STRUCTURAL STABILITY

CE 5251
Dr. T. P. Tezeswi
Assistant Professor
Department of Civil Engineering
NIT-Warangal
email: tezeswi@nitw.ac.in
Syllabus

Code provisions for structural stability

 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
The matrix method is a numerical technique that uses matrix algebra to analyze structural
systems. It idealizes the system as an assembly of discrete elements connected to one
another at points called nodes.

The beam below is shown subdivided into two elements, elements ab and bc.

• Associated with each element are generalized displacements, δ, used to describe its
deformation, and the corresponding generalized forces, q.
• The term generalized denotes the fact that q can be a moment as well as a force and δ
can be a rotation as well as a deflection.
• The q's are internal forces as far as the overall structure is concerned, but they are
external forces when individual elements are considered. Since they refer to the
element, the q's are called element forces, and the δ’s are called element
displacements. [1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
The load-deformation characteristics of an element can be given by means of the
matrix equation:

(1)

in which the stiffness influence coefficient kij is the force q, due to a unit
displacement δj when all other δ 's are zero

The matrix made up of these influence coefficients is called the element stiffness
matrix.

[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
A relation similar to Eq. (1), expressing the load-deformation characteristics of the
overall structure, is:

Force Displacement
matrix Stiffness matrix
matrix

(2)

(3)
[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
The analysis of a structure by the matrix method consists of two steps:

1) The structure stiffness matrix [K] (a.k.a. global stiffness matrix) is synthesized
from the stiffness matrices of the individuaI elements (element stiffness matrix).
This is usually accomplished by applying at each node the conditions of
equilibrium and deformation compatibility.

2) The structure displacements and the external reactions are obtained from the
structure stiffness matrix [K] and the Ioad matrix [W].

[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
Formation of Element Stiffness Matrices
To form the structure stiffness matrix, the stiffness matrices of the individuaI elements
must first be constructed. The stiffness matrix for element ab will be of the form given
by Eq. (1)

The first column of the matrix is obtained by applying a unit translation, δ1 , to the Ieft
end of the element and keeping δ2 , δ3 , and δ4 at zero

Components
of element
stiffness
matrix

[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
Formation of Element Stiffness Matrices : Components of element stiffness matrix

According to the definition of a stiffness influence coefficient, k 21 and k 41 are the

moments induced at the Ieft and right ends of the member by the unit translation
δ1

From the slope deflection equation :

(4)

The shears, k12 and k31 that are also induced by the unit translation δ1 are
obtained by considering moment equilibrium of the element as:

(5)

[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
The entire stiffness matrix, for element ab :

(6)

The stiffness matrix for element bc i.e, [k]bc , is identical to [k]ab

[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
Before synthesis of the structure stiffness matrix [K] from the stiffness matrices of the
individual elements, it is useful to combine them into a single composite element
stiffness matrix [k].

(7)

[q] = [k][δ] (8)

Note: This is not the assembled global stiffness matrix for the entire structure ! [1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
Formation of Structure Stiffness Matrix: Direct Stiffness Method

Direct Stiffness Method: transform each of the element stiffness matrices

individually into structure coordinates, and then combine the transformed
matrices into a structure stiffness matrix.

• The synthesis of [K] (structure/global stiffness matrix) from [k] (element stiffness
matrix) can be seen as a transformation of coordinates.

• The element stiffness matrix is written in terms of element deformations, and

the structure stiffness matrix in terms of structure deformations.

• A transformation from one matrix to the other, can be considered to be a

transformation from element to structure coordinates:

(9)

• The matrix [B], which is used to carry out the transformation, is commonly
referred to as the transformation matrix. It is made up of the direction cosines
between the structure and element coordinate axes.
[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
The matrix [B] transforms structure deformation vectors into element deformation
vectors
(10)

Since the element and structure force vectors coincide with the corresponding element
and structure deformation vectors, the forces must transform in precisely the same
manner at the deformations.
(11)

To obtain [K] (structure/global stiffness matrix) for the beam, transformation matrices
are required for each of the two elements. These matrices are obtained from equations
that relate the element to the structure coordinates as:
(12)

The terms in the transformation matrices are the direction cosines between the element
and structure deformation vectors. Since these vectors coincide for the structure being
considered, the transformation matrices are both identity matrices. [1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
Using the transformation matrices [B]ab and [B]bc and carrying out the
operations indicated by Eq. (9) on each of the two element stiffness matrices
[k]ab and [k]bc one obtains

(13)

[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
The stiffness matrix [K] of the entire structure (global stiffness matrix) is now constructed
by combining [K]ab and [K]bc , i.e., by placing the individual terms from [K]ab and [K]bc , into
their correct positions in the matrix [K].

From the law of matrix multiplication, the row and column of any term in [K]mn is obtained
from the corresponding row in [B]Tmn and the corresponding column in [B]mn

(14)

Note: The dashed lines indicate the manner in which the element matrices have been combined to form the structure matrix.
Whenever a term such as W3 , Δ3 appears in two element matrices, the corresponding terms in the structure matrix is equal to the [1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
Structure Deflections and Element Forces
• Equation (14) relates all possible structure nodal displacements to the corresponding
structure node point loads.
• Before attempting to solve for any of these displacements, it is convenient to
distinguish those that are unknown and those that are fixed as a result of the
boundary conditions. This is accomplished by rearranging and then partitioning the
deformation matrix [Δ] into two submatrices [Δ]a and [Δ]b.
• Included in [Δ]a are the unknown nodal displacements and in [Δ]b the nodal
displacements prescribed by the boundary conditions.
• The load matrix [W] and the stiffness matrix [K] are then rearranged and partitioned
accordingly as:

(15)

• The submatrix [W]a contains the loads corresponding to the unknown nodal
displacements and the submatrix [W]b the reactions at the known nodal
displacements.

[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
In many problems the nodal displacements prescribed by the boundary conditions
are all zero.
Equation (15) can then be separated into the following two relations:
(16-a)
(16-b)

If one multiplies both sides of Eq. (16) by [Kaa]-I, one obtains:

(17)
from which the unknown nodal displacements [Δa] can be determined. Once [Δa] has
been obtained, the reactions [Wb] can be found using Eq. (16-b).
For the fixed-end beam, Δ1 = Δ2 = Δ5 = Δ6 = 0. Consequently, [Δa] consists of Δ3 and Δ4,
and Eq. (16-a) takes the form :

(18)

[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
Inverting the stiffness matrix and noting that W4 = O, one obtains:

(19)

from which the midspan deflection Δ3 is found to be:

(20)
Having obtained [Δa], the reactions [Wb] can be determined using Eq. (16-b). Thus

the fixed-end moments W2 and W6 are:

(21)
(22) [1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD-Compression Members
• The stiffness method utilized previously to analyze a flexural member will now be used
to study the behavior of a member that is subject to axial loading as well as bending.

• Within the limitations of small displacements and elastic stresses, the stiffness of a
flexural member is a constant, and the linear relationship [Q] = [K][Δ] describes its
behavior.

• The stiffness of a member subject to both axial load and bending is a function of the
axial load. The force-deflection relation of such a member takes the form
(23)

in which [Q] contains the transverse loads that cause bending, [Δ] contains the
corresponding bending deformations, and P is the axial load.

• [K] = the standard stiffness matrix of a member subject only to flexure

• [K1]= accounts for the effect that the axial load P has on the stiffness of the flexural
member.
Equation (23) can be used to obtain the deflection of a beam column.

[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
In addition, the equation (23) can be used to obtain the critical load of an axially loaded
member.
By definition, the critical load is that axial load at which the bending stiffness of the
member vanishes.

From rearranging eqn (23) as:

• The bending stiffness vanishes; that is, [Δ] increases without bound for finite values of
[Q] only when the inverse of the stiffness matrix becomes infinitely large.

• Since the inverse of a matrix is obtained by dividing the adjoint matrix by the
determinant, the inverse will blow up when the determinant vanishes.

• The critical load can thus be found by setting the determinant of the stiffness matrix
equal to zero.

[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
Element Stiffness Matrix for Beam Column
Consider an element of a beam column subject to an axial load P and a set of loads [q],
as shown below.

• It is our purpose to find a matrix relationship between the loads [q] and the
deformations [δ] in the presence of the axial load P.
• As long as the deformations are small and the material obeys Hooke's law, the
deformations corresponding to a given set of loads [q] and P are uniquely determined,
regardless of the order of application of the loads.
• The deformations [δ] can therefore be determined by applying first the entire axial
load P and then the loads [q].
• Under these circumstances the relation of [q] to [δ] is linear, and the stiffness matrix
can be evaluated using the principle of conservation of energy. [1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
Element Stiffness Matrix for Beam Column
• The element is assumed to be loaded in two stages. During the first stage only the
axial load P is applied, and during the second stage the element is bent by the [q]
forces while P remains constant.

• Since the element is in equilibrium at the end of stage one as well as at the end of
stage two, the external work must be equal to the strain energy not only for the
entire loading process but also for stage two by itself.

• The external work corresponding to the second loading stage is :

(24)

in which the first term represents the work of the [q] forces and the second term the
work due to P.
• Since the ends of the member approach each other during bending, the axial force
does positive work when it is a compression farce and negative work if it is a
tension farce.

[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
The strain energy stored in the member during stage two is due only to bending. Thus

(25)

Equating the strain energy to the external work gives

(26)

Making use of the relationship [q] = [k][δ], in which [k] is the element stiffness matrix,
Eq. (26) becomes

(27)

To evaluate [k], it is necessary to put the right-hand side of Eq. (27) into matrix form. This
can be accomplished if the deflection y is assumed to be given by

(28)
The choice of a deflection function is an extremely important step. A cubic is chosen in this
instance because such a function satisfies the conditions of constant shear and linearly
varying bending moment that exist in the beam element. [1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD

Taking the coordinate axes in the directions shown in Fig. above, the boundary
conditions for the element are

Substitution of these conditions into Eq. (28) makes it possible to evaluate the four
arbitrary constants and to obtain the following expression for y:

(29)
[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
Equation (29) can be rewritten in matrix form as

(30)

Differentiating eqn (30) gives

(31-a)
(31-b)

Where: (32-a)

(32-b)

From eqn(31) : (33-a)

(33-a)
[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
Substitution of these relations into (27) gives:

(34)

From eqns (32) & (34) we get the stiffness matrix of a beam column element:

(35)

conventional stiffness matrix representing the

matrix of a flexural effect of axial loading on the
element bending stiffness [1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
Calculations involving the stiffness matrix of a beam column element are considerably
simplified if every term in the matrix has the same dimensions. This can be accomplished
by modifying the force and displacement matrices so that the terms in the former all have
the dimension of force and those in the Iatter the dimension of distance. Thus
[k ] [kg]
e

(36)

conventional stiffness
matrix representing the
matrix of a flexural
effect of axial loading on the
element
bending stiffness

This form of the stiffness matrix is especially useful when the stiffness matrix must be
inverted
[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
Calculation of the Critical Load

The member has a length L and a uniform flexural rigidity El and for the purpose of
the analysis is subdivided into two elements. [1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
The stiffness matrix of each element can be constructed in accordance with Eq (36)

(37-a)

(37-b)

[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
The structure stiffness matrix for the entire coIumn will be obtained by transforming the
individuaI element stiffness matrices from element to structure coordinates and then
combining the resuIting matrices.

The transformation matrices that relate the structure deformations to the element
deformations are

[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
Carrying out for each element the transformation of coordinates in accordance with Eqn
(9), one obtains:

(38-a)

(38-b)

Note: Since the transformation matrix for each element is an identity matrix, the transformation from element to
structure coordinates does not bring about a change in the magnitude of any of the terms inside the stiffness
matrices. However, these terms now represent structure loads per unit structure deformations, whereas they
represented element forces per unit element deformations before. [1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
structure stiffness matrix for the entire member is given by:
[Kg]
[Ke]

(39)

This matrix is valid for any boundary conditions. Using it, the critical load will now be
determined for both a fixed-fixed and a hinged-hinged member

[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
Fixed-fixed column

For a fixed-end column, all Δ's with the exception of Δ3 vanish. Consequently, the
stiffness matrix given in Eqn (39) reduces to:

(40)

At the critical load the bending stiffness of the member vanishes; that is, Δ3 becomes
infinitely large for finite values of W3

Hence the quantity inside the parentheses must vanish at the critical load. That is,

The exact solution to this problem is Pcr = 39.44 EI/L2

[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
Hinged-hinged column
For a hinged-hinged column, Δl = Δ4 = Δ5 = O. Crossing out the columns and rows of
the stiffness matrix corresponding to the zero deformations gives

(41)

By recognizing that Δ6 = - Δ2 , the stiffness matrix can be further reduced to

[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
Hinged-hinged column
The criterion for finding the critical load is that the bending stiffness must vanish, or
the inverse of the stiffness matrix must blow up. Hence the denominator of [K]-l,
which is the determinant of [K], must be equal to zero. That is,

(42)

To simplify the numericaI computations, divide each term by El/P and introduce the
notation

Eqn (42) then becomes

(43)

Expansion of Eqn (43) leads

to the characteristic eqn: Whose smallest root is λ =0.49

(44) This gives : Exact solution is Pcr = 9.87EI/L2

[1]
 Eigenvalue Problem: Discrete Systems
In an eigenvalue buckling problem we look for the loads for which the model stiffness
matrix becomes singular, so that the problem has nontrivial solutions.
For a system with n degrees of freedom (DOF), the eigen-system has the following form

Ax = λBx
Where: A and B are square matrices of order n,
the n eigenvalues λi are the critical load factors
the eigenvectors xi are the associated buckling shapes

•Generally only the smallest eigenvalue, which will be denoted by λ1, will be of interest for
engineering design.
•Matrices A and B may be obtained by either equilibrium analysis, or by energy
methods
•If equilibrium approach based on doing Free Body Diagrams (FBD) on a slightly
perturbed configuration is used, the matrices A and B are not necessarily symmetric, but
may be symmetrized if necessary.
•If the energy method is used, the symmetry of both matrices is guaranteed from the
start.
[1]
 Eigenvalue Problem
•How to find P? Solve the eigenvalue problem.
[A]{y}+[B]{y}={0}
•Standard Eigenvalue Problem
[A]{y}={y}
Where,  = eigenvalue and {y} = eigenvector
•Can be simplified to [A-I]{y}={0}
•Nontrivial solution for {y} exists if and only if
| A-I|=0
•One way to solve the problem is to obtain the characteristic
polynomial from expanding | A-I|=0
•Solving the polynomial will give the value of 
•Substitute the value of  to get the eigenvector {y}

Note: This is not the best way to solve the problem, and will not
work for more than 4or 5th order polynomial
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
Eigen value problem Hinged-hinged column
[Ke-λKg]Ø = 0

𝜟𝟐 l
=𝟎
𝜟𝟑
Substituting λ =0.49
3.6733 5.7550 Δ2 l
=0
11.5100 18.1200 Δ3

Setting Δ3 =1.00 and solving we get : Δ2 =3.1475/ L Δ6 = − 3.1475/ L

Ø = [3.1475/ L , 1.00, - 3.1475/ L ]

1.00

[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
STABILlTY OF A FRAME
Stability of the simple portaI frame

Sidesway buckling

[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
STABILlTY OF A FRAME
Each member of the frame has a length L and stiffness El, and the frame is fixed at the
base, free to move laterally at the top, and loaded as indicated in the figure. For the
sake of simplicity we let each member consist of only a single element.

The stiffness matrices for the columns are given by

(45)

and the stiffness matrix for the beam takes the form

(46)

[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
STABILlTY OF A FRAME
To obtain the structure stiffness matrix, the element stiffness matrices are
first transformed to structure coordinates and then combined. The transformation
from element to structure coordinates is carried out as:
(47)

Where [Bn] is the transformation matrix for the nth element, defined as:
(48)
According1y, the transformation matrices for elements l, 2, and 3 of the frame are:

[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
STABILlTY OF A FRAME
Using these transformation matrices and carrying out the operations, the stiffness
matrices become:

(49)

[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
STABILlTY OF A FRAME
The structure stiffness matrix obtained by combining the transformed element
stiffness matrices is:

(45)

Letting :
The structure stiffness matrix reduces to:

At the critical load the determinant of the stiffness matrix must vanish. Expanding the
determinant of the structure matrix and setting the resulting expression equal to zero gives:

The smallest root of this equation is λ1 = 0.25 from which:

This result is very close to the exact value of 7.34EI/l2 for the critical load in the sidesway
mode.
[1]
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
STABILlTY OF A FRAME: Eigenvalue problem
Substituting λ1 = 0.25 into the structure stiffness matrix for non-trivial Ø
[Ke-λKg]Ø = 0

Δ1
Δ2 =0
Δ3

7.7500 2.0000 −5.7500 Δ1

2.0000 7.0000 −5.2500 Δ2 =0
−5.2500 −5.2500 6.0000 Δ3
Setting Δ3 = 1.00 and observing that 𝜟𝟏 = - 𝜟𝟐 by
solving we get:
Δ1 = -Δ2 = 0.2667

Δ1 Δ2 Δ3
Ø = [ 0.2667 l, -0.2667 l, 1.00]

[1]
 Eigenvalue Problem: Numerical Solution
Calculation of Buckling Loads by the Finite Element Method
•The calculation of the linear bifurcation point is based on the eigenvalue problem
constituted from the linear stiffness matrix [ K] and the geometrical stiffness matrix [Kσ ]
as:
([Ko] +λ [Kσ]) δ {D} =0
=

[KT] = [K]+ [Kσ] = Tangent stiffness matrix

[Ko] contains the stiffness of the structure
[Kσ] contains the membrane forces
where the smallest root λ defines the smallest level of load where bifurcation occurs
 Eigenvalue Problem: Numerical Solution
•Inverse Power Method: Numerical Technique to Find Least
Dominant Eigenvalue and its Eigenvector
•Based on an initial guess for eigenvector and iterations Algorithm
1) Compute [E]=-[A]-1[B]
2) Assume initial eigenvector guess {y}0
3) Set iteration counter i=0
4) Solve for new eigenvector {y}i+1=[E]{y}i
5) Normalize new eigenvector {y}i+1={y}i+1/max(yji+1)
6) Calculate eigenvalue = 1/max(yji+1)
7) Evaluate convergence: i+1-i < tol
8) If no convergence, then go to step 4
9) If yes convergence, then = i+1 and {y}= {y}i+1
 FRAME BUCKLING: Eigenvalue Problem
STIFFNESS MATRIX METHOD
Online Resources:
• https://nptel.ac.in/courses/105/108/105108141/ (Lectures 26 to 36 on Structural
Stability Analysis)

• https://www.mathworks.com/matlabcentral/fileexchange/30970-natural-
frequencies-buckling-loads-of-columns-using-finite-element-method
 References
• [1] Principles Of Structural Stability Theory - Alexander Chajes , Prentice-
Hall Inc., Englewood Cliffs, New Jersey