# Module

4
Analysis of Statically
Indeterminate
Structures by the Direct
Stiffness Method
Version 2 CE IIT, Kharagpur

Lesson
26
The Direct Stiffness
Method: Temperature
Changes and
Fabrication Errors in
Truss Analysis
Version 2 CE IIT, Kharagpur

Instructional Objectives
After reading this chapter the student will be able to
1. Compute stresses developed in the truss members due to temperature
changes.
2. Compute stresses developed in truss members due to fabrication members.
3. Compute reactions in plane truss due to temperature changes and fabrication
errors.

26.1 Introduction
In the last four lessons, the direct stiffness method as applied to the truss
analysis was discussed. Assembly of member stiffness matrices, imposition of
boundary conditions, and the problem of inclined supports were discussed. Due
to the change in temperature the truss members either expand or shrink.
However, in the case of statically indeterminate trusses, the length of the
members is prevented from either expansion or contraction. Thus, the stresses
are developed in the members due to changes in temperature. Similarly the error
in fabricating truss members also produces additional stresses in the trusses.
Both these effects can be easily accounted for in the stiffness analysis.

26.2 Temperature Effects and Fabrication Errors

Version 2 CE IIT, Kharagpur

Consider truss member of length L, area of cross section A as shown in
Fig.26.1.The change in length Δl is given by

Δ l = αL Δ T

(26.1)

where α is the coefficient of thermal expansion of the material considered. If the
member is not allowed to change its length (as in the case of statically
indeterminate truss) the change in temperature will induce additional forces in the
member. As the truss element is a one dimensional element in the local
coordinate system, the thermal load can be easily calculated in global coordinate system by

( p1′)t = AE ΔL

(26.2a)

( p2′ )t = − AE ΔL

(26.2b)

or

{(p ) }= AEΔL⎧⎨+− 11⎫⎬
'

t

(26.3)

The equation (26.3) can also be used to calculate forces developed in the truss
member in the local coordinate system due to fabrication error. ΔL will be
considered positive if the member is too long. The forces in the local coordinate
system can be transformed to global coordinate system by using the equation,

⎧( p1 )t ⎫ ⎡cos θ
⎪ ⎢

⎪( p 2 )t ⎪ ⎢ sin θ
⎬=

⎪( p 3 )t ⎪ ⎢⎢ 0
⎪( p ) ⎪ ⎣ 0
⎩ 4 t⎭

0 ⎤
'
0 ⎥⎥ ⎧⎪ p1

cos θ ⎥ ⎪ p 2'
⎥⎩
sin θ ⎦

( ) ⎫⎪
( ) ⎬⎪⎭
t

(26.4a)

t

where ( p1 )t , ( p 2 )t and ( p 3 )t , ( p 4 )t are the forces in the global coordinate system at
nodes 1 and 2 of the truss member respectively Using equation (26.3), the
equation (26.4a) may be written as,
⎧ ( p1 )t ⎫
⎧ cos θ ⎫
⎪( p ) ⎪

⎪ 2 t⎪
⎪ sin θ ⎪

⎬ = AE ΔL ⎨

⎪ ( p3 )t ⎪
⎪− cos θ ⎪
⎪⎩( p4 )t ⎪⎭
⎪⎩ − sin θ ⎪⎭

(26.4b)

Version 2 CE IIT, Kharagpur

The force displacement equation for the entire truss may be written as,

{p} = [k ]{u} + {( p) t }

(26.5)

where , {p} is the vector of external joint loads applied on the truss and {( p )t } is the
vector of joint loads developed in the truss due to change in
temperature/fabrication error of one or more members. As pointed out earlier. in
the truss analysis, some joint displacements are known due to boundary
conditions and some joint loads are known as they are applied
externally.Thus,one could partition the above equation as,
⎧ p k ⎫ ⎡[k11 ]
⎨ ⎬=⎢
⎩ p u ⎭ ⎣[k 21 ]

[k12 ]⎤ ⎧{u u }⎫ ⎧⎪( p k )t ⎫⎪
+
[k 22 ]⎥⎦ ⎨⎩{u k }⎬⎭ ⎨⎪⎩( pu )t ⎬⎪⎭

(26.6)

where subscript u is used to denote unknown quantities and subscript k is used
to denote known quantities of forces and displacements. Expanding equation
(26.6),

{p k } = [k11 ]{u u }+ [k12 ]{u k }+ {( p k )t }

{ pu } = [ k21 ]{uu } + [ k22 ]{uk } + {( pu )t }

(26.7a)
(26.7b)

If the known displacement vector {u k } = {0} then using equation (26.2a) the
unknown displacements can be calculated as

{u u } = [k11 ]−1 ({p k }− {( p k )t })
If {u k } ≠ 0 then
{u u } = [k u ]−1 ({p k }− [k12 ]{u k }− {( p k )t })

(26.8a)

(26.8b)

After evaluating unknown displacements, the unknown force vectors are
calculated using equation (26.7b).After evaluating displacements, the member
forces in the local coordinate system for each member are evaluated by,

{p ′} = [k ′][T ]{u} + {p ′}t

(26.9a)

or

⎧⎪ p1' ⎫⎪
⎨ '⎬
⎪⎩ p 2 ⎪⎭

=

AE
L

⎡ 1 − 1⎤ ⎡cos θ
⎢− 1 1 ⎥ ⎢ 0

⎦⎣

sin θ
0

0
cos θ

⎧u1 ⎫
'
⎪ ⎪
0 ⎤ ⎪v1 ⎪ ⎧⎪ p1
⎨ ⎬+⎨
sin θ ⎥⎦ ⎪u 2 ⎪ ⎪ p 2'

⎪⎩v 2 ⎪⎭

( ) ⎫⎪
( ) ⎬⎪⎭
t

t

Version 2 CE IIT, Kharagpur

Expanding the above equation, yields

{p1′} =

AE
{cosθ
L

sin θ

− cos θ

⎧u1 ⎫
⎪v ⎪
⎪ ⎪
− sin θ }⎨ 1 ⎬ + AEΔL
⎪u 2 ⎪
⎪⎩v 2 ⎪⎭

(26.10a)

⎧ u1 ⎫
⎪v ⎪
⎪ ⎪
sin θ } ⎨ 1 ⎬ − AE ΔL
⎪u2 ⎪
⎪⎩ v2 ⎪⎭

(26.10b)

And,

{ p2′ } =

AE
{− cos θ
L

− sin θ

cos θ

Few problems are solved to illustrate the application of the above procedure to
calculate thermal effects /fabrication errors in the truss analysis:Example 26.1
Analyze the truss shown in Fig.26.2a, if the temperature of the member (2) is
o
raised by 40 C .The sectional areas of members in square centimeters are
5
2
o
shown in the figure. Assume E = 2 × 10 N / mm and α = 1/ 75, 000 per C .

Version 2 CE IIT, Kharagpur

The numbering of joints and members are shown in Fig.26.2b. The possible
global displacement degrees of freedom are also shown in the figure. Note that
lower numbers are used to indicate unconstrained degrees of freedom. From the
figure it is obvious that the displacements u 3 = u 4 = u 5 = u 6 = u 7 = u 8 = 0 due to
boundary conditions.
o
The temperature of the member (2) has been raised by 40 C . Thus,

ΔL = αLΔT
1
ΔL =
3 2 (40 ) = 2.2627 × 10 −3 m
75000

( )

(1)

The forces in member (2) due to rise in temperature in global coordinate system
can be calculated using equation (26.4b).Thus,

⎧( p 5 ) t ⎫
⎧cos θ ⎫

⎪sin θ ⎪

⎪( p 6 )t ⎪

⎬ = AEΔL ⎨

⎪− cos θ ⎪
⎪( p1 )t ⎪
⎪⎩− sin θ ⎪⎭
⎪( p ) ⎪
⎩ 2 t⎭

(2)

For member (2),
A = 20cm 2 = 20 × 10 −4 m 2 and θ = 45 o
Version 2 CE IIT, Kharagpur

⎧ 1 ⎫
⎪ 2 ⎪

⎧( p5 )t ⎫
1

⎪⎪
⎪( p6 )t ⎪
−4
−3
11
3⎪ 2

⎬ = 20 × 10 × 2 × 10 × 2.2627 × 10 /10 ⎨

1
p
(
)
⎪ 1 t⎪
⎪−

⎪ p ⎪
⎪ 2⎪
⎩( 2 )t ⎭

1
⎪−

⎪⎩ 2 ⎭⎪

⎧( p5 )t ⎫
⎧1 ⎫

⎪1 ⎪
⎪( p6 )t ⎪
⎪ ⎪

⎬ = 150.82 ⎨ ⎬ kN
⎪( p1 )t ⎪
⎪−1⎪
⎪ p ⎪
⎪⎩−1⎪⎭
⎩( 2 )t ⎭

(3)

(4)

In the next step, write stiffness matrix of each member in global coordinate
system and assemble them to obtain global stiffness matrix
Element (1): θ = 0 0 , L = 3m, A = 15 × 10 −4 m 2 ,nodal points 4-1

⎡1

15 × 10−4 × 2 × 1011 ⎢ 0
'
⎡⎣ k ⎤⎦ =
⎢ −1
3 × 103

⎣0

0 −1 0 ⎤
0 0 0 ⎥⎥
0 1 0⎥

0 0 0⎦

(5)

Member (2): θ = 45 o , L = 3 2m, A = 20 × 10 −4 m 2 , nodal points 3-1

[k ] = 20 ×10
2

−4

× 2 × 10

3 2

11

0.5 − 0.5 − 0.5⎤
⎡ 0 .5
⎢ 0 .5
0.5 − 0.5 − 0.5⎥⎥

⎢− 0.5 − 0.5 0.5
0.5 ⎥

0.5 ⎦
⎣− 0.5 − 0.5 0.5

(6)

Member (3): θ = 90 o , A = 15 ×10 −4 m 2 , L = 30m, nodal points 2-1

Version 2 CE IIT, Kharagpur

⎡0 0

15 ×10 × 2 × 10 ⎢0 1
⎡⎣ k 3 ⎤⎦ =
3 x 103 ×103 ⎢0 0

⎣0 −1
−4

11

0 0⎤
0 −1⎥⎥
0 0⎥

0 1⎦

(7)

The global stiffness matrix is of the order 8× 8 ,assembling the three member
stiffness matrices, one gets
⎡ 147.14 47.14
⎢ 47.14 147.14

⎢ 0
0

0
− 100
[k ] = 10 3 ⎢⎢
− 47.14 − 47.14

⎢− 47.14 − 47.14
⎢ − 100
0

0
⎢⎣ 0

− 47.14 − 47.14 − 100
0
0
0
0 − 100 − 47.14 − 47.14
0
0
0
0
0
0 100
0
0
0
0
0
0
0

0
0
0
0

47.14
47.14
0
0

47.14
47.14
0
0

0
0
100
0

0⎤
0⎥⎥
0⎥

0⎥
0⎥

0⎥
0⎥

0⎥⎦

(8)

Writing the load displacement equation for the truss
⎡ 147.14

⎧ p1 ⎫
⎢ 47.14
⎪ ⎪

⎪ p2 ⎪
⎢ 0
⎪ p3 ⎪

⎪ ⎪
0
3⎢
⎪ p4 ⎪
⎨ ⎬ = 10 ⎢
− 47.14
⎪ p5 ⎪

⎪ p6 ⎪
⎢− 47.14
⎪ ⎪
⎪ p7 ⎪
⎢ − 100
⎪p ⎪

⎩ 8⎭

⎣⎢
(9)

0

47.14

0

147.14
0

0
0

− 100

0

− 47.14 0
− 47.14 0
0
0
0

0

− 47.14 − 47.14 − 100 0⎤ ⎧u ⎫
1
⎧− 1⎫
− 100 − 47.14 − 47.14
0
0⎥⎥ ⎪u 2 ⎪
⎪− 1⎪
⎪ ⎪
⎪ ⎪
0
0
0
0
0⎥ ⎪u 3 ⎪
⎪0 ⎪
⎥ ⎪u ⎪
⎪ ⎪
100
0
0
0
0⎥ ⎪ 4 ⎪
⎪0 ⎪
⎨ ⎬ + 640 ⎨ ⎬
u
0
47.14
47.14
0
0⎥ ⎪ 5 ⎪
⎪1 ⎪
⎥ ⎪u 6 ⎪
⎪1 ⎪
0
47.14
47.14
0
0⎥ ⎪ ⎪
⎪ ⎪
⎪u 7 ⎪
⎪0 ⎪

0
0
0
100 0 ⎪ ⎪
⎪0 ⎪
⎥ ⎩u8 ⎭
⎩ ⎭
0
0
0
0
0⎦⎥
0

In the present case, the displacements u1 and u 2 are not known. All other
displacements are zero. Also p1 = p 2 = 0 (as no joint loads are applied).Thus,

Version 2 CE IIT, Kharagpur

⎧ p1 ⎫
− 47.14 − 47.14 − 100
0
⎡ 147.14 47.14 0
⎪ ⎪

0
⎪ p2 ⎪
⎢ 47.14 147.14 0 − 100 − 47.14 − 47.14
⎪ p3 ⎪
⎢ 0
0
0
0
0
0
0
⎪ ⎪

− 100 0 100
0
0
0
⎪ p4 ⎪
⎢ 0
⎨ ⎬== ⎢
p

47
.
14
47
.
14
0
0
47
.
14
47
.
14
0
⎪ 5⎪

⎪ p6 ⎪
0
47.14
47.14
0
⎢− 47.14 − 47.14 0
⎪ ⎪

− 100
0
0
0
0
0
100
⎪ p7 ⎪

⎪p ⎪
0
0
0
0
0
0
⎣⎢ 0
⎩ 8⎭

0⎤
0⎥⎥
0⎥

0⎥
0⎥

0⎥
0⎥

0⎦⎥

⎧u1 ⎫
⎧− 1⎫
⎪ ⎪
⎪− 1⎪
⎪u 2 ⎪
⎪ ⎪
⎪u 3 ⎪
⎪0 ⎪
⎪ ⎪
⎪ ⎪
⎪u 4 ⎪
⎪0 ⎪
⎨ ⎬ + 640⎨ ⎬
u
⎪ 5⎪
⎪1 ⎪
⎪u 6 ⎪
⎪1 ⎪
⎪ ⎪
⎪ ⎪
⎪u 7 ⎪
⎪0 ⎪
⎪u ⎪
⎪0 ⎪
⎩ ⎭
⎩ 8⎭

(10)

Thus unknown displacements are

⎧u1 ⎫ 1
⎨ ⎬= 3
⎩u2 ⎭ 10

−1

⎡147.14 47.14 ⎤ ⎧0 ⎫
⎧−1⎫

(
150.82

⎨ ⎬)
⎢ 47.14 147.14 ⎥

⎦ ⎩0 ⎭
⎩−1⎭

(11)

u1 = 7.763 ×10−4 m
u2 = 7.763 × 10−4 m
Now reactions are calculated as
⎧ p3 ⎫
0 ⎤
0
0
0 0⎤ ⎧0⎫
⎧0⎫
⎡0 0
⎡ 0
⎪ ⎪
⎪0⎪

0
0
0 0⎥ ⎪0⎪
− 100 ⎥
⎪ p4 ⎪
⎪ ⎪
⎢0 100
⎢ 0
⎪⎪ p 5 ⎪⎪

u
47
.
14
47
.
14
0
0
47
.
14
47
.
14
0
0
0

⎪⎪1 ⎪⎪

1
3
⎥ ⎨ ⎬ + 640⎨ ⎬
⎥⎨ ⎬ + ⎢
⎨ ⎬ = 10 ⎢
⎢− 47.14 − 47.14⎥ ⎩u 2 ⎭ ⎢0 0 47.14 47.14 0 0⎥ ⎪0⎪
⎪1 ⎪
⎪ p6 ⎪
⎢0 0
⎢ − 100
⎪0⎪
⎪ p7 ⎪
0 ⎥
0
0
100 0⎥ ⎪0⎪
⎥⎪ ⎪

⎪ ⎪
⎪ ⎪
⎪⎩0⎪⎭
0 ⎥⎦
0
0
0 0⎥⎦ ⎪⎩0⎪⎭
⎪⎩ p 8 ⎪⎭
⎢⎣0 0
⎢⎣ 0

⎧ p3 ⎫ ⎧0

⎪p ⎪ ⎪

⎪ 4 ⎪ ⎪−77.63⎪
⎪⎪ p5 ⎪⎪ ⎪⎪77.63 ⎪⎪
⎨ ⎬=⎨
⎬ kN
⎪ p6 ⎪ ⎪77.63 ⎪
⎪ p7 ⎪ ⎪−77.63⎪
⎪ ⎪ ⎪

⎪⎭
⎩⎪ p8 ⎭⎪ ⎪⎩0

(12)

Version 2 CE IIT, Kharagpur

The support reactions are shown in Fig.26.2c.The member forces can be easily
calculated from reactions. The member end forces can also be calculated by
using equation (26.10a) and (26.10b). For example, for member (1),
θ = 0o

⎧0

⎪0

p 2' = 10 3 × 100 [− 1 0 1 0] ⎨7.763 × 10−4 ⎬

⎪7.763 × 10−4 ⎪

(13)

= 77.763 kN. Thus the member (1) is in tension.
Member (2)
θ = 45 o
⎧0

⎪0

p 2' = 10 3 × 94.281 [-0.707 -0.707 0.707 0.707] ⎨
−3 ⎬
3
.
2942
10
×

⎪3.2942 × 10 −3 ⎪

p 2′ = −109.78 kN.
Thus member (2) is in compression

Version 2 CE IIT, Kharagpur

Example 26.2
Analyze the truss shown in Fig.26.3a, if the member BC is made 0.01m too short
before placing it in the truss. Assume AE=300 kN for all members.

Version 2 CE IIT, Kharagpur

Solution
A similar truss with different boundary conditions has already been solved in
example 25.1. For the sake of completeness the member of nodes and members
are shown in Fig.26.3b.The displacements u 3 , u 4 , u 5 , u 6 , u 7 and u 8 are zero due to
boundary conditions. For the present problem the unconstrained degrees of
freedom are u1 and u 2 .The assembled stiffness matrix is of the order 8× 8 and is
available in example 25.1.
In the given problem the member (2) is short by 0.01m.The forces developed in
member (2) in the global coordinate system due to fabrication error is
⎧( p 3 )0 ⎫
⎧cos θ ⎫

(
)
p
⎪ 4 0 ⎪ AE (− 0.01) ⎪sin θ ⎪
=

4
⎪− cos θ ⎪
⎪( p1 )0 ⎪
⎪⎩− sin θ ⎪⎭
⎪( p ) ⎪
⎩ 2 0⎭
⎧0

⎪− 0.75⎪

= ⎪⎨
⎬ kN
0

⎪⎩0.75 ⎪⎭

(1)

Now force-displacement relations for the truss are

Version 2 CE IIT, Kharagpur

⎧ p1 ⎫
0
0
0
− 0.054 0.094 − 0.054 − 0.094⎤ ⎧u1 ⎫ ⎧0

⎡ 0.108
⎪ ⎪
⎥ ⎪u ⎪ ⎪0.75 ⎪

p
0
0
.
575
0
0
.
25
0
.
094
0
.
162
0
.
094
0
.
162

⎪ 2⎪

⎥⎪ 2 ⎪ ⎪

⎪ p3 ⎪

⎢ 0
0
0.866
0
0
0 ⎥ ⎪u 3 ⎪ ⎪0
− 0.433
− 0.433
⎪ ⎪

⎥⎪ ⎪ ⎪

0
0
0.25
0
0
0
0 ⎥ ⎪u 4 ⎪ ⎪− 0.75⎪
− 0.25
⎪ p4 ⎪

⎨ ⎬ = AE ⎢

⎨ ⎬+⎨
0
0.487 − 0.094
0
0 ⎥ ⎪u 5 ⎪ ⎪0
− 0.054 0.094 − 0.433
⎪ p5 ⎪

⎪ p6 ⎪

0
0
.
094
0
.
162
0
0
0
.
094
0
.
162
0
0

u
⎥ 6

⎪ ⎪

⎪ ⎪ ⎪

0
0
0
0.487 0.0934 ⎪u 7 ⎪ ⎪0
− 0.054 − 0.094 − 0.433
⎪ p7 ⎪

⎪p ⎪
0
0
0
0
0.0934 0.162 ⎦⎥ ⎩u 8 ⎭ ⎩0
⎢⎣− 0.094 − 0.162

⎩ 8⎭

(2)
Note that u 3 = u 4 = u 5 = u 6 = u 7 = u 8 = 0
Thus, solving
0 ⎤
⎧u1 ⎫
1 ⎡0.108
⎨ ⎬=

0.575⎥⎦
⎩u 2 ⎭ AE ⎣ 0

−1

⎛ ⎧0⎫ ⎧0 ⎫ ⎞
⎜⎨ ⎬ − ⎨
⎬ ⎟⎟
⎜ 0
0
.
75

⎭⎠

(3)

u1 = 0
and, u 2 = −4.3478 ×10 −3 m

(4)

Reactions are calculated as,
⎧ p3 ⎫
0 ⎤
⎧0

⎡ 0
⎪ ⎪

⎢ 0

p
− 0.25 ⎥
⎪ 4⎪
⎪− 0.75⎪

⎪⎪ p 5 ⎪⎪
⎢− 0.054 0.094 ⎥ ⎧u1 ⎫ ⎪⎪0
⎪⎪
⎥⎨ ⎬ + ⎨

⎨ ⎬ = AE ⎢
⎢ 0.094 − 0.162⎥ ⎩u 2 ⎭ ⎪0

⎪ p6 ⎪
⎢− 0.054 − 0.094⎥
⎪0

⎪ p7 ⎪

⎪ ⎪
⎩⎪0
⎣⎢− 0.094 − 0.162⎦⎥
⎭⎪
⎩⎪ p 8 ⎭⎪

⎧ p 3 ⎫ ⎧0

⎪ ⎪ ⎪

⎪ p 4 ⎪ ⎪− 0.424⎪
⎪⎪ p 5 ⎪⎪ ⎪⎪− 0.123⎪⎪
⎨ ⎬=⎨

⎪ p 6 ⎪ ⎪0.211 ⎪
⎪ p 7 ⎪ ⎪0.123 ⎪
⎪ ⎪ ⎪

⎪⎩ p 8 ⎪⎭ ⎪⎩0.211 ⎪⎭

(5)

(6)

The reactions and member forces are shown in Fig.26.3c. The member forces
can also be calculated by equation (26.10a) and (26.10b). For example, for
member (2),

θ = 90 o

Version 2 CE IIT, Kharagpur

⎧u 3 ⎫
⎪ ⎪
⎪u 4 ⎪ AEΔL
300
'
p2 =
[0 -1 0 1] ⎨u ⎬ − L
4
⎪ 1⎪
⎪⎩u 2 ⎪⎭
=

300(−0.01)
300
− 4.3478 × 10 −3 −
4
4

(

)

= 0.4239 ≅ 0.424 kN

(7)

Example 26.3
Evaluate the member forces of truss shown in Fig.26.4a.The temperature of the
member BC is raised by 40 o C and member BD is raised by 50 o C .Assume
AE=300KN for all members and α =

1
per
75000

o

C.

Version 2 CE IIT, Kharagpur

Solution
For this problem assembled stiffness matrix is available in Fig.26.4b.The joints
and members are numbered as shown in Fig.26.4b. In the given problem
u1 , u 2 , u 3 , u 4 and u 5 represent unconstrained degrees of freedom. Due to support
conditions, u 6 = u 7 = u 8 = 0 .
The temperature of the member (2) is raised by 50 o C.Thus,
ΔL2 = αLΔT =

1
× 5 × 50 = 3.333 × 10 −3 m
75000

(1)

The forces are developed in member (2), as it was prevented from expansion.
⎧( p 7 ) f

⎪( p 8 ) f

⎪( p1 ) f
⎪( p )
⎩ 2 f

⎧cos θ ⎫

−3 ⎪sin θ ⎪
⎬ = 300 × 3.333 × 10 ⎨

⎪− cos θ ⎪

⎪⎩− sin θ ⎪⎭

Version 2 CE IIT, Kharagpur

⎧0 ⎫
⎪1 ⎪
⎪ ⎪
=⎨ ⎬
⎪0 ⎪
⎪⎩− 1⎪⎭

(2)

The displacement of the member (5) was raised by 40 o C . Thus,
ΔL5 = αLΔT =

1
× 5 2 × 40 = 3.771×10 −3 m
75,000

The forces developed in member (5) as it was not allowed to expand is
⎧( p 5 )t ⎫
⎧0.707 ⎫

⎪( p 6 )t ⎪
− 3 ⎪0.707 ⎪
⎬ = 300 × 3.771× 10 ⎨

⎪( p 7 )t ⎪
⎪− 0.707 ⎪
⎪( p ) ⎪
⎪⎩− 0.707 ⎪⎭
⎩ 8 t⎭
⎧1 ⎫
⎪1 ⎪
⎪ ⎪
= 0 .8 ⎨ ⎬
⎪− 1⎪
⎪⎩− 1⎪⎭

(3)

The global force vector due to thermal load is
⎧( p1 )t ⎫ ⎧− 0.8⎫

⎪ ⎪

⎪( p 2 )t ⎪ ⎪− 1.8 ⎪
⎪( p ) ⎪ ⎪

⎪ 3 t ⎪ ⎪0 ⎪
⎪⎪( p 4 )t ⎪⎪ ⎪0 ⎪

⎬=⎨

p
(
)
5
t

⎪ ⎪0.8 ⎪
⎪( p ) ⎪ ⎪0.8 ⎪
⎪ 6 t⎪ ⎪

⎪( p 7 )t ⎪ ⎪0 ⎪

⎪ ⎪1

⎩⎪( p 8 )t ⎭⎪ ⎩

(4)

Writing the load-displacement relation for the entire truss is given below.

Version 2 CE IIT, Kharagpur

⎧ p1 ⎫
0.071 − 0.20
0
0
0 ⎤ ⎧u1 ⎫ ⎧− 0.8⎫
− 0.071 − 0.071
⎡ 0.271
⎪ ⎪
⎪ ⎪

0.271
0
0
0
− 0.071 − 0.071
− 0.2 ⎥⎥ ⎪u 2 ⎪ ⎪⎪− 1.8 ⎪⎪
⎪ p2 ⎪
⎢ 0.071
⎪ p3 ⎪
⎢ − 0.20
0
0.271 − 0.071
0
0
− 0.071 0.071 ⎥ ⎪u 3 ⎪ ⎪0 ⎪
⎪ ⎪

⎥⎪ ⎪ ⎪

0
0
0
0.071
0.071 ⎥ ⎪u 4 ⎪ ⎪0 ⎪
− 0.071 0.129
− 0.2
⎪ p4 ⎪

⎬ (5)
⎨ ⎬ = AE ⎢
⎨ ⎬+⎨
0
0
0.271
0.071
0 ⎥ ⎪u 5 ⎪ ⎪0.8 ⎪
− 0.071 − 0.071
− 0.2
⎪ p5 ⎪

⎪ p6 ⎪
0
0.071
0.271
0
0 ⎥ ⎪u 6 ⎪ ⎪0.8 ⎪
− 0.2
⎢− 0.071 − 0.071

⎪ ⎪
⎪ ⎪ ⎪

0
0
0
0.271 − 0.071⎥ ⎪u 7 ⎪ ⎪0 ⎪
− 0.071 0.071
− 0.2
⎪ p7 ⎪

⎪p ⎪
0.071 − 0.071
0
0
− 0.2
− 0.071 0.271 ⎦⎥ ⎪⎩u 8 ⎪⎭ ⎪⎩1
⎢⎣ 0

⎩ 8⎭

In
the
above
u 6 = u 7 = u8 = 0 .

problem

p1 = p 2 = p 3 = p 4 = p 5 = p 6 = p 7 = p 8 = 0 and

Thus solving for unknown displacements,
− 0.2
− 0.071⎤⎛ ⎧0⎫ ⎧− 0.8⎫ ⎞
0.071
0
⎧u1 ⎫
⎡ 0.271
⎪ ⎪
⎢ 0.071
⎥⎜ ⎪0⎪ ⎪− 1.8 ⎪ ⎟
u

0
.
271
0
0
0
.
071
⎪⎪ 2 ⎪⎪
⎪⎪ ⎟
⎥⎜ ⎪⎪ ⎪⎪ ⎪⎪
1 ⎢

0
=

u
0

0
.
20
0
0
.
271
0
.
071
0
⎨ 3⎬
⎬⎟
⎜⎨ ⎬ ⎨
AE

⎪u ⎪
− 0.071 0.129
0
0 ⎥⎜ ⎪0⎪ ⎪0 ⎪ ⎟
⎢ 0
⎪ 4⎪
⎪⎟
⎜⎪ ⎪ ⎪
⎢− 0.071 − 0.071
0
0
0.271 ⎥⎦⎜⎝ ⎩⎪0⎭⎪ ⎩⎪0.8 ⎭⎪ ⎟⎠

⎩⎪u 5 ⎭⎪

(5)

Solving equation (5), the unknown displacements are calculated as
u1 = 0.0013m, u 2 = 0.0020 m, u 3 = −0.0005m, u 4 = 0

(6)

u 5 = −0.0013m

Now, reactions are computed as,
⎧u1 ⎫
⎪ ⎪
− 0.2 0.071⎤ ⎪u 2 ⎪ ⎧0.8⎫
0
⎧ p 6 ⎫ ⎡− 0.071 − 0.071
⎪ ⎪ ⎢
⎪ ⎪ ⎪ ⎪
− 0.071 0.071 − 0.2 ⎥⎥ ⎨u 3 ⎬ + ⎨0 ⎬
0
⎨ p7 ⎬ = ⎢ 0
⎪p ⎪ ⎢ 0
− 0.2
0.071 − 0.071
0 ⎥⎦ ⎪u 4 ⎪ ⎪⎩1 ⎪⎭
⎩ 8⎭ ⎣
⎪ ⎪
⎪⎩u 5 ⎪⎭

(7)

All reactions are zero as truss is externally determinate and hence change in
temperature does not induce any reaction. Now member forces are calculated by
using equation (26.10b)
Member (1): L=5m, θ = 0 o

Version 2 CE IIT, Kharagpur

⎧u 3 ⎫
⎪ ⎪
u
AE
[-1 0 1 0] ⎪⎨ 4 ⎪⎬
p 2' =
5
⎪u1 ⎪
⎪⎩u 2 ⎪⎭

(8)

p 2' = 0.1080 Kn

Member 2: L=5m, θ = 90 o ,nodal points 4-1
⎧u 7 ⎫
⎪ ⎪
u
AE
[0 -1 0 1] ⎪⎨ 8 ⎪⎬ − 300 × 3.771×10 −5
p 2' =
5
⎪u1 ⎪
⎪⎩u 2 ⎪⎭

(9)

=0.1087 kN
Member (3): L=5m, θ = 0 o ,nodal points 3-4
⎧u 5 ⎫
⎪ ⎪
u
300
p 2' =
[-1 0 1 0] ⎪⎨ 6 ⎪⎬
5
⎪u 7 ⎪
⎪⎩u 8 ⎪⎭

(10)

=0.0780kN
Member (4): θ = 90 o , L = 5m, nodal points 3-2
⎧u 5 ⎫
⎪ ⎪
u
300
p 2' =
[0 -1 0 1] ⎪⎨ 6 ⎪⎬ =0
5
⎪u 3 ⎪
⎪⎩u 4 ⎪⎭

(11)

Member (5): θ = 45 o , L = 5 2 ,nodal points 3-1
⎧u 5 ⎫
⎪ ⎪
u
300
p 2' =
[-0.707 -0.707 0.707 0.707] ⎪⎨ 6 ⎪⎬ − 300 × 3.333 ×10 −3
5 2
⎪u1 ⎪
⎪⎩u 2 ⎪⎭

(12)

=-0.8619 kN
Member (6) : θ = 135 o , L = 5 2 ,nodal points 4-2
Version 2 CE IIT, Kharagpur

⎧u 7 ⎫
⎪ ⎪
u
300
p 2' =
[0.707 -0.707 -0.707 0.707] ⎪⎨ 8 ⎪⎬ = 0.0150 kN.
5 2
⎪u 3 ⎪
⎪⎩u 4 ⎪⎭

(13)

Summary
In the last four lessons, the direct stiffness method as applied to the truss
analysis was discussed. Assembly of member stiffness matrices, imposition of
boundary conditions, and the problem of inclined supports were discussed. Due
to the change in temperature the truss members either expand or shrink.
However, in the case of statically indeterminate trusses, the length of the
members is prevented from either expansion or contraction. Thus, the stresses
are developed in the members due to changes in temperature. Similarly the
errors in fabricating truss members also produce additional stresses in the
trusses. In this lesson, these effects are accounted for in the stiffness analysis. A
couple of problems are solved.

Version 2 CE IIT, Kharagpur

Sign up to vote on this title