Chapter

Thermal Stress
15
Chapter Objectives

■ To formulate the thermal stress problem.

■ To derive the thermal force matrix for a one-dimensional bar.
■ To derive the thermal force matrix for the three-noded triangle for both plane stress
and plane strain.
■ To solve examples of bars, trusses, and plane stress problems for thermal stresses
due to temperature change.
■ To show a finite element computer result for the thermal stress solution of a plate
subject to temperature change.
■ To demonstrate a finite element computer result for the thermal stress solution of a
three-dimensional object subjected to temperature change.

Introduction
In this chapter, we consider the problem of thermal stresses within a body. First, we will discuss
the strain energy due to thermal stresses (stresses resulting from the constrained motion of a
body or part of a body during a temperature change in the body).
The minimization of the thermal strain energy equation is shown to result in the thermal
force matrix. We will then develop this thermal force matrix for the one-dimensional bar ele-
ment and the two-dimensional plane stress and plane strain elements.
We will outline the procedures for solving both one- and two-dimensional problems and
then provide solutions of specific problems, including illustration of a computer program used
to solve thermal stress problems for two- and three-dimensional stress problems.

15.1 Formulation of the Thermal Stress

Problem and Examples
In addition to the strains associated with the displacement functions due to mechanical loading,
there may be other strains within a body due to temperature variations, swelling (moisture
­differential), or other causes. We will concern ourselves only with the strains due to temperature
variation, «T , and will consider both one- and two-dimensional problems.

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 728 15 | Thermal Stress

a1
a2

■■Figure 15 –1  Composite member composed of two materials with different coefficients of
thermal expansion

Temperature changes in a structure can result in large stresses if not considered properly
in design. In bridges, improper constraint of beams and slabs can result in large compressive
stresses and resulting buckling failures due to temperature changes. In statically indeterminate
trusses, members subjected to large temperature changes can result in stresses induced in mem-
bers of the truss. Similarly, machine parts constrained from expanding or contracting may have
large stresses induced in them due to temperature changes. Composite members made of two
or more different materials may experience large stresses due to temperature change if they are
not thermally compatible; that is, if the materials have large differences in their coefficients of
thermal expansion, stresses may be induced even under free expansion (Figure 15 –1).
When a member undergoes a temperature change the member attempts to change
­dimensions. For an unconstrained member AB (Figure 15 –2) undergoing uniform change in
temperature T, the change in the length L is given by

d T 5 aTL (15.1.1)

where a is called the coefficient of thermal expansion and T is the change in temperature. The
coefficient a is a mechanical property of the material having units of 1 ⁄ 8 F (where 8F is degrees
Fahrenheit) in the USCS of units or 1 ⁄ 8 C (where 8C is degrees Celsius) in the SI system. In Eq.
(15.1.1), d T is considered to be positive when expansion occurs and negative when contraction
occurs. Typical values of a are: for structural steel a 5 6.5 3 1026 ⁄ 8 F (12 3 1026 ) ⁄ 8 C and
for aluminum alloys a 5 13 3 1026 ⁄ 8 F (23 3 1026 ) ⁄ 8 C .
Based on the definition of normal strain, we can determine the strain due to a uniform
temperature change. For the bar subjected to a uniform temperature change T (Figure 15 –2),

L
A B

(a)

L dT
A B

(b)

■■Figure 15 –2  (a) Unconstrained member and (b) same member subjected to uniform
temperature increase

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 15.1 Formulation of the Thermal Stress Problem and Examples 729

■■Figure 15 –3  Linear stress/strain law with initial thermal strain

the strain is the change in a dimension due to a temperature change divided by the original
dimension. Considering the axial direction, we then have
«T 5 aT (15.1.2)

Since the bar in Figure 15 –2 is free to expand, that is, it is not constrained by other members or
supports, the bar will not have any stress in it. In general, for statically determinate structures,
a uniform temperature change in one or more members does not result in stress in any of the
members. That is, the structure will be stress free. For statically indeterminate structures, a
uniform temperature change in one or more members of the structure usually results in stress
s T in one or more members. We can have strain due to temperature change «T without stress
due to temperature change, and we can have s T without any actual change in member lengths
or without strains.
We will now consider the one-dimensional thermal stress problem. The linear stress/strain
diagram with initial (thermal) strain («0 5 «T ) is shown in Figure 15 –3.
For the one-dimensional problem, we have, from Figure 15 –3,
sx
«x 5 1 «T (15.1.3)
E
If, in general, we let 1 / E 5 [ D]21, then in general matrix form Eq. (15.1.3) can be written as

{«} 5 [ D]21{s} 1 {«T } (15.1.4)

From Eq. (15.1.4), we solve for {s} as

{s} 5 D({«} 2 {«T }) (15.1.5)

The strain energy per unit volume (called strain energy density) is the area under the s 2«
diagram in Figure 15 –3 and is given by
1
u0 5 {s}({«}) 2 {«T }) (15.1.6)
2
Using Eq. (15.1.5) in Eq. (15.1.6), we have
1
u0 5 ({«} 2 {«T })T [ D]({«} 2 {«T }) (15.1.7)
2
where, in general, the transpose is needed on the strain matrix to multiply the matrices properly.

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 730 15 | Thermal Stress

The total strain energy is then

U5 ∫V u0 dV (15.1.8)
Substituting Eq. (15.1.7) into Eq. (15.1.8), we obtain

1
U5 ∫V 2 ({«} 2 {«T })T [ D]({«} 2 {«T }) dV (15.1.9)
Now, using {«} 5 [B]{d} in Eq. (15.1.9), we obtain

1
2 ∫V
U5 ([ B]{d} 2 {«T })T [ D]([ B]{d} 2 {«T }) dV (15.1.10)

Simplifying Eq. (15.1.10) yields

1
2 ∫V
U5 ({d}T [ B]T [ D][ B]{d} 2 {d}T [ B]T [ D]{«T }
(15.1.11)
2{«T }T [ D][ B]{d} 1 {«T }T [ D]{«T }) dV

The first term in Eq. (15.1.11) is the usual strain energy due to stress produced from mechanical
loading—that is,

1
2 ∫V
UL 5 {d}T [ B]T [ D][ B]{d} dV (15.1.12)

Terms 2 and 3 in Eq. (15.1.11) are identical and can be written together as

UT 5 ∫V {d}T [ B]T [ D]{«T } dV (15.1.13)

The last (fourth) term in Eq. (15.1.11) is a constant and drops out when we apply the principle
of minimum potential energy by setting
U
5 0 (15.1.14)
{d}

Therefore, letting U 5 U L 1 UT and substituting Eqs. (15.1.12) and (15.1.13) into

Eq. (15.1.14), we obtain two contributions as
U L
{d} ∫V
5 [ B]T [ D][ B] dV {d} (15.1.15)

UT
{d} ∫V
and 5 [ B]T [ D]{«T }dV 5 { fT } (15.1.16)

We recognize the integral term in Eq. (15.1.15) that multiplies by the displacement matrix {d}
as the general form of the element stiffness matrix [k], whereas Eq. (15.1.16) is the load or
force vector due to temperature change in the element.

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 15.1 Formulation of the Thermal Stress Problem and Examples 731

One-Dimensional Bar
We will now consider the one-dimensional thermal stress problem.We define the thermal
strain matrix for the one-dimensional bar made of isotropic material with coefficient of ther-
mal expansion a , and subjected to a uniform temperature rise T, as

{«T } 5 {« xT } 5 {aT} (15.1.17)

where the units on a are typically (in. / in.) / 8 F or (mm/mm)/ 8 C.

For the simple one-dimensional bar (with a node at each end), we substitute Eq. (15.1.17)
into Eq. (15.1.16) to obtain the thermal force matrix as
L
{ fT } 5 A ∫ [ B]T [ D]{aT} dx (15.1.18)
0

Recall that for the one-dimensional case, from Eqs. (3.10.15) and (3.10.13), we have

1 1
[ D] 5 [ E ] [ B] 5 2 (15.1.19)
 L L 

Substituting Eqs. (15.1.19) into Eq. (15.1.18) and simplifying, we obtain the thermal force
matrix as

 fT 1  2EaTA 
{ fT } 5  5  (15.1.20)
f  EaTA 
 T 2 

Two-Dimensional Plane Stress and Plane Strain

For the two-dimensional thermal stress problem, there will be two normal strains, « xT and
« yT , along with a shear strain g xyT due to the change in temperature because of the different
mechanical properties (such as E x ± E y ) in the x and y directions for the anisotropic material
(see Figure 15 –4). The thermal strain matrix for an anisotropic material is then

 « xT 
 
{«T } 5  « yT  (15.1.21)
g xyT 
 

p g dy + eyT dy
dy x − xyT
2
dx
dx + exT dx
(a) (b)

■■Figure 15 –4  Differential two-dimensional element (a) before and (b) after being subjected
to uniform temperature change for an anisotropic material

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 732 15 | Thermal Stress

For the case of plane stress in an isotropic material ( E x 5 E y ) with coefficient of thermal
expansion a subjected to a temperature rise T, the thermal strain matrix is

aT 
 
{«T } 5 aT  (15.1.22)
 0 
 

No shear strains are caused by a change in temperature of isotropic materials, only expansion
or contraction.
For the case of plane strain in an isotropic material, the thermal strain matrix is

aT 
 
{«T } 5 (1 1 ν ) aT  (15.1.23)
 0 
 

For a constant-thickness (t), constant-strain triangular element, Eq. (15.1.16) can be

­simplified to

{ fT } 5 [ B]T [ D]{«T } tA (15.1.24)

The forces in Eq. (15.1.24) are contributed to the nodes of an element in an unequal manner
and require precise evaluation. It can be shown that substituting Eq. (6.1.8) for [D], Eq. (6.2.34)
for [B], and Eq. (15.1.22) for {«T } for a plane stress condition into Eq. (15.1.24) reveals the
constant-strain triangular element thermal force matrix to be

 bi 
 
 fTix   gi 
 
 fTiy  a EtT  b j 
{ fT } 5  5   (15.1.25)
   2(1 2 v )  g j 
 fTmy   bm 
   
g m 

where the b  s and g  s are defined by Eqs. (6.2.10).

Axisymmetric Element
For the case of an axisymmetric triangular element of isotropic material subjected to uniform
temperature change, the thermal strain matrix is

 «rT  aT 
  aT 
 « zT 
{«T } 5 
«u T 
5   (15.1.26)
  aT 
g r zT   0 
 

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 15.1 Formulation of the Thermal Stress Problem and Examples 733

The thermal force matrix for the three-noded triangular element is obtained by substituting the
[B] from Eq. (9.1.19) and Eq. (9.1.21) into the following:

{ fT } 5 2π ∫[ B]T [ D]{«T }rdA (15.1.27)

A

For the element stiffness matrix evaluated at the centroid (r , z ), Eq. (15.1.27) becomes

{ fT } 5 2π rA[ B]T [ D]{«T } (15.1.28)

where [ B ] is given by Eq. (9.2.3), A is the surface area of the element which can be found in
general from Eq. (6.2.8) when the coordinates of the element are known, and [D] is given by
Eq. (9.2.6).
We will now describe the solution procedure for both one- and two-dimensional thermal
stress problems.

Step 1
Evaluate the thermal force matrix, such as Eq. (15.1.20) or Eq. (15.1.25). Then treat this
force matrix as an equivalent (or initial) force matrix {F0 } analogous to that obtained when we
replace a distributed load acting on an element by equivalent nodal forces (Chapters 4 and 5
and Appendix D).

Step 2
Apply {F} 5 [ K ]{d} 2 {F0 }, where if only thermal loading is considered, we solve
{F0 } 5 [K ]{d} for the nodal displacements. Recall that when we formulate the set of simulta-
neous equations, {F} represents the applied nodal forces, which here are assumed to be zero.

Step 3
Back-substitute the now known {d} into step 2 to obtain the actual nodal forces,
{F}(5 [ K ]{d} 2 {F0 }).
Hence, the thermal stress problem is solved in a manner similar to the distributed load
problem discussed for beams and frames in Chapters 4 and 5. We will now solve the following
examples to illustrate the general procedure.

Example 15.1
For the one-dimensional bar fixed at both ends and subjected to a uniform temperature rise
T 5 50 8 F as shown in Figure 15 –5, determine the reactions at the fixed ends and the axial stress
in the bar. Let E 5 30 3 106 psi , A 5 4 in 2, L 5 4 ft, and a 5 7.0 3 1026 (in. / in.) / 8 F .

■■Figure 15 –5  Bar subjected to a uniform temperature rise

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 734 15 | Thermal Stress

SOLUTION:
Two elements will be sufficient to represent the bar because internal nodal displacements
are not of importance here. To solve {F0 } 5 [ K ]{d}, we must determine the global stiffness
matrix for the bar. Hence, for each element, we have
1 2 2 3
AE  1 21 lb AE  1 21 lb (15.1.29)
[ k (1) ] 5  [ k (2) ] 5
L / 2 21 1 in. 
L / 2 21 1 in.

where the numbers above the columns in the [k] s indicate the nodal displacements associ-
ated with each element.

Step 1
Using Eq. (15.1.20), the thermal force matrix for each element is given by

2EaTA  2EaTA 
{ f (1)} 5   { f (2)} 5   (15.1.30)
 EaTA   EaTA 
where these forces are considered to be equivalent nodal forces.

Step 2
Applying the direct stiffness method to Eqs. (15.1.29) and (15.1.30), we assemble the global
equations {F0 } 5 [ K ]{d} as

2EaTA   1 21 0  u1 
  AE   
 0 5 21 1 1 1 21 u2  (15.1.31)
 EaTA  L / 2  0 21 1 u3 
    

Applying the boundary conditions u1 5 0 and u3 5 0 and solving the second equation of
Eq. (15.1.31), we obtain
u2 5 0 (15.1.32)

Step 3
Back-substituting Eq. (15.1.32) into the global equation {F} 5 [K ]{d} 2 {F0 } for the nodal
forces, we obtain the actual nodal forces as

 F1x  0  2EaTA   EaTA 

       
 F2 x  5  0  2  0  5  0  (15.1.33)
 F3 x  0   EaTA  2EaTA 
       

Using the numerical quantities for E, a , T, and A in Eq. (15.1.33), we obtain

F1x 5 42,000 lb F2 x 5 0 F3 x 5 242, 000 lb

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 15.1 Formulation of the Thermal Stress Problem and Examples 735

■■Figure 15 –6  Free-body diagram of the bar of Figure 15 –5

as shown in Figure 15 –6. The stress in the bar is then

42,000
s5 5 10,500 psi (compressive) (15.1.34)
4

Example 15.2
For the bar assemblage shown in Figure 15 –7, determine the reactions at the fixed ends
and the axial stress in each bar. Bar 1 is subjected to a temperature drop of 10 8 C. Let
bar 1 be aluminum with E 5 70 GPa , a 5 23 3 1026 (mm/mm)/ 8 C, A 5 12 3 1024 m 2,
and L 5 2 m . Let bars 2 and 3 be brass with E 5 100 GPa , a 5 20 3 1026 (mm/mm)/ 8 C,
A 5 6 3 1024 m 2 , and L 5 2 m .

■■Figure 15 –7  Bar assemblage for thermal stress analysis

SOLUTION:
We begin the solution by determining the stiffness matrices for each element.

Element 1
1 2
(12 3 1024 )(70 3 106 )  1 21  1 21 kN (15.1.35)
[ k (1) ] 5 21 5 42,000 
2  1 21 1 m

Elements 2 and 3

2 3
2 4
(6 3 1024 )(100 3 106 )  1 21  1 21 kN (15.1.36)
[ k (2) ] 5 [ k (3) ] 5 21  5 30, 000 21
2  1   1 m

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 736 15 | Thermal Stress

Step 1
We obtain the element thermal force matrices by evaluating Eq. (15.1.20). First, evaluating
2EaTA for element 1, we have
2EaTA 5 2(70 3 106 )(23 3 1026 )(210)(12 3 1024 ) 5 19.32 kN (15.1.37)
where the 210 term in Eq. (15.1.37) is due to the temperature drop in element 1. Using the
result of Eq. (15.1.37) in Eq. (15.1.20), we obtain

 f1x   19.32 
{ f (1)} 5   5   kN (15.1.38)
f
 2 x  219.32 
There is no temperature change in elements 2 and 3, and so

0 
{ f (2)} 5 { f (3)} 5   (15.1.39)
0 
Step 2
Assembling the global equations using Eqs. (15.1.35), (15.1.36), (15.1.38), and (15.1.39)
into {F0 } 5 [ K ]{d}, we obtain
1 2 3 4
 42 242 0 0  u1  119.32 
   219.32  (15.1.40)
242 42 1 30 1 30 230 230  u2 
1000    5  
0 230 30 0  u3
    0 
 0 230 0 30  u4   0 

where the right-side thermal forces are considered to be equivalent nodal forces. Using the
boundary conditions
u1 5 0 u3 5 0 u4 5 0 (15.1.41)

we obtain, from the second equation of Eq. (15.1.40),

1000(102)u2 5 219.32
Solving for u2, we obtain
u2 5 21.89 3 1024 m (15.1.42)

Step 3
Back-substituting Eq. (15.1.42) into the global equation for the nodal forces,
{F} 5 [ K ]{d} 2 {F0 }, we have
 F1x   42 242 0 0  0   19.32 
  242 102 230 230     
 F2 x  4  
 21.89 3 10  2 219.32  (15.1.43)
2
  5 1000 
 F3 x   0 230 30 0  0   0 
 F4 x   0 230 0 30   0   0 

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 15.1 Formulation of the Thermal Stress Problem and Examples 737

■■Figure 15 –8  Free-body diagram of the bar assemblage of Figure 15 –7

Simplifying Eq. (15.1.43), we obtain the actual nodal forces as

F1x 5 211.38 kN
F2 x 5 0.0 kN
(15.1.44)
F3 x 5 5.69 kN
F4 x 5 5.69 kN
A free-body diagram of the bar assemblage is shown in Figure 15 –8. The stresses in each
bar are then tensile and given by
11.38
s (1) 5 5 9.48 3 103 kN / m 2 (9.48 MPa)
12 3 1024
(15.1.45)
5.69
s (2) (3)
5s 5 3 2
5 9.48 3 10 kN / m (9.48 MPa)
6 3 1024

Example 15.3
For the plane truss shown in Figure 15 –9, determine the displacements at node 1 and the axial
stresses in each bar. Bar 1 is subjected to a temperature rise of 75 8 F. Let E 5 30 3 106 psi,
a 5 7 3 1026 (in./in.)/ 8 F, and A 5 2 in 2 for both bar elements.

x′ x′

■■Figure 15 –9  Plane truss for thermal stress analysis

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 738 15 | Thermal Stress

SOLUTION:
First, using Eq. (3.4.23), we determine the stiffness matrices for each element.

Element 1
Choosing x9 from node 2 to node 1, u 5 908 , and so cos u 5 0, sin u 5 1, and

2 1
 0 0 0 0 
 
(2)(30 3 10 6)
 1 0 21  lb (15.1.46)
[ k (1) ] 5   in.
(8 3 12) 0 0
 
 Symmetry 1 

Element 2
Choosing x9 from node 3 to node 1, u 5 1808 2 53.138 5 126.878, and so cos u 5 20.6,
sin u 5 0.8 and

3 1
 0.36 20.48 20.36 0.48 
 
(2)(30 3 106 )  0.64 0.48 20.64  lb (15.1.47)
[ k (2) ] 5 
(10 3 12) 0.36 20.48  in.
 
 Symmetry 0.64 


Step 1
We obtain the element thermal force matrices by evaluating Eq. (15.1.20) as follows:

2EaTA 5 2(30 3 106 )(7 3 1026 )(75)(2) 5 231,500 lb (15.1.48)

Using the result of Eq. (15.1.48) for element 1, we then have the local thermal force matrix as

 f29 x  231,500 
{ f 9(1)} 5   5   lb (15.1.49)
 f19x   31,500 

There is no temperature change in element 2, so

 f39 x 
0 
{ f 9(2)} 5   5   (15.1.50)
 f19x  0 

Recall that by Eq. (3.4.16), { f 9} 5 [T ]{ f }. Since we have shown that [T ]21 5 [T ]T , we can
obtain the global forces by premultiplying Eq. (3.4.16) by [T ]T to obtain the element nodal
forces in the global reference frame as

{ f } 5 [T ]T { f 9} (15.1.51)

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 15.1 Formulation of the Thermal Stress Problem and Examples 739

Using Eq. (15.1.51), the element 1 global nodal forces are then

 f9 
 f2 x 
0  
2x
 C 2S 0 
  S  
 f2 y  C 0 0   f29 y 
  5     (15.1.52)
 f1x  0 0 C 2S   f19x 
 f1 y   0 0 S C   
   f19y 

where the order of terms in Eq. (15.1.52) is due to the choice of the x9 axis from node 2 to
node 1 and where [T], given by Eq. (3.4.15), has been used.
Substituting the numerical quantities C 5 0 and S 5 1 (consistent with x9 for element 1),
and f19x 5 31,500, f19y 5 0 , f29 x 5 231,500 , and f29 y 5 0 into Eq. (15.1.52), we obtain

f2 x 5 0 f2 y 5 231,500 lb f1x 5 0 f1 y 5 31,500 lb (15.1.53)

These element forces are now the only equivalent global nodal forces, because element 2 is
not subjected to a change in temperature.

Step 2
Assembling the global equations using Eqs. (15.1.46), (15.1.47), and (15.1.53), into
{F0 } 5 [ K ]{d}, we obtain

 0.36 20.48 0 0 0  0
  u1   0 
 1.89 0 21.25 0 0   v1   31,500 
    
0.50 3 106 
0 0 0 0  u2  5  0 
      (15.1.54)
1.25 0 0  v 2 2 31,500 
 
 0.36 20.48   u3   0 
   v3   0 
 
 Symmetry 0.64 

The boundary conditions are given by
u1 5 0 u2 5 0 v2 5 0 u3 5 0 v3 5 0 (15.1.55)

Using the boundary condition Eqs. (15.1.55) and the second equation of Eq. (15.1.54),
we obtain
(0.945 3 106 ) v1 5 31,500

or v1 5 0.0333 in. (15.1.56)

Step 3
We now illustrate the procedure used to obtain the local element forces in local coordinates;
that is, the local element forces are
{ f 9} 5 [ k9]{d9} 2 { f09 } (15.1.57)

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 740 15 | Thermal Stress

We determine the actual local element nodal forces by using the relationship {d9} 5 [T p ]{d}
in Eq. (15.1.57) the usual bar element [ k9] matrix [Eq. (3.1.14)], the transformation matrix
[T p ] [Eq. (3.4.8)], and the calculated displacements and initial thermal forces applicable for
the element under consideration. Substituting the numerical quantities for element 1, from
Eq. (15.1.57) into [ f 9] 5 [ k9][T p ]{d} 2 { f09 }, we have

u2 5 0 
 f9   
 2x  2 (30 3 10 )  1 21  0 1 0 0   v2 5 0
6
 231,500 
 5      2  (15.1.58)
 f19x  8 3 12 21 1  0 0 0 1   u1 5 0   31,500 
 v1 5 0.0333 
 

Simplifying Eq. (15.1.58), we obtain

f29 x 5 10, 700 lb f19x 5 210, 700 lb (15.1.59)

Dividing the local element force f19x (which is the far-end force consistent with the conven-
tion used in Section 3.5) by the cross-sectional area, we obtain the stress as

210, 700
s (1) 5 5 25350 psi (C ) (15.1.60)
2

Similarly, for element 2, we have

0 
 f39 x  2(30 3 106 )  1 21 20.6 0.8 0 0   0 
 5      (15.1.61)
10 3 12 21 1  0 0 20.6 0.8   0 
 f19x 
 0.0333 

Simplifying, Eq. (15.1.61), we obtain

f39 x 5 213,310 lb f19x 5 13,310 lb (15.1.62)

where no initial thermal forces were present for element 2 because the element was not
subjected to a temperature change. Dividing the far-end force f19x by the cross-sectional
area results in

s (2) 5 6660 psi (T ) (15.1.63)

For two- and three-dimensional stress problems, this direct division of force by
cross-sectional area is not permissible. Hence, the total stress due to both applied loading
and temperature change must be determined by

{s} 5 {s L } 2 {s T } (15.1.64)

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 15.1 Formulation of the Thermal Stress Problem and Examples 741

We now illustrate Eq. (15.1.64) for bar element 1 of the truss of Example 15.3. For the
bar, s L can be obtained using Eq. (3.5.6) {s L } 5 [C9]{d}, and s T is obtained from
{s T } 5 [ D]{«T } 5 EaT (15.1.65)

because [ D] 5 E and {«T } 5 aT for the bar element. The stress in bar element 1 is then
determined to be

u2 
 
E  v2 
s (1) 5 [2C 2S C S ]   2 EaT (15.1.66)
L  u1 
 v1 

Substituting the numerical quantities for element 1 into Eq. (15.1.66), we obtain

0 
30 3 106 
0 
s (1) 5 [0 21 0 1]   2 (30 3 106 )(7 3 1026 )(75) (15.1.67)
8 3 12 0 
 0.0333 

or s (1) 5 25350 psi (C ) (15.1.68)

We will now illustrate the solutions of two plane thermal stress problems.

Example 15.4
For the plane stress element shown in Figure 15 –10, determine the element equations. The
element has a 2000 lb/in 2 pressure acting perpendicular to side j–m and is subjected to a
30 8 F temperature rise.

SOLUTION:
Recall that the stiffness matrix is given by [Eq. (6.2.52) or (6.4.1)]

[ k ] 5 [ B]T [ D][ B] tA (15.1.69)

bi 5 y j 2 ym 5 23 g i 5 x m 2 x j 5 21
and
b j 5 ym 2 yi 5 3 g j 5 xi 2 x m 5 21
bm 5 yi 2 y j 5 0 g m 5 x j 2 xi 5 2

(3)(2)
and A5 5 3 in 2 (15.1.70)
2

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 742 15 | Thermal Stress

■■Figure 15 –10  Plane stress element subjected to mechanical loading and a

temperature change

Therefore, substituting the results of Eqs. (15.1.70) into Eq. (6.2.34) for [ B], we obtain

23 0 3 0 0 0
1
[ B] 5 0 21 0 21 0 2  (15.1.71)
6 
 21 23 21 3 2 0 

Assuming plane stress conditions to be valid, we have

1 ν 0  6
1 0.25 0 
[ D] 5
E ν 1 0  5 30 3 10  0.25 1 0 
1 2 ν2   1 2 (0.25)2  
0 0 1 2 ν   0 0 0.375 
 2 


8 2 0 
5 (4 3 106 )  2 8 0  psi (15.1.72)
 
 0 0 3 

23 0 21
 0 21 23 
  8 2 0 
1 3
(4 3 10 )  2 8 0  (15.1.73)
T 0 21 6
Also, [ B] [ D ] 5
6  0 21 3
   0 0 3 
 0 0 2
 0 2 0 

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 15.1 Formulation of the Thermal Stress Problem and Examples 743

Simplifying Eq. (15.1.73), we obtain

224 26 23 
 22 28 29 
6
 
T 4 3 10  24 6 23 
[ B] [ D ] 5 (15.1.74)
6  22 28 9
 
 0 0 6
 4 16 0 

Therefore, substituting the results of Eqs. (15.1.71) and (15.1.74) into Eq. (15.1.69) yields
the element stiffness matrix as

224 26 23 
 22 28 29 
  23 0 3 0 0 0
(3 in 2 ) 4 3 106  24 6 23  
[ k ] 5 (1 in.) 0 21 0 21 0 2  (15.1.75)
6 6  22 28 9 
   21 23 21 3 2 0 
 0 0 6 
 4 16 0 

Simplifying Eq. (15.1.75), we have the element stiffness matrix as

 75 15 269 23 26 212 
 15 35 3 219 218 216 
6

1 3 10 269 3 75 215 26 12  lb
[k ] 5 (15.1.76)
3  23 219 215 35 18 216  in.
 
 26 218 26 18 12 0
212 216 12 216 0 32 

Using Eq. (15.1.25), the thermal force matrix is given by

 bi 
  23  23 
 gi  21 21
   
a EtT  b j  (7 3 1026 )(30 3 106 )(1)(30)  3   3
{ fT } 5  5   5 4200  
2(1 2 ν )  g j  2(1 2 0.25) 21 21
 bm   0  0
   2  2
g    
 m 

212,600 
 
 24200 
 12,600 
or { fT } 5   lb (15.1.77)
 24200 
 0
 8400 
 

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 744 15 | Thermal Stress

The force matrix due to the pressure applied alongside j–m is determined as follows:

L j -m 5 [(2 2 1)2 1 (3 2 0)2 ]1/2 5 3.163 in.

3 
px 5 p cos u 5 2000  5 1896 lb/in 2 (15.1.78)
 3.163
 1 
py 5 p sin u 5 2000  5 632 lb/in 2
 3.163

where θ is the angle measured from the x axis to the normal to surface j–m. Using Eq. (6.3.7)
to evaluate the surface forces, we have

 px 
{ fL } 5 ∫∫ [ N S ]T  py  dS
S j -m  
 Ni 0 
  0 0
 0 Ni  0 0
  (15.1.79)
 Nj 0   px  tL j -m 0   px 
5 ∫∫     dS 5 1
0  
S j -m 
0 N j   py  2 1   py 
  1   
 Nm 0  0
   
 0 1 
 0 N m  evaluated
alongside j -m

Evaluating Eq. (15.1.79), we obtain

0 0  0
0 0  0
  
(1 in.)(3.163 in.) 1 0  1896  3000 
{ fL } 5   5   lb (15.1.80)
2 0 1   632  1000 
  
1 0 3000 
 0 1  1000 
 

Using Eqs. (15.1.76), (15.1.77), and (15.1.80), we find that the complete set of element
equations is
 75 15 269 23 26 212 
   ui  212,600 
 35 3 219 218 216   vi   
    24200 
1 3 106  75 215 26 12   u j   15,600 
  5  (15.1.81)
3 35 18 216   v j   23200 
 
 12 0  um   3000 
   vm  
 9400 
 Symmetry 32   


where the force matrix is { fT } 1 {f L }, obtained by adding Eqs. (15.1.77) and (15.1.80).

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 15.1 Formulation of the Thermal Stress Problem and Examples 745

Example 15.5
For the plane stress plate fixed along one edge and subjected to a uniform temperature rise
of 50 8 C as shown in Figure 15 –11, determine the nodal displacements and the stresses in
each element. Let E 5 210 GPa, ν 5 0.30 , t 5 5 mm, and a 5 12 3 1026 (mm/mm) / 8 C .

■■Figure 15 –11  Discretized plate subjected to a temperature change

SOLUTION:
The discretized plate is shown in Figure 15 –11. We begin by evaluating the stiffness matrix
of each element using Eq. (6.2.52).

Element 1
Element 1 has coordinates x1 5 0, y1 5 0, x 2 5 0.5, y2 5 0, x5 5 0.25, and y5 5 0.25.
From Eqs. (6.2.10), we obtain

b 1 5 y2 2 y5 5 20.25 m b 2 5 y5 2 y1 5 0.25 m b 5 5 y1 2 y2 5 0
(15.1.82)
g1 5 x5 2 x 2 5 20.25 m g 2 5 x1 2 x5 5 20.25 m g 5 5 x 2 2 x1 5 0.5 m

Using Eqs. (6.2.32) in Eq. (6.2.34), we have

 b1 0 b2 0 b5 0 
1  
[ B] 5 0 g1 0 g2 0 g5 
2A 
g
 1
b1 g 2 b2 g5 b 5 

(15.1.83)
20.25 0 0.25 0 0 0 
1   1
5  0 20.25 0 20.25 0 0.5 
0.125  m
20.25 20.25 20.25 0.25 0.5 0 


For plane stress, [D] is given by

1 v 0  1 0.3 0 
E v 1 210 3 109 N
[ D] 5 0  5  0.3 1 0  2 (15.1.84)
(1 2 v ) 
2  0.91  m
0 0 1 2 v   0 0 0.35 
 2 


Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 746 15 | Thermal Stress

We obtain the element stiffness matrix using

[ k ] 5 t A[ B]T [ D][ B] (15.1.85)
Substituting the results of Eqs. (15.1.83) and (15.1.84) into Eq. (15.1.85) and carrying out
the multiplications, we have
u1 v1 u2 v2 u5 v5
 8.4375 4.0625 24.0625 20.3125 24.375 23.75 
 
 4.0625 8.4375 0.3125 4.0625 24.375 212.5 
 24.0625 0.3125 8.4375 24.0625 24.375 3.75  N
[ k ] 5 4.615 3 10 7  
 20.3125 4.0625 24.0625 8.4375 4.375 212.5  m
 
24.375 24.375 24.375 4.375 8.75 0 
23.75 212.5 3.75 212.5 0 25 
 
(15.1.86)
Element 2
For element 2, the coordinates are x 2 5 0.5, y2 5 0, x3 5 0.5, y3 5 0.5, x5 5 0.25, and
y5 5 0.25. Proceeding as for element 1, we obtain

b 2 5 0.25 m b 3 5 0.25 m b 5 520.5 m

γ 2 520.25 m γ 3 5 0.25 m γ5 50

The element stiffness matrix then becomes

u2 v2 u3 v3 u5 v5
 8.4375 24.0625 4.0625 20.3125 212.5 4.375 
 
 24.0625 8.4375 0.3125 24.0625 3.75 24.375 
 4.0625 0.3125 8.437 4.0625 212.5 24.375  N
[ k ] 5 4.615 3 10 7  
 20.3125 24.0625 4.0625 8.4375 23.75 24.375  m
212.5 3.75 212.5 23.75 25 0 
 
 4.375 24.375 24.375 24.375 0 8.75 

(15.1.87)
Element 3
For element 3, using the same steps as for element 1, we obtain the stiffness matrix as
u3 v3 u4 v4 u5 v5
 8.437 4.0625 24.0625 20.3125 24.375 23.75 
 
 4.0625 8.437 0.3125 4.0625 24.375 212.5 
24.0625 0.3125 8.437 24.0625 24.375 3.75  N
[ k ] 5 4.615 3 10 7  
20.3125 4.0625 24.0625 8.4375 4.375 212.5  m
 
24.375 24.375 24.375 4.375 8.75 0 
23.75 212.5 3.75 212.5 0 25 
 
(15.1.88)
Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 15.1 Formulation of the Thermal Stress Problem and Examples 747

Element 4
Finally, for element 4, we obtain
u4 v4 u1 v1 u5 v5
 8.437 24.0625 4.0625 20.3125 212.5 4.375 
 
 24.0625 8.4375 0.3125 24.0625 3.75 24.375 
 4.0625 0.3125 8.4375 4.0625 212.5 24.375  N
[ k ] 5 4.615 3 10 7  
 20.3125 24.0625 4.0625 8.4375 23.75 24.375  m
212.5 3.75 212.5 23.75 25 0 
 
 4.375 24.375 24.375 24.375 0 8.75 

(15.1.89)
Using the direct stiffness method, we assemble the element stiffness matrices, Eqs. (15.1.86)
through (15.1.89), to obtain the global stiffness matrix as

u1 v1 u2 v2
 16.874 8.125 24.0625 20.3125

 8.125 16.874 0.3125 4.0625
 24.0625 0.3125 16.874 28.125

 20.3125 4.0625 28.125 16.875

0 0 4.0625 0.3125
[ K ] 5 4.615 3 10 7 
0 0 20.3125 24.0625

 4.0625 20.3125 0 0
 0.3125 24.0625 0 0

216.875 28.125 216.875 8.125

 28.125 216.875 8.125 216.875

u3 v3 u4 v4 u5 v5

0 0 4.0625 0.3125 216.875 28.125 


0 0 20.3125 24.0625 28.125 216.875 
4.0625 20.3125 0 0 216.875 8.125 

0.3125 24.0625 0 0 8.125 216.875 

16.875 8.125 24.0625 20.3125 216.875 28.125  N
8.125 16.875 0.3125 4.0625 28.125 216.875  m

24.0625 0.3125 16.875 28.125 216.875 8.125 
20.3125 4.0625 28.125 16.875 8.125 216.875 

216.875 28.125 216.875 8.125 67.5 0 

28.125 216.875 8.125 216.875 0 67.5 

(15.1.90)
Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 748 15 | Thermal Stress

Next, we determine the thermal force matrices for each element by using Eq. (15.1.25)
as follows:
Element 1
 b1  20.25 
  20.25 
 g1   
a EtT  b 2  (12 3 1026 )(210 3 109 )(0.005 m)(50)  0.25 
{ fT } 5  5 20.25 
2(1 2 ν )  g 2  2(1 2 0.3)  
b5   0 
   0.5 
 g 5   

20.25   fT 1x  2112,500 
20.25     
 f  2112,500 
T 1 y
 
 0.25   fT 2 x   112,500 
5 450, 000   5  5 N (15.1.91)
20.25
   fT 2 y  2112,500 
 0   fT 5 x   0
 0.5     225, 000 
   fT 5 y   
Element 2

 0.25   fT 2 x   112,500 
20.25     
 f  2112,500 
T 2 y
 
 0.25   fT 3 x   112,500 
{ fT } 5 450, 000  0.25  5  5  N (15.1.92)
  f
 T 3y   112,500 
20.5   fT 5 x  2225, 000 
 0    
   fT 5 y   0 

Element 3
 0.25   fT 3 x   112,500 
 0.25     
 f   112,500 
T 3 y
 
20.25   fT 4 x  2112,500 
{ fT } 5 450, 000  0.25  5  5  N (15.1.93)
  f
 T4y   112,500 
 0   fT 5 x   0
20.5    2225, 000 
   fT 5 y   
Element 4
20.25   fT 4 x  2112,500 
 0.25     
f
 T4y 
   112,500 
20.25   fT 1x  2112,500 
{ fT } 5 450, 000 20.25  5  5  N (15.1.94)
   fT 1 y  2112,500 
 0.5   fT 5 x   225, 000 
 0    
   fT 5 y   0 
Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 15.1 Formulation of the Thermal Stress Problem and Examples 749

We then obtain the global thermal force matrix by direct assemblage of the element force
matrices [Eqs. (15.1.91) through (15.1.94)]. The resulting matrix is

 fT 1x 
  2225, 000 
 fT 1 y   
 fT 2 x  2225, 000 
   225, 000 
 fT 2 y  2225, 000 
f   
 T 3x   225, 000 
  5   N (15.1.95)
 fT 3 y   225, 000 
 fT 4 x  2225, 000 
   
 fT 4 y   225, 000 
 fT 5 x   0
   0 
 fT 5 y  

Using Eqs. (15.1.90) and (15.1.95) and imposing the boundary conditions
u1 5 v1 5 u4 5 v4 5 0 , we obtain the system of equations for solution as

 fT 2 x 5 225, 000 
 
 fT 2 y 5 2225, 000 
 fT 3 x 5 225, 000 
f 5 4.615 3 10 7
 T 3y 5 225, 000 
 fT 5 x 50 
 
 fT 5 y 50 

 16.874 − 8.125 4.0625 − 0.3125 −16.875 8.125   u 
  2
 − 8.125 16.875 0.3125 − 4.0625 8.125 −16.875   v2 
 
 4.0625 0.3125 16.875 8.125 −16.875 − 8.125   u3 
  
 − 0.3125 − 4.0625 8.125 16.875 − 8.125 −16.875   v3 
  u 5 
 −16.875 8.125 −16.875 − 8.125 67.5 0  
 8.125 −16.875 − 8.125 −16.875 0 67.5   v5 (15.1.96)
 

Solving Eq. (15.1.96) for the nodal displacements, we have

u2   3.327 3 1024 

   
 v2 
24
21.911 3 10 
 u3   3.327 3 1024 
 5 24 
m (15.1.97)
 v3   1.911 3 10 
 u5   2.123 3 1024 
v   29 
 5  6.654 3 10 
Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 750 15 | Thermal Stress

We now use Eq. (15.1.64) to obtain the stresses in each element. Using Eqs. (6.2.36) and
(15.1.65), we write Eq. (15.1.64) as

{s} 5 [ D][ B]{d} 2 [ D]{«T } (15.1.98)

Element 1

 u1 
 
v1
s  1 ν 0 
b1 0 b2 0 b5 0  
 x  E ν 1 1    2 
u
s y  5 0  0 g1 0 0
 2A 
g2 g5   
1 2 ν2  v2
τ  0 0 1 2 ν   g b1 g 2 b2 g5 b5   
 
xy
  1   u 5  (15.1.99)
 2   
 v5 
1 ν 0  aT 
E 
ν 1 0  aT 
 
2
1 2 ν2 
 0 0 1 2 ν   0 
 2 


Using Eqs. (15.1.82) and (15.1.97) along with the mechanical properties E, n, and a in
Eq. (15.1.99), we obtain

s  1
 x  0.3 0 
210 3 109  0.3 1
s y  5 0 
0.91  
τ   0 0 0.35 
 xy 
 0 
 
20.25   0 
0 0.25 0 0 0
1    3.327 3 1024 
3  0 20.25 0 20.25 0 0.5   
0.125   21.911 3 1024 
20.25 20.25 20.25 0.25 0.5 0   2.123 3 1024 
 29

 6.654 3 10 
 26 
1 0.3 0   (12 3 10 )(50) 
9
210 3 10    6  (15.1.100)
0.3 1 0 (12 3 10 )(50) 
2

2
0.91 
 0 0 0.35   0 
 

Simplifying Eq. (15.1.100) yields

s   1.800 3 108  1.8 3 108   0 

 x       
8 8 7
 s y  5  1.342 3 10  2 1.8 3 10  5 24.57 3 10  Pa (15.1.101)
τ  21.600 3 10 7   0  21.60 3 10 7 
 xy       

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 15.1 Formulation of the Thermal Stress Problem and Examples 751

■■Figure 15 –12  Discretized plate showing displaced plate superimposed with maximum
principal stress plot in Pa (See the full-color insert for a color version of this figure.)

Similarly, we obtain the stresses in element 2 as follows:

Element 2

s  1.640 3 108  1.8 3 108   21.6 3 10 7 

 x       
 s y  5  2.097 3 10  2 1.8 3 10  5  2.973 3 10 7  Pa (15.1.102)
8 8

τ   22150   0   22150 
 xy       

Stresses in elements 3 and 4 can be determined similarly. The clamped plate subjected
to uniform heating (see the longhand solution, Example 15.5) was also solved using the
Autodesk computer program from Reference [1]. The plate was discretized using the
“automesh” feature of [1]. These results are similar to those obtained from the longhand
solution of Example 15.5 using the very coarse mesh. The computer program solution with
342 elements is naturally more accurate than the longhand solution with only four elements.
Figure 15 –12 shows the discretized plate with resulting displacement superimposed on the
maximum principal stress plot.

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 752 15 | Thermal Stress

■■Figure 15 –13  von Mises stress plot for a solid part subjected to 1008C temperature rise
inside the surface of the hole (See the full-color insert for a color version of this figure.)

Finally, Figure 15 –13 shows a three-dimensional solid part that is fixed on the small front
surfaces and uniformly heated by a temperature increase of 1008 C acting over the entire inside
surface of the hole. The resulting von Mises stress plot is shown with the maximum value of
329.9 MPa occurring inside the hole.

Summary Equations
Unconstrained displacement of bar due to uniform temperature change:

d T 5 aTL (15.1.1)

Strain due to uniform temperature change for a bar:

«T 5 aT (15.1.2)

Thermal strain matrix for a bar:

{«T } 5 {« xT } 5 {aT} (15.1.17)

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 Summary Equations 753

Thermal force matrix for a bar:

 fT 1  2EaTA 
{ fT } 5  5  (15.1.20)
f
 T 2   EaTA 

Thermal strain matrix for isotropic material in plane stress:

aT 
 
{«T } 5 aT  (15.1.22)
 0 
 

Thermal strain matrix for isotropic material in plane strain:

aT 
 
{«T } 5 (1 1 ν ) aT  (15.1.23)
 0 
 

Thermal force matrix for plane stress triangle:

 bi 
 
 f Tix   gi 
 
 f Tiy  a EtT  b j 
{fT} 5  5   (15.1.25)
   2(1 2 ν )  g j 
 f Tmy  b m 
   
 g m 

Thermal strain matrix for axisymmetric triangular element:

 «rT 
  aT 
 «z T  aT 
{«T } 5   5   (15.1.26)
 « uT  aT 
g rzT   0 
 

Thermal force matrix for axisymmetric element evaluated at its centroid:

{ fT } 5 2p rA[ B]T [ D]{«T } (15.1.28)

Reference
[1] Autodesk, Inc., McInnis Parkway, San Rafael, CA 94903.

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 754 15 | Thermal Stress

Problems
15.1 For the one-dimensional steel bar fixed at the left end, free at the right end, and
subjected to a uniform temperature rise T 5 100 8 F as shown in Figure P15 –1, deter-
mine the free-end displacement, the displacement 60 in. from the fixed end, the
reactions at the fixed end, and the axial stress. Let E 5 30 3 106 psi , A 5 4 in 2,
and a 5 7.0 3 1026 (in./in.)/ 8 F .

■■Figure P15 –1 ■■Figure P15 –2

15.2 For the one-dimensional steel bar fixed at each end and subjected to a uniform
temperature drop of T 5 30 8 C as shown in Figure P15 –2, determine the reactions
at the fixed ends and the stress in the bar. Let E 5 200 GPa , A 5 1 3 1022 m 2, and
a 5 11.7 3 1026 (mm/mm)/ 8 C .
15.3 For the plane truss shown in Figure P15 –3, bar element 2 is subjected to a uni-
form temperature rise of T 5 60 8 F. Let E 5 30 3 106 psi , A 5 2 in 2 , and
a 5 7.0 3 1026 (in./in.)/ 8 F . The lengths of the truss elements are shown in the
figure. Determine the stresses in each bar. [Hint: See Eqs. (3.6.4) and (3.6.6) in
Example 3.5 for the global and reduced [K] matrices.]

■■Figure P15 –3 ■■Figure P15 –4

15.4 For the plane truss shown in Figure P15 –4, bar element 1 is subjected to a
uniform temperature drop of 40 8 F. Let E 5 30 3 106 psi , A 5 2 in 2 , and
a 5 7.0 3 1026 (in./in.)/ 8 F . The lengths of the truss elements are shown in the
­figure. Determine the stresses in each bar. (Hint: Use Problem 3.21 for [K].)
15.5 For the structure shown in Figure P15 –5, bar element 1 is subjected to a uni-
form temperature rise of T 5 40 8 C. Let E 5 200 GPa , A 5 2 3 1022 m 2 , and
a 5 12 3 1026 (mm/mm)/ 8 C. Determine the stresses in each bar.

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 Problems 755

■■Figure P15 –5 ■■Figure P15 –6

15.6 For the plane truss shown in Figure P15 –6, bar element 2 is subjected to a uni-
form temperature drop of T 5 30 8 C. Let E 5 70 GPa , A 5 4 3 1022 m 2, and
a 5 23 3 1026 (mm/mm)/ 8 C. Determine the stresses in each bar and the displace-
ment of node 1.
15.7 For the bar structure shown in Figure P15 –7, element 1 is subjected to a uni-
form temperature rise of T 5 30 8 C. Let E 5 210 GPa , A 5 3 3 1022 m 2, and
a 5 12 3 1026 (mm/mm)/ 8 C. Determine the displacement of node 1 and the
stresses in each bar.

■■Figure P15 –7 ■■Figure P15 –8

15.8 A bar assemblage consists of two outer steel bars and an inner brass bar. The three-
bar assemblage is then heated to raise the temperature by an amount T 5 80 8 F.
Let all cross-sectional areas be A 5 2 in 2 and L 5 60 in. , Esteel 5 30 3 106 psi ,
Ebrass 5 15 3 106 psi , asteel 5 6.5 3 1026 / 8 F , and a brass 5 10 3 1026 / 8 F. Deter-
mine (a) the displacement of node 2 and (b) the stress in the steel and brass bars.
See Figure P15 –8.

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 756 15 | Thermal Stress

15.9 For the plane truss shown in Figure P15 –9, bar element 2 is subjected to a uni-
form temperature rise of T 5 10 8 C. Let E 5 210 GPa , A 5 12.5 cm 2 , and
a 5 12 3 1026 / 8 C. What temperature change is needed in bars 1 and 3 to remove
the stress due to the uniform temperature rise in bar 2? Show enough work to prove
your answer. Use a longhand solution.

■■Figure P15 –9

15.10 When do stresses occur in a body made of a single material due to uniform tempera-
ture change in the body? Consider Problem 15.1 and also compare the solution to
Example 15.1 in this chapter.
15.11 Consider two thermally incompatible materials, such as steel and aluminum, attached
together as shown in Figure P15 –11. Will there be temperature-induced stress in
each material upon uniform heating of both materials to the same temperature when
the boundary conditions are simple supports (a pin and a roller such that we have a
statically determinate system)? Explain. Let there be a uniform temperature rise of
T 5 50 8 F.

Steel, E = 30 × 106 psi, a = 6.5 × 10−6/ °F

A
Aluminum, E = 10 × 106 psi, a = 13 × 10−6/ °F
L

■■Figure P15 –11

15.12 A bimetallic thermal control is made of cold-rolled yellow brass and magnesium
alloy bars (Figure P15 –12). The bars are arranged with a gap of 0.005 in. between
them at 72 8 F. The brass bar has a length of 1.0 in. and a cross-sectional area of
0.10 in 2, and the magnesium bar has a length of 1.5 in. and a cross-sectional area
of 0.15 in 2 . Determine (a) the axial displacement of the end of the brass bar and (b)
the stress in each bar after it has closed up due to a temperature increase of 100 8 F.
Use at least one element for each bar in your finite element model.
15.13 For the plane stress element shown in Figure P15 –13 subjected to a uniform tem-
perature drop of T 5 50 8 F, determine the thermal force matrix { fT }.

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 Problems 757

Brass Magnesium

d = 0.005 in.
1.0 in. 1.5 in.

■■Figure P15 –12 Figure P15 –13

Let E 5 10 3 106 psi ν 5 0.30 , and a 5 12.5 3 1026 (in./in.)/ 8 F . The coordinates
(in inches) are shown in the figure. The element thickness is t 5 1 in.
15.14 For the plane stress element shown in Figure P15 –14 subjected to a uniform tempera-
ture rise of T 5 50 8 C, determine the thermal force matrix { fT }. Let E 5 70 GPa ,
ν 5 0.3,a 5 23 3 1026 (mm/mm)/ 8 C, and t 5 5 mm . The coordinates (in millimeters)
are shown in the figure.

■■Figure P15 –14 ■■Figure P15 –15

15.15 For the plane stress element shown in Figure P15 –15 subjected to a uniform tempera-
ture rise of T 5 50 8 F, determine the thermal force matrix { fT }. Let E 5 30 3 106 psi ,
ν 5 0.3, a 5 7.0 3 1026 (in./in.)/ 8 F , and t 5 1 in . The coordinates (in inches) are
shown in the figure.
15.16 For the plane stress element shown in Figure P15 –16 subjected to a uniform tempera-
ture drop of T 5 20 8 C, determine the thermal force matrix { fT }. Let E 5 210 GPa ,
ν 5 0.25, and a 5 12 3 1026 (mm/mm)/ 8 C. The coordinates (in millimeters) are
shown in the figure. The element thickness is 10 mm.
15.17 For the plane stress plate fixed along the left and right sides and subjected to a uni-
form temperature rise of 50 8 F as shown in Figure P15 –17, determine the stresses
in each element. Let E 5 10 3 106 psi , ν 5 0.30 , a 5 12.5 3 1026 (in./in.)/ 8 F , and
t 5 14 in . The coordinates (in inches) are shown in the figure. (Hint: The nodal
displacements are all equal to zero. Therefore, the stresses can be determined from
{s} 5 2[ D]{«T }.)

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 758 15 | Thermal Stress

■■Figure P15 –16 ■■Figure P15 –17

15.18 For the plane stress plate fixed along all edges and subjected to a uniform tempera-
ture decrease of 20 8 C as shown in Figure P15 –18, determine the stresses in each
element. Let E 5 210 GPa , ν 5 0.25, and a 5 12 3 1026 (mm/mm)/ 8 C. The coor-
dinates of the plate are shown in the figure. The plate thickness is 10 mm. (Hint: The
nodal displacements are all equal to zero. Therefore, the stresses can be determined
from {s} 5 2[ D]{«T }.)

■■Figure P15 –18

15.19 If the thermal expansion coefficient of a bar is given by a 5 a 0 (1 1 x / L ), determine

the thermal force matrix. Let the bar have length L, modulus of elasticity E, and
cross-sectional area A.
15.20 Assume the temperature function to vary linearly over the length of a bar as
T 5 t1 1 t 2 x ; that is, express the temperature function as {T} 5 [ N ]{t}, where
[N] is the shape function matrix for the two-node bar element. In other words,
[ N ] 5 [1 2 x /L x /L ]. Determine the force matrix in terms of E, A, a , L, t1, and t2.
[Hint: Use Eq. (15.1.18).]
15.21 Derive the thermal force matrix for the axisymmetric element of Chapter 9. [Also
see Eq. (15.1.27).]

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 Problems 759

Using a computer program, solve the following problems.

15.22 The square plate in Figure P15 –22 is subjected to uniform heating of 50 8 F. Deter-
mine the nodal displacements and element stresses. Let the element thickness be
t 5 0.1 in., E 5 30 3 106 psi , ν 5 0.33, and a 5 10 3 1026 / 8 F. Then fix the left
and right nodes and repeat the problem. Compare answers.

■■Figure P15 –22 ■■Figure P15 –23

15.23 The square plate in Figure P15 –23 has element 1 made of steel with E 5 30 3 106 psi ,
ν 5 0.33, and a 5 10 3 1026 / 8 F and element 2 made of a material with
E 5 15 3 106 psi , ν 5 0.25, and a 5 50 3 1026 / 8 F . Let the plate thickness be
t 5 0.1 in . Determine the nodal displacements and element stresses for both ele-
ments subjected to an 80 8 F temperature increase. Then fix the left and right nodes
and repeat the problem. Compare answers.

15.24 Solve Problem 15.3 using a computer program.

15.25 Solve Problem 15.6 using a computer program.

15.26 The aluminum tube shown in Figure P15 –26 fits snugly into a hole (with surround-
ing material aluminum) at room temperature. If the temperature of the tube is then
increased by 40 8 C, determine the deformed configuration and the stress distribution
of the tube. Let E 5 70 GPa , ν 5 0.33, and a 5 23 3 1026 / 8 C for the tube.

50-mm diameter
30-mm diameter
C
D

40 mm
20 mm 20 mm
z

40 mm

A B y

30 mm

■■Figure P15 –26

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 760 15 | Thermal Stress

15.27 For the solid model of a fixture shown in Figure P15 –27, the inside surface of the
hole is subjected to a temperature increase of 80 8 C. The right end surfaces are fixed.
Determine the von Mises stresses throughout the fixture due to this temperature
increase. What is the largest von Mises stress? Is it a concern against yielding of the
material? Assume the material is AISI 1020 cold-rolled steel.

30 dia.

20 rad.
40 mm
mm
30

15 mm
d
Fixe
d
Fixe
75 m
m m
50 m

(All dimensions in mm units)

■■Figure P15 –27

15.28 For the fixture shown in Figure P15 –28, the inside surfaces of the eight holes are
increased in temperature by 80 8 F. Determine the von Mises stresses throughout the
fixture. What is the largest von Mises stress in the fixture? Is there concern for fail-
ure due to yielding of the material? Assume the material is aluminum alloy 6061-O
(annealed). Fix the inside surface of the upper hole.

0.50 in. R chamfer typ.

0.30
ϕ 0.50
0.10
0.25 R

0
3.0
ϕ 0.60
0.50 dia.
0.30 R 0.60

1.500
R.
15

0.40 .500 ia.

0

.625 0.60 d
0.625
0
0.6 0.150
R
0.775
Y
X 2.50 (All holes 0.60 in. dia. (ϕ)
with 0.50 in. rad. chamfer)

(All dimensions in inch units)

■■Figure P15 –28

Copyright 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.