Linear Algebra Application Example

Stress Analysis
As you have learned from CVE 220 and/or MCE 301, when an elastic body is
subjected to applied loadings, stresses are created inside the body. In general
these stresses often vary in complicated ways from point to point and from plane
to plane within the structure. To help characterize this situation, stresses are
normally defined with respect to a given coordinate system. The illustration
below shows that at a typical point within a loaded body, the state of stress can
be characterized on a small cube of material defined with respect to a Cartesian
coordinate system. These nine components are called the stress components,
with x, y, z referred to as normal stresses and xy, yx, yz, zy, zx, xz called the
shearing stresses.

P3 y
P2

yz yx
xy
zy
y x
zx xz
z
p
x
P1
(Externally Loaded Body) z (State of Stress at a Point)

Because of the number of components and some other transformation

properties, the stress can be expressed as a 3 x 3 matrix

x xy xz

[ ] yx y yz (1)
zx zy z

and since the shearing stresses have the equalities xy = yx , yz =zy , zx = xz , the
stress matrix is symmetric.

Again from your previous courses, you studied how the stresses vary by
orientation of the plane of action. For example, as we change the orientation of a
particular plane the normal stress component (say x) will vary. There exists a
special orientation where the normal stress will be a maximum, and these are
called principal planes and the normal stresses acting on them are called the
principal stresses. It turns out that for the general three-dimensional case, the
theory to determine principal stresses and the planes on which they act is
formulated by the eigenvalue problem

[ ]{n} {n} (2)

where [] is the stress matrix, {n} is the principal direction vector and (the
eigenvalue) is the principal stress. Thus solving the eigenvalue problem will
determine up to three distinct principal stresses and the corresponding three
principal directions. It turns out for this application (3 x3, symmetric real matrix)
the principal directions are mutually orthogonal. Again you should remember
that the shear stress components will vanish on these three principal planes and
so for a coordinate system that is aligned with the principal directions the stress
matrix takes on the simplified diagonal form

1 0 0
[ ] 0 2 0 (3)
0 0 3

where 1, 2, 3 are the problems eigenvalues (roots of the characteristic

equation) and are generally referred to as the principal stresses. This concept is
graphically illustrated by showing the state of stress in the original given system
and in the principal coordinate system which will be rotated to align with the
principal directions as shown below

y
y 2
2
1

yz yx
xy 1
zy
x
zx xz
3
z x

z 3
(General Coordinate System) (Principal Coordinate System)

Using various methods like mechanics of materials, theory of elasticity, or finite

elements, the stresses can be determined with respect to some x,y,z coordinate
system. However, to further understand the possible failure of the structure, it is
often necessary to find the principal stresses and see how they are distributed.
Trying to solve the eigenvalue problem (3) for many different locations within the
body is beyond hand calculation and thus numerical/computational methods
must be used. MATLAB can easily handle this problem and we now look at a
simple example problem

Two-Dimensional Beam Example

Consider the following two-dimensional example of a simply-supported beam
carrying a concentration force at mid-span. Assume the beam has unit thickness.

y
P

x
2c

2L
From mechanics of materials theory the in-plane stress components for the right
half of the beam (x > 0) are given by

3P 3P 2
x 3
( L x ) y , y 0 , xy 3
(c y 2 ) (4)
4c 8c
Thus the stress matrix for this problem reduces to the 2x2 form

3 3P

x ( L x ) y ( c 2
y 2
)
xy 2c 3
8c 3
[] 3P

xy y 3 (c y )
2 2
0
(5)
8c
We can now employ MATLAB to calculate the principal stresses and make a
contour plots of the distribution in the region 0 < x < L, -c < y < c. Contour plots
are shown below along with the MATLAB code.
First Principal Stress Contours, 1

0.4

0.2

-0.2

-0.4

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

Second Principal Stress Contours, 2

0.4
0.3
0.2
0.1
0
-0.1
-0.2
-0.3
-0.4

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

% MCE 372 Engineering Analysis Example Code
% Eigenvalue Problem Application
% Determine and Plot Principal Stresses in Beam Example
clc;clear all;clf;
% Input Beam Parameters
c=0.5;L=5;P=1;
% Set up grid of x and y values
x=[0:0.1:L];y=[-c:0.01:c];
[X,Y]=meshgrid(x,y);
% Calculate Individual Stresses
sx=-(3/(4*c^3))*(L-X).*Y;
sy=zeros(length(y),length(x));
txy=-(3/(8*c^3))*(c^2-Y.^2);
% Create Stress Matrix
for i=1:length(y)
for j=1:length(x)
s=[sx(i,j),txy(i,j);txy(i,j),sy(i,j)];
% Determine Eigenvalues(Principal Stresses)
p=eig(s);
s1(i,j)=p(2);
s2(i,j)=p(1);
end
end
% Plot Distributions of Principal Stresses s1 and s2
figure(1)
contour(X,Y,s1,[0.01,0.05,0.1,0.5,1,3,5,7,9,11])
title('First Principal Stress Contours, s1')
axis tight
figure (2)
contour(X,Y,s2,[-0.01,-0.05,-0.1,-0.5,-1,-3,-5,-7,-9,-11])
axis tight
title('Second Principal Stress Contours, s2')