27

Bifurcation:
Linearized
Prebuckling I

271

Chapter 27: BIFURCATION: LINEARIZED PREBUCKLING I

272

TABLE OF CONTENTS
Page

27.1. Introduction
273
27.2. Stability Assessment Criteria
274
27.2.1. Static Criterion . . . . . . . . . . . . . . . . . 274
27.2.2. Dynamic Criterion . . . . . . . . . . . . . . .
274
27.3. The Tangent Stiffness Test
275
27.4. Linearized Prebuckling
276
27.5. The LPB Eigensystem
277
27.6. Stability Eigenproblem
278
27.7. LPB Analysis Example
278
27.8. Summary of LPB Steps
2710
27. Exercises . . . . . . . . . . . . . . . . . . . . . . 2712

272

273

27.1

INTRODUCTION

27.1. Introduction
This Chapter starts a systematic study of the stability of elastic structures. We shall postpone the
more rigorously mathematical denition of stability (or lack thereof) until later because the concept
is essentially dynamic in nature. For the moment the following physically intuitive concept should
sufce:

A structure is stable at an equilibrium position if it returns to that position

upon being disturbed by an extraneous action

Note that this informal denition is dynamic in nature, because the words returns and upon
convey a sense of time and/or history. But it does not imply that the inertial and damping effects of
true dynamics are involved. So real time is not involved in the static case.
A structure that is initially stable may lose stability as it moves to another equilibrium position
when the control parameter(s) change. Under certain conditions, that transition is associated with
the occurrence of a critical point. These were classied into limit points and bifurcation points in
Chapter 5.
For the slender structures that occur in aerospace, civil and mechanical engineering, bifurcation
points are more practically important than limit points. Consequently, attention will be initially
directed to the phenomena of bifurcation or branching of equilibrium states, a set of phenomena
informally call (in structural engineering) as buckling. The analysis of what happens to the structure
after it crosses a bifurcation point is called post-buckling analysis.
The study of bifurcation and post-buckling while carrying out a full nonlinear analysis is a mathematically demanding subject. But in important cases the loss of stability of a geometrically
nonlinear structure by bifurcation can be assessed by solving linear algebraic eigenvalue problems
or eigenproblems for short. This eigenanalysis provides the magnitude of the loads (or, more
generally, of the control parameters) at which buckling is expected to occur. The analysis yields no
information on post-buckling behavior. Information on the buckling load levels is often sufcient,
however, for design purposes.
The present Chapter covers the source of such eigenproblems for conservatively loaded elastic
structures. Chapters 28 through 31 discuss stability in the context of full nonlinear analysis. The
two nal Chapters (32 and 33) extend these concepts to structures under nonconservative loading.
Following a brief review of the stability assessment criteria the singular-stiffness test is described.
Attention is then focused on the particular form of this test that is most used in engineering practice:
the linearized prebuckling (LPB) analysis. The associated buckling eigenproblem is formulated.
The application of LPB on a simple problem is worked out using the bar element developed in
the previous three sections. The assumptions underlying LPB and its range of applicability are
discussed in the next Chapter.

273

Chapter 27: BIFURCATION: LINEARIZED PREBUCKLING I

274

27.2. Stability Assessment Criteria

For elastic, geometrically-nonlinear structures under static loading we can distinguish the following
techniques for stability assessment.

Conservative Static criterion (Euler method): singular stiffness

Dynamic criterion: zero frequency

Loading

zero frequency (divergence)

Nonconservative Dynamic criterion

frequency coalescence (utter)

27.2.1. Static Criterion
The static criterion is also known as Eulers method, since Euler introduced it in his famous
investigations of the elastica published in 1744. Other names for it are energy method and method
of adjacent states. To apply this criterion we look at admissible static perturbations of an equilibrium
position1 . These perturbations generate adjacent states or congurations, which are not generally
in equilibrium.
Stability is assessed by comparing the potential energy of these adjacent congurations with that
of the equilibrium position. If all adjacent states have a higher potential energy, the equilibrium is
stable. If at least one state has a lower (equal) potential energy the equilibrium is unstable (neutrally
stable). This comparison can be expressed in terms of the second variation of the potential energy
and hence can be reduced to the assessment of the positive denite character of the tangent stiffness
matrix.
Although stability is a dynamic phenomenon, no true-dynamics concepts such as mass or damping
are involved in the application of the static criterion, which is a key reason for its popularity. But
the reasoning behind it makes it strictly applicable only to conservatively loaded systems, because
a load potential function is assumed to exist.
27.2.2. Dynamic Criterion
The dynamic criterion looks at dynamic perturbations of the static equilibrium position. In informal
terms, give the structure a (little) kick and see how it moves. More precisely, we consider small
oscillations about the equilibrium position, and pose an eigenproblem that determines characteristic
exponents and associated eigenmodes. The characteristic exponents are generally complex numbers. If all characteristic exponents have no positive real components the equilibrium is dynamically
stable, and unstable otherwise.
These exponents change as the control parameter is varied. For sufciently small values the
structure is stable. Loss of stability occurs when a characteristic exponent enters the right-hand
complex plane. If that happens, the associated mode viewed as a displacement pattern will amplify
exponentially in the course of time. A deeper study of the stable-to-unstable transition mechanism
reveals two types of instability phenomena, which are associated with the physically-oriented terms
terms divergence and utter.
1

Admissible in the sense of the Principle of Virtual Work: variations of the state parameters that are consistent with the
essential boundary conditions (kinematic constraints)

274

275

27.3

Apply an
allowed
perturbation
Equilibrium
configuration

THE TANGENT STIFFNESS TEST

Motion oscillates about

equilibrium configuration
or decays toward it

Subsequent
motion outcome

Transition
between stable
and unstable

Motion is either unbounded, or

oscillates about, or decays toward,
another equilibrium configuration

S: Stable*

N: Neutrally stable

U: Unstable

* Strictly speaking, S requires stability

for all possible admissible perturbations
Figure 27.1. Stability outcomes upon perturbing a static equilibrium conguration.

Divergence occurs when the characteristic exponent enters the right-hand plane through the origin.
It can therefore be directly correlated with both the zero frequency test of dynamics and the singular
stiffness test of statics.
The dynamic criterion is applicable to both conservative and nonconservative systems. This wider
range of application is counterbalanced by the need of incorporating additional information (mass
and possibly damping) into the problem. Furthermore, unsymmetric eigenproblems arise in the
nonconservative case, and these are the source of many computational difculties.
Figure 27.1, which is extracted from an undergraduate course offering, depicts the three possible
outcomes of applying a perturbation to an equilibrium conguration.
27.3. The Tangent Stiffness Test
The stability of conservative systems can be assessed by looking at the spectrum2 of the tangent
stiffness matrix K at an equilibrium conguration. Let i denote the i th eigenvalue of K. The set
of i s are the solution of the algebraic eigenproblem
K zi = i zi .

(27.1)

Since K is real symmetric3 all of its eigenvalues are real. As a result, we can administer the
following test:
Case

Condition

The equilibrium conguration is

(I)
(II)
(III)

All i > 0
All i 0 and at least one i = 0
At least one i < 0

Strongly stable
Neutrally stable
Unstable

The spectrum of a matrix is the set of its eigenvalues.

Because K = 2 /uu is the Hessian of a total potential energy

275

Chapter 27: BIFURCATION: LINEARIZED PREBUCKLING I

276

Mathematically, K is said to be positive denite, nonnegative and indenite in cases (I), (II) and
(III), respectively.
In engineering applications one is especially interested in the behavior of the structure as the stage
control parameter is varied. Consequently
K = K().

(27.2)

Given this dependence, a key information is the transition from stability to instability at the value
of closest to stage start, which is usually = 0. This is called the critical value of , which we
shall denote as cr .
If the entries of K depend continuously on the eigenvalues of K also depend continuously4 on
, although the dependence is not necessarily continuously differentiable. It follows that transition
from strong stability case (I) to instability case (III) has to go through case (II), i.e. a
zero eigenvalue. Thus a necessary condition is that K be singular, that is
det K(cr ) = 0,

(27.3)

K(ucr , cr ) zcr = 0,

(27.4)

or, equivalently,
in which zcr = 0 is the buckling mode introduced in Chapter 5, where it was called a null eigenvector.
Either (27.3) or (27.4) expresses the static test for nding the transition of stability to instability.
Remark 27.1. Equation (27.3) is a nonlinear eigenvalue problem because: (a) K has to be evaluated at an
equilibrium position, and (b) K is a nonlinear function of u, which in turn depends on as dened by the
primary equilibrium path. It follows that in general a complete response analysis has to be conducted to
solve (27.3) Such techniques were called indirect methods in the context of critical point location methods
discussed in Chapter 26. This involves evaluating K at each computed equilibrium position, and then nding
the spectrum of K. An analysis of this nature is obviously computationally expensive. One way of reducing
part of the cost is noted in the following remark.
Remark 27.2. If K is known at a given , an explicit solution of the eigenproblem (27.1) is not necessary for
assessing stability. It is sufcient to factor K as

K = LDLT ,

(27.5)

in which L is unit lower triangular and D is diagonal. The number of negative eigenvalues of K is equal to
the number of negative diagonal elements (pivots) of D. Matrix factorization is considerably cheaper than
carrying out a complete eigenanalysis because sparseness can be exploited more effectively.
Remark 27.3. The condition (27.3) is not sufcient for concluding that a system that is stable for < cr will go
unstable as exceeds cr . A counterexample is provided by the stable-symmetric bifurcation point discussed
in later Chapters. The Euler column furnishes a classical example. At such points (27.3) holds implying
neutral stability but the system does not lose stability as the bifurcation point is traversed. Nonetheless the
displacements may become so large that the structure is practically rendered useless.
4

Continuous dependence of eigenvalues on the entries is guaranteed by the perturbation theory for symmetric and Hermitian
matrices. This continuous dependence does not hold, however, for eigenvectors.

276

277

27.5

THE LPB EIGENSYSTEM

27.4. Linearized Prebuckling

We investigate now the rst critical state of an elastic system if the change in geometry prior to it
can be neglected. We shall see that in this case the nonlinear equilibrium equations can be partly
linearized, a process that leads to the classical stability eigenproblem or buckling eigenproblem.
The eigenstability analysis procedure that neglects prebuckling displacements is known as linearized prebuckling (LPB). The modeling assumptions that are tacitly or explicitly made in LPB
are discussed in some detail in the next Chapter, as well as the practical limitations that emanate
from these assumptions. In the present Chapter we discuss the formulation of the LPB eigenproblem
and illustrate these techniques on a simple problem using the bar elements developed in previous
Chapters.
27.5. The LPB Eigensystem
The two key results from the LPB assumptions (which are studied in the next Chapter) can be summarized as follows. Recall from Chapters 815 that the tangent stiffness matrix can be decomposed
as the sum of material and geometric stiffness matrices:
K = K M + KG .

(27.6)

Then the LPB assumptions lead to the following simplications:

(1) The material stiffness is the stiffness evaluated at the reference conguration:
K M = K0 .

(27.7)

(2) The geometric stiffness is linearly dependent on the control parameter :

KG = K1 .

(27.8)

in which K1 is constant, and is also evaluated at the reference conguration.

Now the stability test (27.3) requires that K be singular, which leads to the eigenproblem
K zi = (K0 + i K1 ) zi = 0.

(27.9)

This is called the LPB stability eigenproblem. The eigenvalue i closest to the stage start = 0 is
the critical control parameter cr .
In the following Chapter we shall prove that under certain restrictions the critical states determined
by solving (27.9) bifurcation points and not limit points. That is, they satisfy the orthogonality test
ziT q = 0.
277

(27.10)

278

Chapter 27: BIFURCATION: LINEARIZED PREBUCKLING I

P=q
(2) k (2)

(2)

Y, y
(1)

L0

(1)

L0

;
3

(1)

X, x

;;

(2)

<< k

(1)

Figure 27.2. LPB example involving two bar elements displacing on the X x, Y y
plane.

for all i. The eigenproblem (27.9) bets the generalized symmetric algebraic eigenproblem
Axi = i B xi ,

(27.11)

in which both matrices A K0 and B K1 are real symmetric, and x z are the buckling
mode eigenvectors. If (as usual) the material stiffness K0 is positive denite, eigensystem theory
says that all eigenvalues of (27.11) are real. We cannot in general make statements, however, about
the sign of these eigenvalues. That will depend on the physics of the problem as well as on the sign
conventions chosen for the control parameter(s).
27.6. Stability Eigenproblem
In production FEM codes, the stability eigenproblem (27.9) is generally treated with special solution
techniques that take full advantage of the sparsity of both K0 and K1 ; for example subspace iteration
or Lanczos-Arnoldi type of methods.
For small systems an expedient solution method consists of reducing it to canonical form by
premultiplying both sides by the inverse of K0 . This is possible if K0 is nonsingular, which means
that the = 0 conguration is not a critical one. Renaming for convenience A = K1 K1 and
0
i = 1/i one gets
(27.12)
Azi = i zi
This is a standard algebraic eigenproblem, which can be solved by standard library routines for the
eigenvalues i and eigenvectors zi . For example, Eigensystem in Mathematica or eig in Matlab.
The i farthest away from zero gives the i = 1/i closest to zero.
One disadvantage of this reduction is that A is unsymmetric even if K0 and K1 are. There are
more complicated reduction methods that preserve symmetry if at least one of those matrices is
nonnegative denite. These may be studied in standard numerical analysis textbooks covering linear
algebra; for example Golub and Van Loan [269]. But in production FEM programs the original
eigensystem (27.9) is treated directly.
278

279
27.7.

27.7 LPB ANALYSIS EXAMPLE

LPB Analysis Example

To illustrate the application of LPB to a very simple example, the 2-bar assembly shown in Figure 27.1 is chosen.
The bars can only displace on the x, y plane, thus the problem is two dimensional. The equivalent-spring
stiffness of the bars is denoted by
E A(1)
0

k (1) =

L (1)
0

k (2) =

E A(2)
0
L (2)
0

(27.13)

in which A(e) and L (e) denote the cross sectional areas and lengths, respectively, of the eth bar in the reference
0
0
conguration, and E is the elastic modulus common to both bars.
The gure shows the reference conguration C0 for the two bars. That conguration is taken when the applied
load is zero, that is, = 0. We shall assume that the stiffness of bar 1 is much greater than that of bar 2, i.e.,
k (1) >> k (2) and is such that the vertical displacement u Y 2 of node 2 under the load is very small compared to
the dimensions of the structure.
Now let the load p = q be gradually applied by increasing . The structure assumes a deformed current
conguration in equilibrium, that is, a target conguration C. According to the LPB basic assumption, the
displacements prior to the buckling load level characterized by cr are negligible. Therefore C C0 as long
as || < |cr |.
The linear nite element equations for the example problem are as follows. For element (1):

0
0
k (1)
0
0

0
1
0
1

For element (2):

1
0
k (2)
1
0

0
0
0
0

0
u X1
0
1 u Y 1 0
=
.
0 u X2 0
uY 2
1
q

(27.14)

1 0
u X2
0
0 0 uY 2 0
=
.
1 0 u X3 0
uY 3
0 0
0

0
0
0
0

(27.15)

Assembling and applying the boundary conditions u X 1 = u Y 1 = u X 3 = u Y 3 = 0 we get

k (2)
0

0
k

u X2
uY 2

(1)

The linear solution is

u X 2 = 0,

uY 2

0
.
q

(27.16)

q L (1)
q
0
= (1) =
.
(1)
k
E A0

(27.17)

The axial linear strain and Cauchy (true) stress developed in element (1) are
(1)

uY 2
L (1)
0

q
E A(1)
0

(1) = E

(1)

q
A(1)
0

(27.18)

According to the assumptions stated above the change in geometry prior to buckling is neglected. Consequently
e(1)

(1)

s (1) (1)

279

(27.19)

2710

Chapter 27: BIFURCATION: LINEARIZED PREBUCKLING I

The axial strain and stress of element (2) are zero.

The simplied nonlinear nite element equations are, for element (1)

0
0
k (1)
0

0 0
0
1
N (1) 0
1 0 1
+
0 0
0 L (1) 1
0
1 0
1
0

1
0
1
0

0
1
0
1

0 u X1
0

1 u Y 1 0
=
.
0 u X 2 0

uY 2
1
q

(27.20)

where N (1) = A(1) s (1) = p = q denotes the axial force in bar element (1).
0
For element (2) we have the same linear matrix equations as before because its geometric stiffness vanishes.
Assembling and applying displacement boundary conditions we get the equations

k (2)

q
L (2)
0

0
k

(1)

q
(1)
L0

u X2 =
uY 2

0
.
q

(27.21)

One now regards K in (27.21) as unaffected by the displacements u X 2 and u Y 2 , which is consistent with the
assumption that the change of geometry prior to buckling is neglected. This having being done, setting the
determinant of K to zero yields the buckling eigenproblem:

det

k (2)

q
L (1)
0

k (1)

0
q
L (1)
0

= 0.

(27.22)

This matrix is singular if either diagonal element vanishes, which yields the two eigenvalues
cr 1 = k (1) L (1) /q,
0

cr 2 = k (2) L (1) /q,

0

(27.23)

as critical values of the load parameter. Since k (2) << k (1) the lowest critical load will be
pcr = cr 2 q = k (2) L (1) .
0
This is the buckling load obtained under the LPB assumptions.

2710

(27.24)

2711

27.8

SUMMARY OF LPB STEPS

27.8. Summary of LPB Steps

The foregoing example illustrates the key steps of LPB analysis. Those are summarized below for
completeness.

1.

Assemble the linear stiffness K0 and solve the linear static problem
K0 u = q0 ,

(27.25)

for = 1 and obtain the internal force (stress) distribution. Note: In statically
determinate structures, such as Exercise 27.4, the internal forces and stresses may
be obtained directly from equilibrium. However K0 is still necessary for step 3.
2.

Form the reference geometric stiffness K1 for that internal force distribution. The
geometric stiffness is KG = K1 .

3.

Solve the stability eigenproblem

(K0 + K1 ) zi = 0,

or

K0 zi = i K1 zi ,

(27.26)

The eigenvalue i closest to zero is the critical load multiplier, and the associated
eigenvector zi gives the corresponding buckling mode.

2711

2712

Chapter 27: BIFURCATION: LINEARIZED PREBUCKLING I

Homework Exercises for Chapter 27

Bifurcation: Linearized Prebuckling I
EXERCISE 27.1 [A:15] Find the buckling mode (null eigenvector) z (normalized to unit length) for the

example problem of 27.6, and verify the orthogonality condition zT q = 0.

EXERCISE 27.2 [A:15] Find the buckling load for the two-bar problem if bar (2) forms an angle 0 < 90

with the x axis. Assume still that the system displaces only on the x y plane and that k (2) << k (1) so that
the stress in bar (2) can be neglected in forming the geometric stiffness. Determine z and verify orthogonality.
EXERCISE 27.3 [A:20] Suppose the load of the two-bar example problem of 27.6 depends on the vertical
displacement of point 2 as p = cu 2 2 , where c is a constant with dimensions of stress. Show that even if
Y
prebuckling deformations are neglected, the singular stiffness test leads to a nonlinear eigenvalue problem.
EXERCISE 27.4 [A:25] The Euler column shown in Figure E27.1 is modelled by one 2-node Euler-Bernoulli

beam column element along its length:

Y, y

;
;
z

E, I constant

;;
2

X, x
L

Figure E27.1. One-element model of Euler column.

The state variables are the nodal displacements degrees of freedom arranged as

X1

uY 1

u = z1
u X2

(E27.1)

uY 2
z2
where z1 and z2 are (to rst order) the end rotations, positive counterclockwise about z.
The linear material matrix in the reference (undeformed) conguration is

EA
L

K0 =

0
12E I
L3

0
6E I
L2
4E I
L

EA
L

symm

2712

0
0
EA
L

0
12E I
L3
6E I
2
L
0
12E I
L3

0
6E I
L2
2E I
L
0
6E I
2
L
4E I
L

(E27.2)

2713

Exercises

in which E is the elastic modulus, L the element length, A the cross section area, and I the moment of inertia
of the cross section about the z neutral axis.
The exactly-integrated geometric stiffness at the reference conguration is5

KG = K1 ,

P
K1 =
30L

0
36

0
3L
4L 2

0
0
0 36
0 3L
0
0
36

symm

0
3L
L 2

3L
4L 2

(E27.3)

where N is the axial force in the element (here obviously equal to the applied force P because the structure
is statically determinate.)
For this problem:
Check that K0 and K1 satisfy translational innitesimal rigid body motion conditions u X 1 and u y 1
if the six degrees of freedom are left unconstrained. (Convert those modes to node displacements, then
premultiply by the stiffness matrices.)

(a)

buckling mode

z1

2
z2 = z1

Figure E27.2. Symmetric buckling of one-FE model of Euler column.

(b)

Set up the linearized prebuckling eigenproblem

(K0 + K1 )z = 0

(E27.4)

Apply the support end conditions to remove u X 1 , u Y 1 and u Y 2 as degrees of freedom.

(c)

Justify that freedom u X 2 can be isolated from the eigenproblem, and proceed to drop it to reduce the
eigenproblem to 2 2.

(d)

Reduce the 2 2 eigenproblem to a scalar one for the symmetric buckling mode sketched in Figure
E27.2. by setting z1 = z2 (see Figure), and get the rst critical load parameter 1 .

(e)

Repeat (e) for the antisymmetric buckling mode sketched in Figure E27.3 by setting z1 = z2 (see
Figure) and obtain the second critical load parameter 2 .

(f)

Compare the results of (d)(e) to the exact critical load values

P1E = 2

EI
,
L2

See e.g., Przemienieckis book [504].

2713

P2E = 4 2

EI
,
L2

(E27.5)

2714

Chapter 27: BIFURCATION: LINEARIZED PREBUCKLING I

z1

z2 = z1

buckling mode
Figure E27.3. Antisymmetric buckling of one-FE model of Euler column.

The rst one was determined by Euler in 1744 and therefore is called the Euler critical load. For P1E the
FEM result should be within 25%, which is good for one element.
(g)

Repeat the calculations of 1 and 2 with the following reduced-integration geometric stiffness matrices6

K
1

P
=
8L

symm
0

K =
1
24L

0
9

0
24

0
3L
L2

0
0
2L 2

symm

0
0
0 9
0 3L
0
0
9
0
0
0
0

0
3L
L 2

3L
L2

0
0
24
0
0
2L 2

0
0

24
0
2L 2

(E27.6)

(E27.7)

and comment on the relative accuracy obtained against the exact values.
(h)

Repeat the calculations of steps (b) through (e) for a two equal-element discretization. Verify that
the symmetric-mode buckling load is now 10E I /L 2 , which is (surprisingly) close to Eulers value
P1E = 2 E I /L 2 .

EXERCISE 27.5 [A:25] This is identical in all respects to Exercise 27.4, except that the 2-node Timoshenko
beam model is used, with Mac Nealss RBF correction for the material stiffness matrix. The net result is
that K0 is identical to (E27.2) but K1 is different. In fact K1 is given by Equation (9.45) where V = 0
and N = P. Note: dont be surprised if the one-element results are poor when compared to the analytical
buckling load.
EXERCISE 27.6 [A/C: 20] The column shown in Figure E27.4 consists of 3 rigid bars of equal length L
connected by hinges and stabilized by two lateral springs of linear stiffness k. The applied axial load is
P, where > 0 means compression. Compute the two buckling loads 1 P and 2 P in terms of k and L,
assuming the LPB model of innitesimal displacements from the initial state. Show that one load corresponds
to a symmetric buckling mode and the other to an antisymmetric buckling mode, and nd which one is critical.

Hint. The two degrees of freedom are the small lateral displacements u 1 and u 2 of hinges 1 and 2, where
u 1 << L, u 2 << L. Write the total potential energy as
P 2
= U W,
U = 1 ku 2 + 1 ku 2 ,
W =
u + (u 1 u 2 )2 + u 2 .
(E27.8)
1
2
2
2
2
2L 1
6

These matrices are obtained by one-point and two-point Gauss integration, respectively, whereas the K1 of (E27.3) is
obtained by either 3-point Gauss or analytical integration.

2714

2715

;;
;;
;;

Exercises

;
;
;
;

A
k

All 3 bars are

considered rigid

;;

Figure E27.4. Buckling of a segmented-hinged column propped by two springs.

Explain where the expression of W comes from. Once

2715

is in hand, it is smooth sailing.