des endroits sur de vastes étendues, convient
admirablement au petit fermier et à l’arboriculteur, qui peut envoyer
son camion aux villes. De tous les côtés j’entendis réclamer de la
main-d’œuvre de toute espèce. Et d’autre part, dans nul autre lieu
sur le Continent je ne rencontrai tant d’hommes qui décriaient la
valeur et les possibilités de leur pays, ni qui s’appesantissaient plus
volontiers sur les souffrances et les privations que l’immigrant blanc
avait à endurer. Je crois que deux ou trois messieurs se sont rendus
en Angleterre pour expliquer les inconvénients de vive voix. Il est
possible qu’ils encourent une très grande responsabilité
actuellement, et une plus terrible encore dans l’avenir.
LES VILLES FORTUNÉES