Joint Displacements and Forces

1. Coordinate Systems
y z

x x

z y
Fig. 1: Coordinate System1 Fig. 2: Coordinate System2
(widely used and also applied in this course) (used in some formulations)

2. Sign Convention for Joint Displacements and Forces

uy Fy

y My

ux Fx
x Mx
z Mz

uz Fz
Fig. 3: Sign convention for Displacements Fig. 4: Sign convention for Forces

3. Sign Convention for Two-dimensional Problems

uy Fy

z ux Mz Fx

Fig. 5: Two-Dimensional Displacements Fig. 6: Two-Dimensional Forces

1
 Stiffness Matrix for Truss Members in the Local Axes System

Consider a truss member AB subjected to forces (XA, YA) and (XB, YB) at joints A and B.

YA YB

XA XB
A B

Assume that the length of the member is L, its modulus of elasticity is E and cross-sectional area A.
The axial stiffness of the member, Sx = Load to produce unit deflection = EA/L
Also assume that the member has no flexural or shear stiffness.

If the displacements of joints A and B are (uA, vA) and (uB, vB), the effect of the external forces may result in
the following cases.

(uA=1) (vA=1)
uA=1
vA=1
Sx uA Sx uA
A B

(uB=1) (vB=1)
uB=1
vB=1
S x uB S x uB
A B

Equilibrium equations:

Fx(A) = 0  XA = Sx uA + 0  Sx uB + 0 ……………….(1)
Fy(A) = 0  YA = 0 + 0 + 0 + 0 ……………….(2)
Fx(B) = 0  XB = Sx uA + 0 + Sx uB + 0 ……………….(3)
Fy(B) = 0  YB = 0 + 0 + 0 + 0 ……………….(4)
Eqs. (1)~(4) can be summarized in matrix form as

Sx 0 Sx 0 uA XA
0 0 0 0 vA = YA
Sx 0 Sx 0 uB XB
0 0 0 0 vB YB
 KmL umL = pmL ………………(5)
L
where Km = The stiffness matrix of member AB in the local axis system,
umL = The displacement vector of the member in the local axis system, and
pmL = The force vector of the member in the local axis system

2
 Transformation of Stiffness Matrix from Local to Global Axes

The member matrices formed in the local axes system can be transformed into the global axes system by
considering the angles they make with the horizontal.
The local vectors and global vectors are related by the following equations.
v BG
v BL u BL
B u BG

vA G 
vAL
A uAG
uA L
Local and global joint displacements of a truss member

uAL = uAG cos  + vAG sin  …………………(6)

vAL = uAG sin  + vAG cos  …………………(7)
uBL = uBG cos  + vBG sin  …………………(8)
vBL = uBG sin  + vBG cos  …………………(9)
In matrix form
uAL cos  sin  0 0 uAG
vAL sin  cos  0 0 vAG
=
uBL 0 0 cos  sin  uBG
vBL 0 0 sin  cos  vBG
 umL = Tm umG …………………(10)
where Tm is called the transformation matrix for member AB, which connects the displacement vector
umL in the local axes of AB with the displacement vector umG in the global axes.
A similar expression can be obtained for the force vectors pmL and pmG; i.e.,
 pmL = Tm pmG …………………(11)

Eq. (5) can be rewritten as  KmL Tm umG = Tm pmG …………….……(12)

 (Tm-1 KmL Tm) umG = pmG
 (TmT KmL Tm) umG = pmG ...………….……(13)
where Tm is the transpose of the transformation matrix Tm, which is also = Tm-1
T

If (TmT KmL Tm) is written as KmG, the member stiffness matrix in the global axis system, then

C2 CS C2 CS
CS S2 CS S2 [where C = cos , S = sin ]
KmG = Sx
C2 CS C2 CS
CS S 2
CS S2

3
 Assembly of Stiffness Matrix and Load Vector of a Truss

Assemble the global stiffness matrix and write the global load vector of the truss shown below.
Also write the boundary conditions [EA/L = Constant = 500 kip/ft].

10k u8
D 20k
u7

10
u2 u4 u6

u1 u3 u5
A B C

10 10

Member AB: (C = 1, S = 0) Member BC: (C = 1, S = 0) Member BD: (C = 0, S = 1)

1 2 3 4 3 4 5 6 3 4 7 8
1 0 1 0 1 1 0 1 0 3 0 0 0 0 3
KABG = 500 0 0 0 0 2 KBCG = 500 0 0 0 0 4 KBDG = 500 0 1 0 1 4
1 0 1 0 3 1 0 1 0 5 0 0 0 0 7
0 0 0 0 4 0 0 0 0 6 0 1 0 1 8

Member AD: (C = 1/2, S = 1/2) Member CD: (C = 1/2, S = 1/2)

1 2 7 8 5 6 7 8
0.5 0.5 0.5 0.5 1 0.5 0.5 0.5 0.5 5
KADG = 500 0.5 0.5 0.5 0.5 2 KCDG = 500 0.5 0.5 0.5 0.5 6
0.5 0.5 0.5 0.5 7 0.5 0.5 0.5 0.5 7
0.5 0.5 0.5 0.5 8 0.5 0.5 0.5 0.5 8

1 2 3 4 5 6 7 8
1+0.5 0+0.5 1 0 0.5 0.5 1 XA
0+0.5 0+0.5 0 0 0.5 0.5 2 YA
1 0 1+1+0 0+0+0 1 0 0 0 3 XB
0 0 0+0+0 0+0+1 0 0 0 1 4 YB
KG = 500 1 0 1+0.5 00.5 0.5 0.5 5 pG = 0
0 0 00.5 0+0.5 0.5 0.5 6 YC
7
0.5 0.5 0 0 0.5 0.5 0+0.5+0.5 0+0.50.5 20
8
0.5 0.5 0 1 0.5 0.5 0+0.50.5 1+0.5+0.5 10

Boundary Conditions: u1 = 0, u2 = 0, u3 = 0, u4 = 0, u6 = 0

4
 Boundary Conditions, Support Reactions and Member Forces

After assembly of the member stiffness matrices, the equilibrium equations were

1.5 0.5 1 0 0 0 0.5 0.5 u1 XA

0.5 0.5 0 0 0 0 0.5 0.5 u2 YA

1 0 2 0 1 0 0 0 u3 XB

500 0 0 0 1 0 0 0 1 u4 = YB

u5 0
0 0 1 0 1.5 0.5 0.5 0.5
u6 YC
0 0 0 0 0.5 0.5 0.5 0.5
u7 20
0.5 0.5 0 0 0.5 0.5 1 0
u8 10
0.5 0.5 0 1 0.5 0.5 0 2

Applying the boundary conditions (u1 = 0, u2 = 0, u3 = 0, u4 = 0, u6 = 0), the equations are modified to

1.5 0.5 0.5 u5 0 u5 = 22.22  10-3 ft

500 0.5 1 0 u7 = 20  u7 = 51.11  10-3 ft
0.5 0 2 u8 10 u8 = 15.56  10-3 ft

Once displacements are known, support reactions can be calculated from equilibrium equations; i.e.,
XA = 750 u1 + 250 u2 500 u3 + 0 u4 + 0 u5 + 0 u6 250 u7 250 u8 = 0 + 0 + 0 + 0 + 0 + 0 +12.78 3.89
= 8.89k
Similarly, YA = 12.78 3.89 = 8.89k, XB = 11.11k, YB = 7.78k, YC = 5.56 12.78 3.89 = 11.11k

The bar forces can be calculated from the equation PAB = (EA/L) {(uBGuAG) cos  + (vBGvAG) sin }
PAB = 500{(u3u1) cos 0+(u4u2) sin 0}= 0, PBC = 500 {(u5u3) cos 0+(u6u4) sin 0}= 11.11 k
PBD = 500{(u7u3) cos 90+(u8u4) sin 90}=7.78k, PAD= 500{(u7u1)cos 45+(u8u2)sin 45}= 12.57k
PCD = 500 {(u7u5) cos 135+(u8u6) sin 135}= 15.71k

In addition to the externally applied forces if the support C settles 0.10, then u6 = 0.10 is known
Applying boundary conditions (u1 = 0, u2 = 0, u3 = 0, u4 = 0, u6 = 0.10), the equations become

1.5 0.5 0.5 u5 0+250 u6 25 u5 = 33.33  10-3 ft

500 0.5 1 0 u7 = 20250 u6 = 5  u7 = 6.67  10-3 ft
0.5 0 2 u8 10+250 u6 15 u8 = 6.67  10-3 ft

PAB = 0, PBC = 16.67 k, PBD = 3.33k, PAD = 4.71k, PCD = 23.57k

5
 Stiffness Formulation using Equilibrium of Joints

10k u8
D 20k
u7

10
YA YB YC u2 u4 u6

XA XB u1 u3 u5
A B C
EA/L = Constant = 500 kip/ft
10 10

For the truss ABCD, the equilibrium equations of joints A, B, C and D take the following forms when the
equation for member force [i.e., PAB = (EA/L) {(uBG uAG) cos  + (vBG vAG) sin }] is applied
FxA= 0  XA+PAB+PAD cos 45= 0
 XA + 500(u3u1) + 500{(u7u1)cos 45 + (u8u2)sin 45}cos 45 = 0
 500 {(1.0 + 0.5) u1 + 0.5 u2 1.0 u3  0.5 u7  0.5 u8} = XA …………………(1)
FyA = 0  YA + PAD sin 45 = 0  YA + 500 {(u7u1) cos 45 + (u8u2) sin 45} sin 45 = 0
 500 {0.5 u1 + 0.5 u2  0.5 u7  0.5 u8} = YA …………………(2)
FxB = 0  XB  PAB + PBC = 0  XB  500 (u3u1) + 500 (u5u3) = 0
 500 { 1.0 u1 + 2.0 u3  1.0 u5} = XB …………………(3)
FyB = 0  YB + PBD = 0  YB + 500 (u8u4) = 0
 500 {1.0 u4  1.0 u8} = YB …………………(4)
FxC = 0   PBC  PCD cos 45 = 0
 500 (u5u3) 500{(u7u5)cos 135 + (u8u6)sin 135}cos 45 = 0
 500 { 1.0 u3 + (1.0 + 0.5) u5  0.5 u6  0.5 u7 + 0.5 u8} = 0 …………………(5)
FyC = 0  YC + PCD sin 45 = 0  YC + 500 {(u7u5) cos 135 + (u8u6) sin 135} sin 45 = 0
 500 {0.5 u5 + 0.5 u6 + 0.5 u7  0.5 u8} = YC …………………(6)
FxD = 0   20  PAD cos 45 + PCD cos 45 = 0   20  500 {(u7u1)cos 45 + (u8u2)sin 45} cos 45 +
500 {(u7u5) cos 135 + (u8u6) sin 135} cos 45 = 0
 500 {0.5 u1  0.5 u2  0.5 u5 + 0.5 u6 + (0.5 + 0.5) u7 + (0.5  0.5) u8} = 20 ………….……(7)
FyD = 0  10  PBD  PAD sin 45  PCD sin 45= 0
 10  500 (u8u4)  500 {(u7u1) cos 45 + (u8u2) sin45} sin45
 500 {(u7u5) cos 135 + (u8u6) sin 135} sin 45 = 0
 500{0.5 u10.5 u2  1.0 u4 + 0.5 u5  0.5 u6 + (0.5  0.5) u7+ (1.0 + 0.5 + 0.5) u8}= 10 .…………….(8)
Eqs. (1)~(8) are the same equations given by the Stiffness Matrix assembled earlier.
After applying boundary conditions for the known displacements u1, u2, u3, u4 and u6, Eqs. (5), (7) and (8)
can be solved for the three unknown displacements u5, u7 and u8, whereupon Eqs. (1)~(4) and (6) can be
used to calculate the support reactions XA, YA, XB, YB and YC.

6
 Problems on Stiffness Method for Trusses
1. Calculate the horizontal and vertical deflections at joints c and d of the plane truss loaded as shown
below if P = 50 kN, Q = 50 kN [Given: Sx = constant = 10 kN/mm].

a b
10m
c d
P
10m
Q
e f
10m 10m 10m
2. In the truss shown below, ignore the zero-force members and formulate the stiffness matrix, load vector
and write down the boundary conditions [Given: EA/L = constant = 1000 kip/ft].
A B C

25
D E
10 k 10 k
25
45 G 45
F H
4@25 = 100
3. In the truss shown below, ignore the zero-force members and formulate the stiffness matrix, load vector
and write down the boundary conditions [Given: EA/L = constant = 1000 kip/ft].
P P
B C D E F
28.9
J 30
50 I
45
K
H
86.6
60
A
6 @50 = 300 G

4. The tensile forces in members ac, ce and bd of the plane truss described in Question 1 are all equal to
100 kN. Without formulating the stiffness matrix, calculate the
(i) Deflections at joints c and d, (ii) Applied forces P and Q [Given: Sx = constant = 10 kN/mm].
5. For the truss described in Question 2, the force in member DE is 8 k (tension). Calculate the forces in
the other members of the truss and deflections of joints D and E.
6. In the truss abcdef shown below, calculate the applied forces Px and Py needed for joint a to deflect 0.10
rightward and e to deflect 0.10 downward [Given: Sx = constant = 500 kip/ft].

43.3

a b d e
Px

Py 43.3
f

50 25 25 50

7
 Stiffness Matrix for 2-Dimensional Frame Members in the Local Axes System

Consider a frame member AB subjected to forces (XA, YA, MA) and (XB, YB, MB) at joints A and B.

YA YB
MA MB
XA XB
A B

Assume that the length of the member = L, its modulus of elasticity = E, cross-sectional area = A and
moment of inertia about z-axis = I.
The axial stiffness of the member, Sx = Load to produce unit deflection = EA/L
Also assume S1 = shear stiffness = 12EI/L3, S2 = 6EI/L2, S3 = flexural stiffness = 4EI/L, S4 = 2EI/L

If the displacements and rotations of joints A and B are (uA, vA, A), (uB, vB, B) and the fixed-end reactions
are denoted by ‘FE’, the external forces may result in the following cases.
FEYA FEYB

FEXA FEXB
FEMA FEMB

(uA=1) (vA=1) (A=1)

uA=1 S1 S2
vA =1 S2 S2 A=1 S4
Sx Sx S3
A B S1 S2

(uB=1) (vB=1) (B=1) S2

uB=1 S2 S3
S2 vB =1 B=1
Sx Sx S1 S4
A B
S1 S2
Equilibrium equations:
Fx(A) = 0  XA = FEXA + Sx uA + 0 + 0  Sx uB + 0 + 0 ……………….(1)
Fy(A) = 0  YA = FEYA + 0 + S1 vA + S2 A + 0  S1 vB + S2 B ……………….(2)
Mz(A) = 0  MA = FEMA + 0 + S2 vA + S3 A + 0  S2 vB + S4 B ……………….(3)
Fx(B) = 0  XB = FEXB Sx uA + 0 + 0 + Sx uB + 0 + 0 ……………….(4)
Fy(B) = 0  YB = FEYB + 0  S1 vA  S2 A + 0 + S1 vB  S2 B ……………….(5)
Mz(B) = 0  MB = FEMB + 0 + S2 vA + S4 A + 0  S2 vB + S3 B ……………….(6)

Sx 0 0 Sx 0 0 uA XA FEXA
0 S1 S2 0 S1 S2 vA YA FEYA
0 S2 S3 0 S2 S4 A MA FEMA
 = 
Sx 0 0 Sx 0 0 uB XB FEXB
0 S1 S2 0 S1 S2 vB YB FEYB
0 S2 S4 0 S2 S3 B MB FEMB

 KmL umL = qmL  fmL = pmL ………………(7)

where KmL = The stiffness matrix of member AB in the local axis system,
umL = The displacement vector of the member in the local axis system, and
pmL = The force vector of the member in the local axis system
(= qmL  fmL = Imposed load vector  Fixed end reaction vector)

8
 Transformation of Stiffness Matrix from Local to Global Axes

The member matrices formed in the local axes system can be transformed into the global axes system by
considering the angles they make with the horizontal. The local displacements/rotations and global
displacements/rotations are related by the following equations.
v BG
v BL u BL
B u BG
BL, BG

vA G 
L
vA
A uAG
uAL AL, AG
Local and global joint displacements and rotations of a frame member
uAL = uAG cos  + vAG sin  …………………(8)
vAL = uAG sin  + vAG cos  …………………(9)
AL = AG ..………………(10)
uB = uB cos  + vB sin 
L G G
………………..(11)
vBL = uBG sin  + vBG cos  ………………..(12)
BL = BG ………………..(13)
In matrix form, using C = cos , S = sin 

uAL C S 0 0 0 0 uAG
vAL S C 0 0 0 0 vAG
AL 0 0 1 0 0 0 AG
uBL
= 0 0 0 C S 0 uBG
vBL 0 0 0 S C 0 vBG
BL 0 0 0 0 0 1 BG
 umL = Tm umG …………………(14)
where Tm is the transformation matrix for member AB, which connects the displacement vector umL in the
local axes of AB with the displacement vector umG in the global axes.

A similar expression can be obtained for the force vectors pmL and pmG; i.e.,
 pmL = Tm pmG …………………(15)
Eq. (7) can be rewritten as  Km Tm um = Tm pm
L G G
…………….……(16)
 (Tm1 KmL Tm) umG = pmG
 (TmT KmL Tm) umG = pmG ...………….……(17)
where Tm is the transpose of the transformation matrix Tm, which is also = Tm1
T

If (TmT KmL Tm) is written as KmG, the member stiffness matrix in the global axis system, then
KmG umG = pmG ...………….……(18)

9
 Assembly of Stiffness Matrix and Load Vector of a 2D Frame

The general form of the stiffness matrix for any member of a 2-dimensional frame is

SxC2+S1S2 (SxS1)CS S2S (SxC2+S1S2) (SxS1)CS S2S

(SxS1)CS SxS2+S1C2 S2 C (SxS1)CS (SxS2+S1C2) S2 C
G
Km = S2S S2 C S3 S2S S2C S4
(SxC2+S1S2) (SxS1)CS S2S SxC2+S1S2 (SxS1)CS S2S
(SxS1)CS (SxS2+S1C2) S2C (SxS1)CS (SxS2+S1C2) S2C
S2S S2 C S4 S2S S2C S3

Example
Assemble the global stiffness matrix and write the global load vector of the frame shown below.
Also write the boundary conditions [E, A, I are constant for all the members].
10 k
Since E, A, I and L are uniform, so are Sx, S1, S2, S3 and S4
A B 20 k
3 2 4
If E = 50010 ksf, A = 1 ft , I = 0.10 ft , L = 10 ft
Sx = EA/L = 50,000 k/ft
S1 = 12EI/L3 = 600 k/ft, S2 = 6EI/L2 = 3,000 k/rad 5 5
S3 = 4EI/L = 20,000 k-ft/rad, S4 = 2EI/L =10,000 k-ft/rad 1 k/ft
10
For member AB, C = 1, S = 0
For member BC, C = 0, S = 1
C

d.o.k.i = 33 = 9, which are (uA, vA, A), (uB, vB, B) and (uC, vC, C), denoted by u1~u9.

1 2 3 4 5 6 4 5 6 7 8 9
Sx 0 0 Sx 0 0 S1 0 S2 S1 0 S2
0 S1 S2 0 S1 S2 0 Sx 0 0 Sx 0
0 S2 S3 0 S2 S4 KBC = S2 0
G S3 S2 0 S4
KABG =
Sx 0 0 Sx 0 0 S1 0 S2 S1 0 S2
0 S1 S2 0 S1 S2 0 Sx 0 0 Sx 0
0 S2 S4 0 S2 S3 S2 0 S4 S2 0 S3

1 2 3 4 5 6 7 8 9
Sx 0 0 Sx 0 0 u1 XA 0
0 S1 S 2 0 S1 S2 u2 YA 5.0
0 S2 S 3 0 S2 S4 u3 0 12.5
Sx 0 0 Sx+S1 0+0 0+S2 S1 0 S2 u4 20 0 + 5.0
0 S1 S2 0+0 S1+Sx S2+0 0 Sx 0 u5 = 0  5.0 + 0
0 S2 S4 0+S2 S2+0 S3+S3 S2 0 S4 u6 0 12.5+8.33
S1 0 S2 S1 0 S2 u7 XC 5.0
0 Sx 0 0 Sx 0 u8 YC 0
S2 0 S4 S2 0 S3 u9 MC 8.33

Boundary Conditions: u1 = 0, u2 = 0, u7 = 0, u8 = 0, u9 = 0

Therefore, the matrices and vectors can be modified accordingly (similar to the analysis of truss).
Solving the resulting (44) matrix, the following displacements and rotations are obtained
u3 = 8.12104 rad, u4 = 5.14104 ft, u5 = 1.27104 ft, u6 = 3.36104 rad

10
 Stiffness Method for 2-D Frame neglecting Axial Deformations
If axial deformations are neglected in the problem shown before, the displacements u4 and u5 are zero and
the only unknown displacements are the rotations u3 and u6. In that case, modified equilibrium equations are
S3 u3 + S4 u6 = 12.5  20103 u3 + 10103 u6 = 12.5
and S4 u3 + 2S3 u6 = 4.17  10103 u3 + 40103 u6 = 4.17
[Note: S1 = 600 k/ft, S2 = 3,000 k/rad, S3 = 20,000 k-ft/rad, S4 = 10,000 k-ft/rad]
Solving, u3 = 7.74104 rad, u6 = 2.98104 rad [instead of 8.12104, 3.36104 found before]
If the axial deformations are neglected, the calculations and formulations are much simplified without
significant loss of accuracy.
Neglecting the axial deformations, the earlier problem can be formulated as shown below

u3 10 k u6
A B Here, d.o.k.i. = 2
20 k There can be three cases of response
(i) Case0: The fixed-end reactions
(ii) Case1: The reactions due to u3
(iii) Case2: The reactions due to u6

5k 5k S2 S2
12.5 k 12.5 k S4 S3
25 k S2
8.33 k S3 S4
S2 S2 S3

8.33 k S4
Case 0 (FER) 5k Case 1 (u3 = 1) Case 2 (u6 = 1) S2

Equilibrium equations:
Mz(A) = 0  12.5 + S3 u3 + S4 u6 = 0  20103 u3 + 10103 u6 = 12.5
Mz(B) = 0  12.5 + 8.33 + S4 u3 + (S3+S3) u6 = 0  10103 u3 + 40103 u6 = 4.17
Solving the two equations, u3 = 7.7410 rad, u6 = 2.98104 rad
4

Calculation of Internal Forces (SF and BM):

SF(A) = 5 + S2 u3 + S2 u6 = 5 + 3,000(7.74104) + 3,000(2.98104) = 3.54 k
SF(B) (in AB) = 5  S2 u3  S2 u6 = 5  3,000(7.74104)  3,000(2.98104) = 6.46 k
SF(B) (in BC) = 25 + 0 + S2 u6 = 25 + 3,000(2.98104) = 25.89 k
SF(C) (in BC) = 5 + 0  S2 u6 = 5  3,000(2.98104) = 4.11 k should be zero

BM(A) = 12.5 + S3 u3 + S4 u6 = 12.5 + 20,000(7.74104) + 10,000(2.98104) = 0

BM(B) (in AB) = 12.5+ S4 u3 + S3 u6 =12.5+ 10,000(7.74104)+ 20,000(2.98104) = 14.28 k
BM(B) (in BC) = 8.33 + 0 + S3 u6 = 8.33 + 20,000(2.98104) = 14.29 k
BM(C) = 8.33 + 0 + S4 u6 = 8.33 + 10,000(2.98104) = 5.35 k should be equal

11
 Problems on Stiffness Method for Beams/Frames

1. 1 k/ Support A settles 0.05

A Neglect axial deformations and assume
EI = 40,000 k-ft2
10 20
5k
B
D
2. 5k 1 k/ 3. A C
A 20 20 5

A and B are guided roller supports; EIAB = 2 EI

10 20
4. B in an Internal Hinge 5. 1 k/
1 k/ 10 k EIDE = 2 EI

A C D E
B
10 5 5 5 15 10 20

6. Assemble the stiffness matrix, load vector and calculate the unknown joint deflections and rotations of
the beam ABC shown below, considering flexural and axial deformations as well as boundary conditions
[Given: P = 250 k, w = 1 k/ft, F = 10 k, E = 400  103 k/ft2].
w F
A B C 15
P P/2
15 15
10
A and C are Guided Rollers Beam Section

7. 100 k 10 k 8. 1 k/

Guided Roller
8
10

Support A settles 0.05 Joint A rotates 1º anticlockwise

6 10
5 5

9. Use the Stiffness Method (considering flexural deformations only) to calculate the unknown joint
deflections and rotations of the frame loaded as shown below [Given: EI = constant = 10  103 kN-m2].

a b and k are Guided Rollers j

d g
5m
50 kN/m

100 kN
b e h k

5m
f i

c l

5m 5m 5m

12
 Analysis of Three-Dimensional Trusses and Frames

1. Three-Dimensional Trusses
Three-dimensional trusses have 3 unknown displacements at each joint; i.e., the deflection u along x-axis,
deflection v along y-axis and deflection w along z-axis. Therefore the size of the member stiffness matrix is
(66). If Sx = EA/L, then the stiffness matrix in the local axes system is

1 0 0 1 0 0
0 0 0 0 0 0
KmL = Sx 0 0 0 0 0 0
1 0 0 1 0 0
0 0 0 0 0 0
0 0 0 0 0 0

The member stiffness matrix in the global axes system is

Cx2 CxCy CxCz Cx 2 CxCy CxCz

CyCx Cy 2 CyCz CyCx Cy2 CyCz
KmG = Sx CzCx C zCy Cz2 CzCx CzCy Cz2
Cx 2 CxCy CxCz Cx 2 CxCy CxCz
CyCx Cy2 CyCz CyCx Cy2 CyCz
CzCx CzCy Cz2 CzCx CzCy Cz2

where Cx = cos , Cy = cos , Cz = cos 

[,  and  are the angles the member makes with the coordinate axes x, y and z respectively]
After assembling the stiffness matrix and load vector and applying known boundary conditions, the
unknown displacements are calculated by any standard method of solving simultaneous equations.
Once the displacements are known, the member forces are calculated by the following equation
PAB = Sx [(uBuA) Cx + (vBvA) Cy + (wBwA) Cz]

2. Three-Dimensional Frames
Three-dimensional frames have 6 unknown displacements at each joint; i.e., the deflections (u, v, w) along
the x, y and z-axis and rotations (x, y, z) around the x, y and z-axis. Therefore the size of the member
stiffness matrix is (1212), which has the following form in the local axes system

Sx Sx
S1z S2z S1z S2z
S1y S2y S1y S2y
Tx Tx
S2y S3y S2y S4y
S2z S3z S2z S4z
KmL = Sx Sx
S1z S2z S1z S2z
S1y S2y S1y S2y
Tx Tx
S2y S4y S2y S3y
S2z S4z S2z S3z

The transformation matrix Tm and the transformed stiffness matrix KmG in the global axes system are
complicated and not written here. However, the method of applying boundary conditions and solving for the
unknown displacements are similar to the methods mentioned earlier.

13
 Assembly of Stiffness Matrix and Load Vector of a Three-Dimensional Truss
Assemble the global stiffness matrix and write the global load vector of the three dimensional truss shown
below. Also write the boundary conditions [EA/L = Constant = 500 kip/ft].

10 u8
C
y u7
u11 u9 u5
D B
10k u10 u4
u12 u6
x
10 20k
z
u2
A
u1
10 10 u3

Member DB: (Cx = 1, Cy = 0, Cz = 0) Member DC: (Cx = 0, Cy = 0, Cz = −1)

10 11 12 4 5 6 10 11 12 7 8 9
1 0 0 1 0 0 10 0 0 0 0 0 0 10
0 0 0 0 0 0 11 0 0 0 0 0 0 11
KDB = 500 0
G 0 0 0 0 0 12 KDCG = 500 0 0 1 0 0 1 12
1 0 0 1 0 0 4 0 0 0 0 0 0 7
0 0 0 0 0 0 5 0 0 0 0 0 0 8
0 0 0 0 0 0 6 0 0 1 0 0 1 9

Member DA: (Cx = −0.707, Cy = −0.707, Cz = 0)

10 11 12 1 2 3
0.5 0.5 0 0.5 0.5 0 10
0.5 0.5 0 0.5 0.5 0 11
KDAG = 500 0 0 0 0 0 0 12
0.5 0.5 0 0.5 0.5 0 1
0.5 0.5 0 0.5 0.5 0 2
0 0 0 0 0 0 3

1 2 3 4 5 6 7 8 9 10 11 12
0.5 0.5 0 −0.5 −0.5 0 1 XA
0.5 0.5 0 −0.5 −0.5 0 2 YA
0 0 0 0 0 0 3 ZA
1 0 0 −1 0 0 4 XB
0 0 0 0 0 0 5 YB
0 0 0 0 0 0 6 ZB
KG = 500 pG =
0 0 0 0 0 0 7 XC
0 0 0 0 0 0 8 YC
0 0 1 0 0 −1 9 ZC
−0.5 −0.5 0 −1 0 0 0 0 0 1+0+0.5 0+0+0.5 0+0+0 10 10
−0.5 −0.5 0 0 0 0 0 0 0 0+0+0.5 0+0+0.5 0+0+0 11 −20
0 0 0 0 0 0 0 0 −1 0+0+0 0+0+0 0+1+0 12 0

Boundary Conditions: u1 = 0, u2 = 0, u3 = 0, u4 = 0, u5 = 0, u6 = 0, u7 = 0, u8 = 0, u9 = 0

14
Applying boundary conditions

1.5 0.5 0 u10 10

500 0.5 0.5 0 u11 = −20
0 0 1 u12 0

Solving the three equations  u10 = 0.06, u11 = −0.14, u12 = 0

Support Reactions
XA = 250 u1 + 250 u2 − 250 u10 − 250 u11 = 20 k
YA = 250 u1 + 250 u2 − 250 u10 − 250 u11 = 20 k
ZA = 0
XB = 500 u4 − 500 u10 = −30 k
YB = 0
ZB = 0
XC = 0
YC = 0
ZC = 500 u9 − 500 u12 = 0

Member Forces
FDA = 500 {−0.707 (u1 − u10) −0.707 (u2 − u11) + 0 (u3 − u12)} = −28.28 k
FDB = 500 {1 (u4 − u10) + 0 (u5 − u11) + 0 (u6 − u12)} = −30 k
FDC = 500 {0 (u7 − u10) + 0 (u8 − u11) − 1 (u9 − u12)} = 0

15
 Problems on the Analysis of Three-Dimensional Trusses
Apply boundary conditions and ignore zero-force members whenever necessary/convenient
[Assume Sx = constant = 500 k/ft, if not mentioned otherwise]
1. Calculate the joint deflections, support reactions and member forces of the space truss analyzed in class
if support A settles 0.10 vertically downwards.

2. Calculate the member forces of the space truss abcd loaded as shown below, if Px = 0, Py = 10 k, Pz = 0.
Py z Hinge Support
a Px d

20 Pz x 20
c, d b
y b Px
a

20 40 20

x
c
3. Calculate the member forces and applied loads Px, Py, Pz in the space truss abcd shown in Question 2, if
the joint a moves 0.10 right wards and 0.05 downwards due to the applied loads (i.e., no displacement
in z-direction).

4. Ignore zero-force members to form the stiffness matrix, load vector and write down the boundary
conditions of the truss abcdefg shown below [Given: Sx = constant = 500 k/ft].
y y
a
Nodal Coordinates (ft)
d a (0, 30, 0), b (0, 0, 20) , c (0, 0, 20), d (50, 20, 0),
10 k
c e (50, 0, 20), f (50, 0, 20), g (70, 0, 0)
f
30
20
20 20 g
50 x
o o 20
20 20
b e
z
z
5. Ignore zero-force members to form the stiffness matrix, load vector and write down the boundary
conditions of the 3D truss abcdef shown below [Given: Sx = constant = 5000 kN/m].
c
y

50 kN Nodal Coordinates (m)

b a (0, 9, 6), b (0, 9, 0), c (0, 9, 6)
a d (0, 0, 6) , e (0, 0, 6), f (15, 0, 0)
e
9m
6m f
x
o 15m
6m
d

16
 Stiffness Method for Grids
E

E y
5 EI = 4010 k-ft3 2 10 k
D GJ = 30103 k-ft2
10 k x
5 D
A C
A C
B B z

5 5

Top View Isometric View

doki = 3 (u1 = vB, u2 = xB, u3 = zB)

5k
K1 K2
K2 K4 T2
12.5 k
K2
5k K1 K3 T2
S1 S 2 S2 S1 K2
12.5 k S4 S3 S3 S4

S2 S1 S1 S2 T1 T1 T1 T1 S2 S2 S2 S2

Case 0 (FER) Case 1 (u1 = 1) Case 2 (u2 = 1) Case 3 (u3 = 1)

Here, S1 = 1240103/53 = 3840 k/ft S2 = 640103/52 = 9600 k/rad

S3 = 440103/5 = 32000 k-ft/rad S4 = 240103/5 = 16000 k-ft/rad
K1 = 1240103/103 = 480 k/ft K2 = 640103/102 = 2400 k/rad
K3 = 440103/10 = 16000 k-ft/rad K4 = 240103/10 = 8000 k-ft/rad
T1 = 30103/5 = 6000 k-ft/rad T2 = 30103/10 = 3000 k-ft/rad

Fy(B) = 0  5 + (2S1+K1) u1 + K2 u2 + (S2S2) u3 = 10  8160 u1 + 2400 u2 + 0 = 15

Mx(B) = 0  12.5 + K2 u1 + (K3+2T1) u2 + 0 = 0  2400 u1 + 28000 u2 + 0 = 12.5
Mz(B) = 0  0 + (S2S2) u1 + 0 + (2S3+T2) u3 = 0  0 + 0 + 67000 u3 = 0

 u1 = 1.751103 ft u2 = 0.296103 rad u3 = 0

SF(A) = 0  S1 u1 + 0 + S2 u3 = 6.72 k SF(B)/AB = 0 + S1 u1 + 0  S2 u3 = 6.72 k

SF(B)/BC = 0 + S1 u1 + 0 + S2 u3 = 6.72 k SF(C) = 0  S1 u1 + 0  S2 u3 = 6.72 k
SF(B)/BE = 5 + K1 u1 + K2 u2 + 0 = 3.44 k SF(E)/BE = 5  K1 u1  K2 u2 + 0 = 6.55 k

BM(A) = 0  S2 u1 + 0 + S4 u3 = 16.81 k BM(B)/AB = 0  S2 u1 + 0 + S3 u3 = 16.81 k,

BM(B)/BC = 0 + S2 u1 + 0 + S3 u3 = 16.81 k BM(C) = 0 + S2 u1 + 0 + S4 u3 = 16.81 k,
BM(B)/BE = 12.5 + K2 u1 + K3 u2 + 0 = 3.56 k BM(E)/BE = 12.5 + K2 u1 + K4 u2 + 0 = 19.07 k

T(A) = 0 + 0  T1 u2 + 0 = 1.78 k T(B)/AB = 0 + 0 + T1 u2 + 0 = 1.78 k

T(B)/BC = 0 + 0 + T1 u2 + 0 = 1.78 k T(C) = 0 + 0  T1 u2 + 0 = 1.78 k
T(B)/BE = 0 + 0 + 0 + T2 u3 = 0 T(E)/BE = 0 + 0 + 0  T2 u3 = 0

17
 Problems on Stiffness Method for Grids

Given: EI = 40103 k-ft2, GJ = 30103 k-ft2 for all questions.

1,2. Use the Stiffness Method to calculate the rotations at joint C for the grids shown below.
A A

15 10 10 15

10 10
1 k/ft
D B C
5 B 5 D

E C E
10 k
Question 1 Question 2

3. Using the Stiffness Method, calculate the deflection and rotation at joint C for the grid shown in the
figure below.

15
10 10
B C D
5
E
10 k

4,5. Formulate the stiffness matrix and load vector for the grids shown in the figures below.

A
1 k/ft
15 10
10 k
B C D

10 10 5 5

Question 4 Question 5

18
 Calculation of Degree of Kinematic Indeterminacy (Doki)
Determine the doki (i.e., size of the stiffness matrix) for the structures shown below, considering boundary
conditions. For the frames, also determine the doki if axial deformations are neglected.

3D Truss 3D Frame

3D Frame 3D Truss

2D Frame 2D Frame 3D Frame

2D Frame 3D Frame

2D Frame 3D Frame

19
 Energy Formulation of Geometric Nonlinearity
Linear structural analysis is based on the assumption of small deformations and linear elastic behavior of
materials. The analysis is performed on the initial undeformed shape of the structure. As the applied loads
increase, this assumption is no longer accurate, because the deformations may cause significant changes in
the structural shape. Geometric nonlinearity is the change in the elastic load-deformation characteristics of
the structure caused by the change in the structural shape.
Among various types of geometric nonlinearity, the structural instability or moment magnification caused by
large compressive forces, stiffening of structures due to large tensile forces, change in structural parameters
due to applied dynamic loads are significant. Rather than using equilibrium equations, it is often more
convenient to formulate geometrically nonlinear problems by the Method of Virtual Work.

Method of Virtual Work

Another way of representing Newton’s equation of equilibrium is by energy methods, which is based on the
law of conservation of energy. According to the principle of virtual work, if a system in equilibrium is
subjected to virtual displacements u, the virtual work done by the external forces (WE) is equal to the
virtual work done by the internal forces (WI)
WI = WE ...…………………(1)
where the symbol  is used to indicate ‘virtual’. This term is used to indicate hypothetical increments of
displacements and works that are assumed to happen in order to formulate the problem.

Energy Formulation and Buckling of Beams-columns

u(x) = u0 (x)

q(x) u0
P P
A B
L

Transversely Loaded Member and Assumed Shape

Applying the method of virtual work to flexural members subjected to transverse load of q(x) per unit length
and axial (tensile) force P   uE I u dx +  u P u dx =  q(x) dx u ..…...……………(2)
Using the energy formulation assuming u(x) = u0(x) provides the following equation
 u0 (x) EI u0 (x) dx +  u0 (x) P u0 (x) dx =  q(x) dx u0 (x)
 { EI [(x)]2 dx +  P [(x)]2 dx} u0 =  q(x) (x) dx ……….………….(3)
Carrying out the integrations after knowing (or assuming) (x), Eq. (3) can be rewritten as,
kTotal* u0 = f* ……………….…...(4)
where k*, f*are the ‘effective’ stiffness and force of the system, with
kTotal* =  EI [(x)]2 dx +  P [(x)]2 dx ……...…………..(5.1)
f* =  q(x) (x) dx ……………….…(5.2)
Therefore a tensile force P (i.e., positive P) will further stiffen the beam-column (i.e., increase its stiffness)
and a compressive force (i.e., negative P) will make it more flexible and increase the resulting deflection and
internal forces compared to linear analysis. In the extreme case, the beam-column will buckle if the effective
stiffness k* becomes zero, which is possible only for a compressive force
Pcr = −{EI [(x)]2 dx}/ [(x)]2 dx ……………….…...(6)
It is obvious that the accuracy of the formulation depends on the accuracy of the assumed shape function
(x), which must at least satisfy the natural boundary conditions. However, other than assuming a more
appropriate shape function, its accuracy cannot be improved by any other means.

20
Stiffness Matrix and Geometric Stiffness Matrix of Beams-columns

u1 u3
u2 q(x) u4
P P
A B
L
Transversely Loaded Beam-Column
Two-noded elements with cubic interpolation functions for u1, u2, u3 and u4 are typically chosen in such
cases, so that u(x) = u1 1 + u2 2 + u3 3 + u4 4 ……………(7)
where 1(x) = 13(x/L)2 +2(x/L)3, 2(x) = x{1 (x/L)}2
3(x) = 3(x/L)2 2(x/L)3, 4(x) = (xL)(x/L)2 ……………(8)
u= u1 1 + u2 2 + u3 3 + u4 4; u= u1 1 + u2 2 + u3 3 + u4 4 .……….…..(9)
u= u1 1 + u2 2 + u3 3 + u4 4; u= u1 1 + u2 2 + u3 3 + u4 4 .…...…..…(10)

1(x)
2(x)

x x

3(x)
x 4(x)
x

Shape functions 1(x), 2(x), 3(x) and 4(x)

Inserting the values of u, u, u and u in Eq. (1), and equating the coefficients of 1 
(E I 1 1 dx + P 1 1 dx) u1 + (E I 1 2 dx + P 1 2 dx) u2
+ (E I 1 3 dx + P 1 3 dx) u3 + (E I 1 4 dx + P 1 4 dx) u4 = q(x) 1 dx …...…...(11)

Similarly, equating the coefficients of 2, 3 and 4 will produce two (44) matrices Km and Gm, along with
a (41) load vector pm here, and their elements are given by
Kmij = EI i j dx Gmij =  P ij dx pmi =  q(x) i dx ……..…(12)
The equations of the stiffness matrix and geometric stiffness matrix for flexural members guarantee that for
‘linear’ problems,
(i) The stiffness and geometric stiffness matrices are symmetric [i.e., Element (i,j) = Element (j,i)],
(ii) The diagonal elements of the matrices are positive [as the element (i,i) involves squares].

As mentioned, for structural analysis the effect of axial load on flexural behavior can be approximated by
simplified formulations of the geometric nonlinearity problem. For this purpose, a new matrix called the
geometric stiffness matrix (G) has been added to the original stiffness matrix K obtained from linear
analysis of the undeformed deflected shape of the structure. Therefore, the total stiffness matrix of a flexural
member is the sum of these two matrices; i.e.,
Ktotal = K + G ……………………..(13)
Using the same shape functions i (i = 1~4) as done for the linear analyses of beams and frames, the
following geometric stiffness matrix is formed in the local axes system of a member of length L.

21
 36 3L 36 3L
GmL = (P/30L) 3L 4L2 3L L2 ……….....(14)
36 3L 36 3L
3L L2 3L 4L2

This geometric stiffness matrix can be added to the linear stiffness matrix shown before, and the total
stiffness matrix is transformed and assembled using the equations and formulations mentioned in earlier
lectures. Once the total stiffness matrix Ktotal is obtained in the global axes after applying appropriate
boundary conditions, the structural analyses can be carried out using the procedures mentioned before.

The governing equations of motion can be written in matrix form as

Ktotal u = f ……………….….(15)
However it should be noted that the presence of axial force P in the geometric stiffness matrix makes the
problem nonlinear because P is obtained from member deformations, which cannot be found before
performing the structural analysis. Therefore the system properties and output are interdependent, which
calls for iterative methods of structural analysis. However, P is known for special cases (e.g., single column
subjected to a known axial load) so that the problem is not nonlinear any more.

Buckling occurs when the structure loses its stiffness, i.e., when the total stiffness matrix Ktotal becomes
singular. Therefore, the buckling load can be obtained by solving the eigenvalue problem
Ktotal= 0  K+G= 0 ………………………..(16)
Since the stiffness and geometric stiffness matrix are derived from approximate shape functions, the critical
buckling load obtained from Eq. (16) is also approximate and can be improved if the beam-column is
divided into more segments throughout its length.

Just as an axial compressive load can reduce the effective stiffness of a structural member, a tensile load
may increase it. This will cause stiffening of the member and a corresponding decrease in deformations.

22
Example
For EI = 40103 k-ft2, L = 10 ft, calculate the approximate first buckling load for
(i) a simply supported beam, (ii) a cantilever beam
(iii) Suggest how to improve the results.
(iv) Also calculate the tip deflection and rotation of the cantilever beam when subjected to a uniformly
distributed transverse load of 1 k/ft along with a compressive load of 400 kips.
Solution
(i) For the simply supported beam, the two d.o.f. are A and B, so the K and G matrices are
16 8 400 100
K = 103 and G = (P/300) A B
8 16 100 400
For critical buckling load P, the determinant of (K + G) is = 0
 (16000 + 4P/3)2 (8000  P/3)2 = 0  16000 + 4P/3 =  (8000  P/3)
 5P/3 = 8000; i.e., P = 4800 k, or P + 24000 = 0; i.e., P = 24000 k (negative  compression)
Compared to the first two ‘exact’ buckling loads, 2EI/L2 and 42EI/L2 ; i.e., 3948 k and 15791 k
(ii) For the cantilever beam, the two d.o.f. are vA and A, the K and G matrices being
0.48 2.4 36 30
K = 103 and G = (P/300) A B
2.4 16 30 400
For critical buckling load P, the determinant of (K + G) is = 0
 (480 + 0.12 P) (16000 + 4 P/3)  (2400 + 0.10 P)2 = 0  0.15 P2 + 2080 P +192104 = 0
 P = [2080  {(2080)2 40.15192104}]/0.30 = 994 k and 12872 k
Compared to the first two ‘exact’ buckling loads, 2EI/(2L)2 and 92EI/(2L)2; i.e. 987 k, 8883 k
(iii) The predictions can be improved by dividing the beams into more segments or using more appropriate
shape functions. For the simply supported beam
Dividing into two segments, half of the symmetric beam takes the form
For the simply supported beam, the two d.o.f. are A and vC, so the K and G matrices are
32 9.6 100 15 A C
K = 103 and G = (P/150)
9.6 3.84 15 36
For critical buckling load P, the determinant of (K + G) is = 0
 (32000 + 2P/3) (3840 + 0.24P) (9600 P/10)2 = 0  0.15 P2 + 8320 P + 30.72 × 106 = 0
 P = 3978 k, or 51489 k (negative  compression)
They only represent the first two ‘odd’ buckling loads, 2EI/L2 and 92EI/L2; i.e., 3948 k, 35531 k
Assuming (x) = sin(x/L) [with (0) = (L) = 0, (x) = (/L) cos(x/L), (x)= −(/L)2 sin(x/L)]
Effective stiffness k* =  EI [(x)]2 dx = (/L)4 EI L/2
Effective geometric stiffness g* =  P [(x)]2 dx = (/L)2 P L/2
kTotal*= k* + g* = 0  Buckling load Pcr = −{(/L)4 EI L/2}/{(/L)2 L/2}= −2 EI//L2 = −3948 k, which is
the exact first buckling load.
(iv) If P = 400 kips for the cantilever beam, the total stiffness matrix Ktotal and load vector f are
480  48 2400  40 432 2360 5.00
Ktotal = = f=
2400  40 16000  533.33 2360 15466.67 8.33
3 3
Solving the two equations, vA = 51.86 10 ft, and A = 7.374 10 rad
(compared to the results when P = 0, i.e., vA = 31.25 103 ft, and A = 4.167 103 rad)
Assuming (x) = 1sin(x/2L), (x) = (/2L) cos(x/2L)
Effective stiffness k* =  EI [(x)]2 dx = (/2L)4 EI L/2 = 121.76 k/ft
Effective geometric stiffness g* =  P [(x)]2 dx = (/2L)2 P L/2 = 49.35 k/ft
Effective force f* =  q(x) (x) dx = qL(1 2/) = 3.63 kips
(121.76  49.35) u2 = 3.63  u2 = 50.18  103 ft
 vA = u2 (0) = u2 = 50.18  103 ft, and A = u2 (0) = u2 (/2L) = 7.882 103 rad

23
 Assembling Stiffness Matrix and Geometric Stiffness Matrix

Assume EI = 40103 k-ft2 for the following problems

1.

A B C
P

20 20 Guided Roller

Stiffness Matrix K = 8000 + 8000 = 16000

Geometric Stiffness Matrix G =  (P/600) 1600 + 1600 =  (P/600) 3200

2.
A B C
P

20 20 Roller

8000 + 8000 4000 16000 4000

Stiffness Matrix K = =
4000 8000 4000 8000

1600 + 1600 400 3.2 0.4

Geometric Stiffness Matrix G =  (P/600) =  (P/0.6)
400 1600 0.4 1.6

3.
A B C
P

20 20 Slope = 0, Deflections 0

8000 + 8000 600 16000 600

Stiffness Matrix K = =
600 60 600 60

1600 + 1600 60 3200 60

Geometric Stiffness Matrix G =  (P/600) =  (P/600)
60 36 60 36

24
 Practice Problems on Geometrically Nonlinear Structures
1. Use Stiffness Method to calculate the unknown joint deflections and rotations of beam ABC loaded as
shown below, considering flexural deformations only with geometric nonlinearity [E = 400  103 k/ft2].
A and B are Guided Rollers 10 k 10 k
B 100 k-ft
A C 24
500 k 250 k 12
20 10 15 15
Beam Section

2. Calculate the value of P needed to cause buckling of the beam abcd shown below
[Given: EIabc = 20  103 k-ft2, EIcd = 2EIabc].
b P P c
a d

10 20 20

3. Use Stiffness Method to calculate the force P needed to cause buckling of frame abcde shown below,
considering flexural deformations only with geometric nonlinearity [EI = constant = 10  103 kN-m2].
50 kN 10 kN/m

P
a b
c

10m

a and c are Guided Rollers

d e
5m 5m 10m

4. Calculate the force P to cause buckling of the frame shown below, using w = 0.15P [EI = 40  103 k-ft2].

P w k/ft P
C D
B E

10

A F
5 10 10 5

5. Use the Stiffness Method (considering geometric nonlinearity) to calculate the joint deflections and
rotations of the frame loaded as shown below [Given: EI = constant = 50  103 k-ft2].

B C

A, B, C are Guided Rollers

10 5 k/ 10 k

A
50 k
20

25
 Material Nonlinearity and Plastic Moment
As mentioned in the previous section, structural properties cannot be assumed to remain constant in many
practical situations. In addition to the geometric nonlinearity that may lead to instability of structures with
linearly materials properties, the variation in material properties itself can make the structural analysis
nonlinear. For example, yielding of the structural materials, a likely situation in a severe loading conditions
or ground vibrations, may alter the stiffness properties, which needs to be updated with structural
deformations.

Material Nonlinearity in Concrete, Steel and Reinforced Concrete

Concrete and steel are the most common among the construction materials used for Civil Engineering
constructions. Among them, concrete is much stronger in compression than in tension (tensile strength is of
the order of one-tenth of compressive strength). While its tensile stress-strain relationship is almost linear,
the stress-strain relationship in compression is nonlinear from the beginning (Fig. 1).

Steel on the other hand, has similar stress-strain properties in tension and compression. After an initial
linearly-elastic portion, the stress remains almost constant while the strain increases significantly (a
phenomenon called yielding). This is typically followed by some increase in stress (strain hardening) at a
reducing elasticity, and finally a decrease in stress leading to breaking of the specimen (Fig. 2).

Reinforced Concrete or RC is a unique combination of these two materials where the complexities of their
constitutive behavior come into effect. The behavior of RC cannot be modeled properly by linear elastic
behavior. Recognizing this, the design of RC structures has gradually shifted over the years from the
‘elastic’ Working Stress Design (WSD) to the more rational Ultimate Strength Design (USD). The design of
steel structures has also undergone similar transition from the Allowable Stress Design (ASD) method to the
Load and Resistant Factor Design (LRFD) method.

Parabola Typical  vs.  fult

fbrk
fc Straight Line fy

fc fs E Elastic-Perfectly Plastic Model

0 0 u
 
Fig. 1: Stress-Strain Model for Concrete (Compression) Fig. 2: Typical Stress vs. Strain for Steel

Analysis of Linearly Elastic and Inelastic Systems

For a linearly elastic system the relationship between the applied force fs and the resulting deformation u is
linear, i.e.,
fs = k u ……………………(1)
where k is the linear stiffness of the system; its units are force/length. Implicit in Eq. (1) is the assumption
that the linear fs-u relationship determined for small deformations of structure is also valid for large
deformations. Because the resisting force is a single valued function of u, the system is elastic; hence the
term k can be used in linearly elastic system. This is however not valid when the load-deformation
relationship is nonlinear, i.e., when the stiffness itself is not constant but is a function of u. Thus the resisting
force can be expressed as
fs = fs(u) .……………………....(2)
and the system is called inelastic dynamic system. The structural analysis of such systems can only be
performed by iterative methods.

26
Plastic Moment of Typical Sections
The iterative method required to analyze nonlinear systems is quite laborious, time consuming and its
convergence to the exact solution is not always guaranteed, it is usually not followed in typical structural
analyses other than for very important projects. However, the calculation of the ultimate moment capacity of
a cross-section or the ultimate load carrying capacity of a structure is usually much simpler, and is of more
interest to a structural designer.

The following examples show the calculation of yielding and ultimate moment capacities of typical steel and
RC sections.

Example 1
Calculate the Yield Moment and Plastic Moment capacity of the sections shown below if they are made of
elastic-fully plastic material (e.g., steel model shown in Fig. 2).

b
For the rectangular section, the neutral axis divides the area into two segments of (b  h/2)
Compressive force = Tensile force = yp (bh/2)
h
Plastic moment Mp = Tensile (or compressive) force  Moment arm = yp (bh/2)  h/2
Mp = yp (bh2/4)
The yield moment is My = yp (S) = yp (bh3/12)/(h/2) = yp (bh2/6) 2

For the T-section, the equal-area axis divides the area along the flange line.
Compressive force = Tensile force = yp (12  2) = 24yp 12
Plastic moment Mp = Tensile (or compressive) force  Moment arm = 24yp  (1 + 6)
Mp = yp (168) = 6048 k-in = 504 k-ft [assuming yp = 36 ksi]
2
Also,y = (24  1 + 24  8)/48 = 4.5; c = 14 − 4.5 = 9.5
12
I = 12  23/12 + 24 (1 − 4.5)2 + 2  123/12 + 24 (8 − 4.5)2 = 884 in4
S = 884/9.5 = 93.05 in3  My = yp (93.05) = 279.15 k-ft 5″/16

For the I-section, the equal-area axis divides the area symmetrically.
Compressive force = Tensile force = yp {6  5/16 + (6 −5/16)  0.25} = 3.297yp 0.25
12
Plastic moment Mp = yp {1.875  (6 −5/32) + 1.422  (6 −5/16)/2}  2 = 30yp
Mp = yp (30) = 1800 k-in = 150 k-ft [assuming yp = 60 ksi]

6
Example 2
Calculate the Ultimate Moment capacity of the rectangular RC beam section shown below
[Given: fc = 4 ksi, fy = 60 ksi].

15.5
Steel Bars
2.5 (1 diameter)
12

For b = 12, d = 15.5, As = 3 (1)2/4 = 2.36 in2

 a = Asfy/(0.85f c b) = 2.36  60/(0.85  4  12) = 3.46
Mult = Asfy (d − a/2) = 2.36  60 (15.5 − 3.46/2) = 1946 k-in = 162.2 k-ft

27
 Ultimate Load of Simple Beams
Plastic Hinge and Ultimate Load
Since Plastic Moment of a section is its ultimate moment capacity, it cannot take any more moment beyond
this. As such, the section behaves almost like an internal hinge within a structure. Such a hypothetical
internal hinge is called Plastic Hinge; and by adding a new equation of statics, it reduces by one the degree
of statical indeterminacy of the structure. Therefore, formation of such hinges can make the structure
statically determinate, and eventually lead to its instability, which can cause the ultimate collapse of the
structure, at the formation of Collapse Mechanism.
By calculating the external loads necessary to form such hinges, it is possible to calculate the loads needed
to form Collapse Mechanism of the structure. This load is called the Ultimate Load of the structure and is
important to a designer because it provides information about the load that the structure can possibly sustain,
as demonstrated by the following examples.
Example 3
Calculate the ultimate load capacity of the simply supported beams loaded as shown below
[Given: Plastic Moment (Mp) of the section = 150 k-ft, as calculated for the I-section in Example 1].
P w per unit length

L/2 L/2 L
P1 L/4 = Mp w1 L2/8= Mp

BMD BMD

-Shape
-Shape

When P = P1, Plastic Hinge forms at the When w = w1, Plastic Hinge forms again at
midspan of the beam at a moment P1L/4. the midspan at a moment of w1L2/8.
P1L/4 = Mp  P1 = 4Mp/L w1L2/8 = Mp  w1 = 8Mp/L2
L = 25 and Mp= 150 k L = 25 and Mp= 150 k
 Pult = P1 = 4  150/25 = 24 k  wult = w1 = 8  150/252 = 1.92 k/ft

Example 4
Calculate the ultimate load capacity of the fixed-ended beams loaded as shown below.
P w per unit length

L/2 L/2 L

P1 L/8
wL2/24

P1 L/8 P1 L/8 wL2/12 wL2/12

When P = P1, Plastic Hinges form at both ends When w = w1, the first Plastic Hinges form at both
and midspan of the beam at moments of P1L/8. ends at moments of w1L2/12.
P1L/8 = Mp  P1 = 8Mp/L, when a Collapse w1L2/12 = Mp  w1 = 12Mp/L2
Mechanism is formed But a Collapse Mechanism is not formed until
L = 25, Mp= 150 k another hinge forms at midspan at a load w = w2;
 Pult = P1 = 8  150/25 = 48 k i.e., when w2L2/8 − Mp = Mp w2 = 16Mp/L2
L = 25, Mp= 150 k  wult = w2 = 3.84 k/ft

28
 Energy Formulation of Collapse Mechanism
The calculation of ultimate load capacity based on bending moment diagrams demonstrates the actual
sequence of plastic hinge formulation in a structure leading to its ultimate failure. However, it requires the
bending moment diagram after each hinge formation, which may not always be convenient to form. A more
direct (though not as detailed) calculation of the ultimate load capacity is possible by using the virtual work
method on assumed collapse mechanisms of structures. As mentioned in previous formulations, if a system
in equilibrium is subjected to virtual displacements u, the virtual work done by the external forces (WE) is
equal to the virtual work done by the internal forces (WI); i.e., WE = WI
Example 5
Use Energy Formulation to calculate the ultimate load capacity of the simply supported beams shown below.
P w per unit length

L/2 L/2 L

1-Shape -Shape
     
   

2-Shape x (Lx)
   
 
x
External work done = wL /2
(Lx)
Internal work done = Mp(+)
For the deflected shape 1 wL /2 = Mp( + ) = Mp{/x + /(Lx)}
External work done = P   w = (2Mp/L){1/x + 1/(Lx)}
Internal work done = Mp( +) = 2Mp w/x = 0  1/x2 + 1/(Lx)2 = 0  x = L/2
P  = 2Mp  = 2Mp{/(L/2)}  wult = wmin = (2Mp/L2) (2 +2) = 8Mp/L2
 P = 4Mp/L
For the deflected shape 2
External work done = P  = P L/2
Internal work done = Mp(+)
P L/2 = Mp(+)
 P{/(Lx)}L/2 = Mp{/x + /(Lx)}
 P = 2Mp/L{(L/x 1) +1}= 2Mp/x
Pmin = 2Mp/(L/2) = 4Mp/L

Example 6
Use Energy Formulation to calculate the ultimate load capacity of the beams shown below.
P w per unit length

L/2 L/2
x (Lx)
-Shape
     
   

External work done = P  External work done = wL /2

Internal work done = Mp(2) + Mp + Mp Internal work done = Mp(+) + Mp = Mp(+2)
= 4Mp wL /2 = Mp(+2) = Mp{/x + 2/(Lx)}
P  = 4Mp  = 2Mp{/(L/2)}  w = (2Mp/L){1/x + 2/(Lx)}
 Pult = 8Mp/L w/x = 0  1/x2 + 2/(Lx)2 = 0  x = L/(2+1)
 wult = wmin = (2Mp/L2){2+1+2+2} = 11.66Mp/L2

29
 Ultimate Load of Continuous Beams, Frames

Example 7
Use the Energy Method to calculate the plastic moment Mp needed to prevent the development of plastic
hinge mechanism in the beam ABCD loaded as shown below [Given: Mp(AB) = Mp (BC) = Mp, Mp(CD) = 2Mp].

5 k/ 50 k

A
B C D
10 15 10 10

For span AB: w = 11.66Mp(AB)/L2  5 = 11.66 Mp/102  Mp = 42.88 k-ft

For span BC: w = 16Mp(BC)/L2  5 = 16Mp/152  Mp = 70.31 k-ft
For span CD: P = 8Mp(CD)/L  50 = 8(2Mp)/20  Mp = 62.50 k-ft
 Mp(req) = 70.31 k-ft
If Mp(req) = 70.31 k-ft, w(all)AB = 11.66  70.31/102 = 8.20 k/ft
w(all)BC = 16  70.31/152 = 5.00 k/ft
P(all)CD = 8  2  70.31/20 = 56.25 k

Example 8
Use the Energy Method to calculate the load (i) w needed to form beam mechanism, (ii) P needed to form
the sidesway mechanism in the frames ABCD loaded as shown below [Given: Mpb  Mpc].
w per length w per length
B B
P C C
P

H H

A D A D

L L

    
   



     
 

For beam mechanism, wult = 16Mpb/L2 For beam mechanism, wult = 16Mpb/L2
For sidesway mechanism, For sidesway mechanism,
P = Mpc + Mpc = 2Mpc = 2Mpc/H P = Mpc + Mpc + Mpc + Mpc
Pult = 2Mpc/H = 4Mpc = 4Mpc/H
Pult = 4Mpc/H

30
 Practice Problems on Material Nonlinearity and Plastic Moment
1. Calculate the distributed load w k/ft needed to develop plastic hinge mechanism of the beam abc loaded
as shown below (by using its bending moment diagram) [Given: yp = 40 ksi].
8
w k/ft
3/8 
c
a
b 0.25 16

5 15
Beam X-section

2. Use the bending moment diagram of beam ABCDE shown below to calculate the Mp (and corresponding
fy) needed to prevent formation of plastic hinge mechanism [Given: Mp(BCD) = Mp, Mp(AB) = Mp(DE) = 2Mp].

C and D are 50 kN 10 kN/m

Internal Hinges
0.5m
C D E
A B
5m 5m 5m 10m 5m 0.3m Beam X-section

3. Answer Question 1 and 2 using the Energy Method of Collapse Mechanism.

4. Use the bending moment diagram to calculate the force P needed to develop plastic hinge mechanism in
the reinforced concrete beam ABCD loaded as shown below [Given: fc = 3 ksi, fy = 50 ksi].

2 #7 bars 3
P P/2
12
A C
D
B 3

10 10 10 2#7 bars 12

3.5P 2P Beam cross-section

3P BMD

5. Use Energy Method to calculate the plastic moment Mp necessary to prevent plastic hinge mechanism
from developing in the beam abcdef loaded as shown below [Given: Mp(abc) = 2 Mp (cdef) = Mp].

2 k/ft 30 k 50 k 1 k/ft
c is an
Internal
Hinge a c
b d e f
30 15 15 20 10 30

6. Use the Energy Method to calculate the plastic moment Mp of the cross-sections necessary to prevent the
development of collapse mechanism in the (i) continuous bridge ABCD, and (ii) balanced cantilever
bridge ABEFCD loaded as shown below.
3 k/ft

E and F are Internal Hinges

A B E F C D

80 20 60 20 80

31
 Dynamic Equations of Motion for Lumped Mass Systems
Formulation of the Single-Degree-of-Freedom (SDOF) Equation
A dynamic system resists external forces by a combination of forces due to its stiffness (spring force),
damping (viscous force) and mass (inertia force). For the system shown in Fig. 1.1, k is the stiffness, c the
viscous damping, m the mass and u(t) is the dynamic displacement due to the time-varying excitation force
f(t). Such systems are called Single-Degree-of-Freedom (SDOF) systems because they have only one
dynamic displacement [u(t) here].

m f(t), u(t) f(t) ma

fS fV
k c

Fig. 1.1: Dynamic SDOF system subjected to dynamic force f(t)

Considering the free body diagram of the system, f(t)  fS  fV = ma …………..(1.1)

where fS = Spring force = Stiffness times the displacement = k u …..………(1.2)
fV = Viscous force = Viscous damping times the velocity = c du/dt …..………(1.3)
fI = Inertia force = Mass times the acceleration = m d2u/dt2 ..…………(1.4)

Combining the equations (1.2)-(1.4) with (1.1), the equation of motion for a SDOF system is derived as,
m d2u/dt2 + c du/dt + ku = f(t) …..………(1.5)
nd
This is a 2 order ordinary differential equation (ODE), which needs to be solved in order to obtain the
dynamic displacement u(t). As will be shown subsequently, this can be done analytically or numerically.

Eq. (1.5) has several limitations; e.g., it is assumed on linear input-output relationship [constant spring (k)
and dashpot (c)]. It is only a special case of the more general equation (1.1), which is an equilibrium
equation and is valid for linear or nonlinear systems. Despite these, Eq. (1.5) has wide applications in
Structural Dynamics. Several important derivations and conclusions in this field have been based on it.

Governing Equation of Motion for Systems under Seismic Vibration

The loads induced by earthquake are not body-forces; rather it is a ground vibration that induces certain
forces in the structure. For the SDOF system subjected to ground displacement ug(t)

m u(t) u(t) ma

k c
fS fV
ug(t) ug(t)
Fig. 1.2: Dynamic SDOF system subjected to ground displacement ug(t)
fS = Spring force = k (uug), fV = Viscous force = c (du/dtdug/dt), fI = Inertia force = m d2u/dt2
Combining the equations, the equation of motion for a SDOF system is derived as,
m d2u/dt2 + c (du/dtdug/dt) + k (uug) = 0  m d2u/dt2 + c du/dt + k u = c dug/dt + k ug ...……(1.6)
 m d2ur/dt2 + c dur/dt + k ur = m d2ug/dt2 ..…..……………(1.7)
where ur = uug is the relative displacement of the SDOF system with respect to the ground displacement.
Eqs. (1.6) and (1.7) show that the ground motion appears on the right side of the equation of motion just like
a time-dependent load. Therefore, although there is no body-force on the system, it is still subjected to
dynamic excitation by the ground displacement.

32
Formulation of the Two-Degrees-of-Freedom (2-DOF) Equation
The simplest extension of the SDOF system is a two-degrees-of-freedom (2-DOF) system, i.e., a system
with two unknown displacements for two masses. The two masses may be connected to each other by
several spring-dashpot systems, which will lead to two differential equations of motion, the solution of
which gives the displacements and internal forces in the system.

m2 f2(t), u2(t)

k2 c2 k2 (u2u1) c2 (v2v1)
m2
m1 f1(t), u1(t) f2(t), u2(t)
f1(t), u1(t)
m1
k 1u 1 c1v1
k2 c2
k1 c1 k1 c1
k2 (u2u1) c2 (v2v1)

Fig. 2.1: Dynamic 2-DOF system and free body diagrams of m1 and m2

Fig. 2.1 shows a 2-DOF dynamic system and the free body diagrams of the two masses m1 and m2. In the
figure, ‘u’ stands for displacement (i.e., u1 and u2) while ‘v’ stands for velocity (v1 and v2). Denoting
accelerations by a1 and a2, the differential equations of motion are formed by applying Newton’s 2nd law of
motion to m1 and m2; i.e.,
m1a1 = f1(t) + k2(u2–u1) + c2(v2–v1) – k1u1 – c1v1
 m1a1 + (c1+c2) v1 + (k1+k2)u1 – c2v2 – k2u2 = f1(t) ……..…(2.1)
and m2a2 = f2(t) – k2 (u2–u1) – c2(v2–v1)  m2a2 – c2v1 + c2v2 – k2u1 + k2u2 = f2(t) ……..…(2.2)
Putting v = du/dt (i.e., v1 = du1/dt, v2 = du2/dt) and a = d2u/dt2 (i.e., a1 = d2u1/dt2, a2 = d2u2/dt2) in Eqs. (2.1)
and (2.2), the following equations are obtained
m1 d2u1/dt2 + (c1+c2) du1/dt – c2 du2/dt + (k1+k2) u1 – k2 u2 = f1(t) ………....(2.3)
m2 d2u2/dt2 – c2 du1/dt + c2 du2/dt – k2 u1 + k2 u2 = f2 (t) …………(2.4)
Eqs. (2.3) and (2.4) can be arranged in matrix form as
m1 0 d2u1/dt2 c1 + c2 –c2 du1/dt k1+k2 –k2 u1 f1(t)
+ + =
2 2
0 m2 d u2/dt –c2 c2 du2/dt –k2 k2 u2 f2(t)
………....(2.5)
Eqs. (2.5) represent in matrix form the set of equations [i.e. (2.3) and (2.4)] to evaluate the displacements
u1(t) and u2(t). In this set, the matrix consisting of the masses (m1 and m2) is called the mass matrix, the one
consisting of the dampings (c1 and c2) is called the damping matrix and the one consisting of the stiffnesses
(k1 and k2) is called the stiffness matrix of this particular system. These matrices are different for various 2-
DOF systems, so that Eq. (2.5) cannot be taken as a general form for any 2-DOF system.
For a lumped 2-DOF system subjected to ground displacement ug(t), velocity vg(t) and acceleration ag(t), the
following equations are obtained in matrix form

m1 0 d2u1/dt2 c1 + c2 c2 du1/dt k1 + k2 k2 u1 c1vg + k1ug

+ + =
0 m2 d2u2/dt2 c2 c2 du2/dt k2 k2 u2 0

…………(2.6)

m1 0 d2u1r/dt2 c1 + c2 c2 du1r/dt k1 + k2 k2 u1r m1 a g

+ + =
0 m2 d2u2r/dt2 c2 c2 du2r/dt k2m2 a g k2 u2r
…………(2.7)
For a MDOF system, Eq. (2.5) can be written in the general form of the dynamic equations of motion,
M d2u/dt2 + C du/dt + K u = f(t) ……….….(2.8)

33
 Numerical Solution of SDOF Equation
The equation of motion for a SDOF system can be solved analytically for different loading functions. Even
if the assumptions of linear structural properties are satisfied; the practical loading situations can be more
complicated and not convenient to solve analytically. Numerical methods must be used in such situations.

The most widely used numerical approach for solving dynamic problems is the Newmark- method.
Actually, it is a set of solution methods with different physical interpretations for different values of . The
total simulation time is divided into a number of intervals (usually of equal duration t) and the unknown
displacement (as well as velocity and acceleration) is solved at each instant of time. The method solves the
dynamic equation of motion in the (i + 1)th time step based on the results of the ith step.

The equation of motion for the (i +1)th time step is

m (d2u/dt2)i+1 + c (du/dt)i+1 + k (u)i+1 = f i+1  m ai+1 + c vi+1 + k ui+1 = f i+1 …..………(3.1)

where ‘a’ stands for the acceleration, ‘v’ for velocity and ‘u’ for displacement.

To solve for the displacement or acceleration at the (i + 1)th time step, the following equations are assumed
for the velocity and displacement at the (i + 1)th step in terms of the values at the ith step.

vi+1 = vi + {(1) ai +  ai+1}t ….…………(3.2)

ui+1 = ui + vi t + {(0.5) ai +  ai+1}t2 …….………(3.3)

By putting the value of vi+1 from Eq. (3.2) and ui+1 from Eq. (3.3) in Eq. (3.1), the only unknown variable ai+1
can be solved from Eq. (3.1).

In the solution set suggested by the Newmark- method, the Constant Average Acceleration (CAA) method
is the most popular because of the stability of its solutions and the simple physical interpretations it
provides. This method assumes the acceleration to remain constant during each small time interval t, and
this constant is assumed to be the average of the accelerations at the two instants of time ti and ti+1. The CAA
is a special case of Newmark- method where  = 0.50 and  = 0.25. Thus in the CAA method, the
equations for velocity and displacement [Eqs. (3.2) and (3.3)] become

vi+1 = vi + (ai + ai+1)t/2 ……………(3.4)

ui+1 = ui + vi t + (ai + ai+1)t2/4 ……………(3.5)

Inserting these values in Eq. (3.1) and rearranging the coefficients, the following equation is obtained,

(m + ct /2 + kt2/4)ai+1 = fi+1 – kui – (c + kt)vi – (ct/2 + kt2/4)ai ….….…..(3.6)

(meff) ai+1 = fi+1 – kui –(ceff) vi – (meff1) ai ….….…..(3.6)

To obtain the acceleration ai+1 at an instant of time ti+1 using Eq. (3.6), the values of ui, vi and ai at the
previous instant ti have to be known (or calculated) before. Once ai+1 is obtained, Eqs. (3.4) and (3.5) can be
used to calculate the velocity vi+1 and displacement ui+1 at time ti+1. All these values can be used to obtain the
results at time ti+2. The method can be used for subsequent time-steps also.

The simulation should start with two initial conditions, like the displacement u0 and velocity v0 at time t0 = 0.
The initial acceleration can be obtained from the equation of motion at time t0 = 0 as

a0 = (f0 – cv0 – ku0)/m ……………(3.7)

34
Example 3.1
For the undamped SDOF system described before (m = 1 k-sec2/ft, k = 25 k/ft, c = 0 k-sec/ft), calculate the
dynamic response for a Ramped Step Loading with p0 = 25 kips and t0 = 0.5 sec [i.e., p(t) = 50 t  25 kips]

Results using the CAA Method (for time interval t = 0.05 sec) as well as the exact analytical equation are
shown below in tabular form.

Table 3.1: Acceleration, Velocity and Displacement for t = 0.05 sec

m (k-sec2/ft) c (k-sec/ft) k (k/ft) t0 (sec) dt (sec) meff (k-sec2/ft) ceff (k-sec/ft) meff1 (k-sec2/ft)
1.00 0.00 25.00 0.50 0.05 1.0156 1.2500 0.0156

i t (sec) fi (kips) ai (ft/sec2) vi (ft/sec) ui (ft) uex (ft) 20.0

0 0.00 0.0 0.0000 0.0000 0.0000 0.0000

Acceleration (ft/sec )
1 0.05 2.5 2.4615 0.0615 0.0015 0.0010

2
10.0
2 0.10 5.0 4.7716 0.2424 0.0091 0.0082
3 0.15 7.5 6.7880 0.5314 0.0285 0.0273 0.0
4 0.20 10.0 8.3867 0.9107 0.0645 0.0634 0.0 0.5 1.0 1.5 2.0
5 0.25 12.5 9.4693 1.3571 0.1212 0.1204 -10.0
6 0.30 15.0 9.9692 1.8431 0.2012 0.2010
7 0.35 17.5 9.8556 2.3387 0.3058 0.3064 -20.0
8 0.40 20.0 9.1354 2.8135 0.4346 0.4363
Time (sec)
9 0.45 22.5 7.8531 3.2382 0.5859 0.5888
10 0.50 25.0 6.0876 3.5867 0.7565 0.7606
Fig. 3.1:Acceleration
Fig. 6.1: Acceleration vs. Time
vs. Time
11 0.55 25.0 1.4858 3.7760 0.9406 0.9463
12 0.60 25.0 -3.2073 3.7330 1.1283 1.1353
13 0.65 25.0 -7.7031 3.4603 1.3081 1.3159 4.0
14 0.70 25.0 -11.7249 2.9746 1.4690 1.4769
15 0.75 25.0 -15.0251 2.3058 1.6010 1.6082 2.0
Velocity (ft/sec)

16 0.80 25.0 -17.4007 1.4952 1.6960 1.7017

17 0.85 25.0 -18.7055 0.5925 1.7482 1.7516 0.0
18 0.90 25.0 -18.8592 -0.3466 1.7544 1.7547 0.0 0.5 1.0 1.5 2.0
19 0.95 25.0 -17.8523 -1.2644 1.7141 1.7109 -2.0
20 1.00 25.0 -15.7468 -2.1044 1.6299 1.6230
21 1.05 25.0 -12.6723 -2.8149 1.5069 1.4962 -4.0
22 1.10 25.0 -8.8179 -3.3521 1.3527 1.3387 Time (sec)
23 1.15 25.0 -4.4209 -3.6831 1.1768 1.1600
24 1.20 25.0 0.2481 -3.7874 0.9901 0.9715 Fig. 3.2:Velocity
Fig. 6.2: Velocity vs. Time
vs. Time
25 1.25 25.0 4.9019 -3.6586 0.8039 0.7846
26 1.30 25.0 9.2540 -3.3048 0.6298 0.6112 2.0
27 1.35 25.0 13.0367 -2.7475 0.4785 0.4620 ui (ft) uex (ft)
28 1.40 25.0 16.0171 -2.0211 0.3593 0.3462 1.5
Displacement (ft)

29 1.45 25.0 18.0118 -1.1704 0.2795 0.2711

30 1.50 25.0 18.8981 -0.2477 0.2441 0.2412 1.0
31 1.55 25.0 18.6214 0.6903 0.2551 0.2586
32 1.60 25.0 17.1989 1.5858 0.3120 0.3220
0.5
33 1.65 25.0 14.7179 2.3837 0.4113 0.4276
34 1.70 25.0 11.3312 3.0350 0.5468 0.5688
35 1.75 25.0 7.2472 3.4994 0.7101 0.7368 0.0
36 1.80 25.0 2.7172 3.7485 0.8913 0.9212 0.0 0.5 1.0 1.5 2.0
37 1.85 25.0 -1.9800 3.7670 1.0792 1.1105 Time (sec)
38 1.90 25.0 -6.5553 3.5536 1.2622 1.2929 Fig.3.3:
Fig. 6.3:Displacement
Displacement vs.vs.
Time
Time
39 1.95 25.0 -10.7273 3.1215 1.4291 1.4570
40 2.00 25.0 -14.2391 2.4974 1.5696 1.5928

35
 Stiffness and Mass Matrices of Continuous Systems
Axial Members
Applying the method of virtual work to undamped members subjected to axial load of p(x,t) per unit length,
WI = WE   m dx d2u/dt2 u +  u E A u dx =  p(x,t) dx u …………….….(4.1)

u1A p(x,t) u1B

A B

Fig. 4.1: Axially Loaded Member

If the displacements of a member AB (Fig. 4.1) are assumed to be interpolating functions [1(x) and 2(x)]
of two nodal displacements u1A and u1B,
u = u1A 1 + u1B 2  u = u1A 1 + u1B 2 ….…….…..………(4.2), (4.3)
d2u/dt2 = d2u1A/dt2 1 + d2u1B/dt2 2 ……………………..……(4.4)
u = u1A 1 + u1B 2  u = u1A 1 + u1B 2 …………...……….(4.5), (4.6)
Eq. (4.1) can be written in matrix form as,

m 1 1dx m 1 2dx d2u1A/dt2 EA 11dx EA 12dx u1A p(x,t)1dx

+ =
m 2 1dx m 2 2dx 2
d u1B/dt 2
EA 21dx EA 22dx u1B p(x,t)2 dx

...…...…..(4.7)
For concentrated loads p(x,t) is a delta function of x, as mentioned before. If loads XA and XB are applied at
joints A and B, they can be added to the right side of Eq. (4.7).
Eq. (4.7) can be rewritten as, Mm d2um/dt2 + Km um = fm ………………(4.8)
where Mm and Km are the mass and stiffness matrices of the member respectively, while d2um/dt2, um and fm
are the member acceleration, displacement and load vectors. They can be formed once the shape functions 1
and 2 are known or assumed.
Mmij = m i j dx, and Kmij = EA ij dx ………………………(4.9)

Flexural Members

u2A u2B
3A q(x,t) 3B
A B
L

Fig. 4.2: Transversely Loaded Member

Applying the method of virtual work to undamped members subjected to flexural load of q(x,t) per unit
length   m dx d2u/dt2 u +  uE I u dx =  q(x,t) dx u …….……………..(4.10)

Following the same type of formulation as for axial members, the member equations for undamped flexural
members subjected to transverse load of q(x,t) per unit length (Fig. 4.2) can be written in matrix form like
Eq. (4.8), but the member matrices are different here.

Interpolation functions for u2A, 3A, u2B and 3B are typically chosen in such cases, so that
u(x) = u2A 1 + 3A 2 + u2B 3 + 3B 4 ……….………………(4.11)

The size of the matrices is (44) here, due to transverse joint displacements (u2A, u2B) joint rotations (3A,
3B) and their elements are given by
Mmij = m i j dx, and Kmij = EI i j dx ………………………(4.12)

36
Example 4.1
For modulus of elasticity E = 450000 ksf, cross-sectional area A = 1 ft2, length L = 10 ft, mass per length m
= 0.0045 k-sec2/ft2, calculate the natural frequencies of a cantilever beam in axial direction, analyzing with
(i) one lumped-mass element, (ii) one consistent-mass element, (iii) two lumped-mass elements.

Solution
(i) For lumped-mass elements

mL/2 0 EA/L –EA/L

Mm = Km =
0 mL/2 –EA/L EA/L

Assuming one linear element with properties mentioned, mL/2 = 0.0225 k-sec2/ft, EA/L = 45000 k/ft

0.0225 0 45000 –45000

Mm = Km =
0 0.0225 –45000 45000

Applying the boundary conditions that the only non-zero DOF is the axial deformation at B (u1B), the mass
and stiffness matrices are reduced to (11) matrices M = 0.0225, K = 45000
 K–n2 M  = 0  45000 –  n2 0.0225 = 0  n2 = 2  106  n = 1414 rad/sec
(ii) For linear functions 1(x) = 1–x/L, 2(x) = x/L, the mass and stiffness matrices obtained from Eq. (4.7)

mL/3 mL/6 EA/L –EA/L

Mm = Km =
mL/6 mL/3 –EA/L EA/L

Assuming one linear element with properties mentioned, mL/3 = 0.015 k-sec2/ft, EA/L = 45000 k/ft

0.0150 0.0075 45000 –45000

Mm = Km =
0.0075 0.0150 –45000 45000

Applying the boundary conditions that the only non-zero DOF is the axial deformation at B (u1B), the mass
and stiffness matrices are reduced to (11) matrices M = 0.015, K = 45000
 K–n2 M  = 0  45000 –  n2 0.015 = 0  n2 = 3  106  n = 1732 rad/sec
(iii) For two lumped-mass elements of length 5 each, mL/2 = 0.01125 k-sec2/ft, EA/L = 90000 k/ft
The following mass and stiffness matrices are obtained for each element

0.01125 0 90000 –90000

Mm = Km =
0 0.01125 –90000 90000
Applying the boundary conditions that axial deformation at A (u1A) is zero, only the axial deformations at B
(u1B) and C (u1C) are non-zero, the mass and stiffness matrices are reduced to (22) matrices.

u1A= 0 u1B u1C

A B C
5 5

0.01125 0 90000 –90000

M= K=
0 0.0225 –90000 180000
  K-n2 M  = 0  (90000 –n2 0.01125) (180000 –n2 0.0225) – (–90000)2 = 0
 n = 1531 rad/sec, 3696 rad/sec3
Analytical solutions for the first two natural frequencies are 1571 rad/sec, 4712 rad/sec respectively.

37
Example 4.2
For the member properties E = 450000 ksf, I = 0.08 ft4, L = 10 ft, m = 0.0045 k-sec2/ft2, calculate the
approximate first natural frequency of the cantilever beam in transverse direction, analyzing with
(i) one lumped-mass element, (ii) one consistent-mass element.
Solution
(i) For cubic polynomial functions
1(x) = 13(x/L)2 +2(x/L)3, 2(x) = x{1(x/L)}2, 3(x) = 3(x/L)2 2(x/L)3, 4(x) = (xL)(x/L)2
with lumped mass mL/2 at both ends and constant EI, the following matrices are obtained from Eq. (4.12)
1 0 0 0 12 6L -12 6L
Mm = (mL/2) 0 0 0 0 Km = (EI/L3) 6L 4L2 -6L 2L2
0 0 1 0 -12 -6L 12 -6L
0 0 0 0 6L 2L2 -6L 4L2
In this case, mL = 0.045 k-sec2/ft, EI/L3 = 36 k/ft

1 0 0 0 12 60 -12 60
Mm = 0.0225 0 0 0 0 Km = 36 60 400 -60 200
0 0 1 0 -12 -60 12 -60
0 0 0 0 60 200 -60 400
Applying the boundary conditions that the only non-zero degrees of freedom are the vertical deflection and
rotation at B (u2B and 3B), the mass and stiffness matrices are reduced to (2  2) matrices

1 0 432 -2160
M = 0.0225 K=
0 0 -2160 14400

u2A= 0 u2B
3A= 0 3B
A B

10

 K  n2 M  = 0  (432  n2 0.0225) 14400 – (2160)2 = 0  n = 69.28 rad/sec

(ii) For cubic polynomial functions with uniform m

156 22L 54 -13L 156 220 54 -130

Mm = (mL/420) 22L 4L2 13L -3L2 = 1.07110 -4
220 400 130 -300
54 13L 156 -22L 54 130 156 -220
-13L -3L2 -22L 4L2 -130 -300 -220 400
Applying the boundary conditions, the mass and stiffness matrices are reduced to

156 -220 432 -2160

M = 1.07110-4 K=
-220 400 -2160 14400

 K  n2 M  = 0  (432  n2 0.01671) (14400  n2 0.04286) – (2160 + n2 0.02357)2 = 0
 n = 99.92 rad/sec, 984.49 rad/sec

The exact results for the first two natural frequencies are 99.45 rad/sec and 623.10 rad/sec respectively.
Therefore, as was the case for axial vibrations, the natural frequencies are under-estimated for lumped-mass
element and over-estimated for consistent-mass element.

38
 Dynamic Analysis of Trusses and Frames
Two-dimensional Trusses
The mass and stiffness matrices derived for axially loaded members can be used for the dynamic analysis of
two-dimensional trusses. One difference is that here the transverse displacements (u2A, u2B) are also
considered in forming the matrices, so that the size of the matrices is (4  4) instead of (2  2).

mL/2 0 0 0 mL/3 0 mL/6 0

MmL = 0 0 0 0 or MmL = 0 0 0 0 ……….(5.1)
(Lumped) 0 0 mL/2 0 (Consistent) mL/6 0 mL/3 0
0 0 0 0 0 0 0 0
The member matrices formed in the local axes system by Eq. (5.1) can be transformed into the global axes
system by considering the angles they make with the horizontal.

 0  /2
MmG = (mL/2) or MmG = (mL/3) . ……..….(5.2)
0  /2 

where  is a (2  2) matrix of coefficients given by

C2 CS
= ………………..(5.3)
CS S2
The mass and stiffness matrices (from previous formulations) and load vector of the whole structure can be
assembled from the member matrices and vector (MmG, KmG and fmG). They are obtained in their final forms
only after applying appropriate boundary conditions.

Two-dimensional Frames
The matrices formed for flexural members and already used for a cantilever beam can be used for the
dynamic analysis of two-dimensional frames. The elements of the ith row and jth column of the mass and
stiffness matrices are given by Eq. (4.12) in integral form and can be evaluated once the shape functions i
and j are known or assumed [as shown in Example 4.2]. However, the axial displacements of joints (u1A,
u1B) are also considered for frames in addition to the transverse displacements (u2A, u2B) and rotations (3A,
3B) about the out-of-plane axis considered in forming the matrices for beams, so that the size of the matrices
is (6  6) instead of the (4  4) matrices shown for beams.

If shape functions of Example 4.2 are assumed for frame members of uniform cross-section, the member
mass and stiffness matrices take the following forms in the local axes system

1 0 0 0 0 0 140 0 0 70 0 0
0 1 0 0 0 0 0 156 22L 0 54 -13L
2 2
0 0 0 0 0 0 Mm = (mL/420) 0 22L 4L 0 13L -3L
L ...…...(5.4)
MmL = (mL/2)
0 0 0 1 0 0 70 0 0 140 0 0
(Lumped) 0 0 0 0 1 0 (Consistent) 0 54 13L 0 156 -22L
0 0 0 0 0 0 0 -13L -3L2 0 -22L 4L2

Denoting the global structural matrices by M and K respectively and assuming appropriate damping ratios,
the damping matrix C can be obtained as,
C = a0 M + a1 K ………………..(5.5)
The dynamic analysis can be carried out once these matrices and vector are formed.

39
Example 5.1
For the plane truss shown below, modulus of elasticity E = 30000 ksi, cross-sectional area A = 2 in2, mass
per length m = 1.510-6 k-sec2/in2. Calculate its natural frequencies using consistent mass matrices.
Solution
u6 u8
C D
u5 u7

7.5
u2 u4
45 45
A u3
u1 B

15 7.5
The truss has 8 DOF. The displacements u1~u4 and u7, u8 are restrained, so that only two DOF (u5, u6) are
non-zero. There are five members in the truss (including two zero-force members), all with the same cross-
sectional properties, but different lengths. The member mass and stiffness matrices can be obtained from
C2 CS C2/2 CS/2 C2 CS -C2 -CS
G
Mm = (mL/3) CS S2 CS/2 S2/2 G
Km = (EA/L) CS S2 -CS -S2
C2/2 CS/2 C2 CS -C2 -CS C2 CS
CS/2 S2/2 CS S2 -CS -S2
CS S2
For member AB, C = 1, S = 0, L = 15 = 180, mL/3 = 9.010-5 k-sec2/in, EA/L = 333.33 k/in
1.0 0 0.5 0 1.0 0 -1.0 0
MABG = 9.010-5 0 0 0 0 KABG = 333.33 0 0 0 0
0.5 0 1.0 0 -1.0 0 1.0 0
0 0 0 0 0 0 0 0
DOF [1 2 3 4]
The matrices for AB and CD are the same, but the latter connects displacements 5, 6, 7 and 8
For member AC, C = 0.707, S = 0.707, L = 10.607 = 127.28
mL/3 = 6.3710-5 k-sec2/in, EA/L = 471.41 k/in
0.5 0.5 0.25 0.25 0.5 0.5 -0.5 -0.5
MACG = 6.3710-5 0.5 0.5 0.25 0.25 KACG = 471.41 0.5 0.5 -0.5 -0.5
0.25 0.25 0.5 0.5 -0.5 -0.5 0.5 0.5
0.25 0.25 0.5 0.5 -0.5 -0.5 0.5 0.5
[1 2 5 6]
The matrices for AC and BD are the same, but the latter connects displacements 3, 4, 7 and 8
For member BC, C = -0.707, S = 0.707, L = 10.607 = 127.28
mL/3 = 6.3710-5 k-sec2/in, EA/L = 471.41 k/in
0.5 -0.5 0.25 -0.25 0.5 -0.5 -0.5 0.5
-5
MBCG = 6.3710 -0.5 0.5 -0.25 0.25 KBC G
= 471.41 -0.5 0.5 0.5 -0.5
0.25 -0.25 0.5 -0.5 -0.5 0.5 0.5 -0.5
0.25 -0.25 -0.5 0.5 0.5 -0.5 -0.5 0.5
[3 4 5 6]
Applying boundary conditions, the mass and stiffness matrices for the whole truss can be assembled as

15.37 0 804.74 0
M = 10-5 K=
0 6.37 0 471.41

 Kn2 M  = 0  (804.74 n2 0.0001537) (471.41 n2 0.0000637) = 0

 n = 2288 rad/sec, 2720 rad/sec

40
 Practice Problems on Structural Dynamics
1. For a (20  20  20) overhead water tank supported by a (25  25) square column, use the CAA
method (c = 0) to calculate the relative displacement at time t = 0.01 second, when subjected to ground
acceleration ag = 10 Cos(50t) (ft/sec2) [use k = 3EI/L3].
Assume the total weight of the system to be concentrated in the tank, and initial displacement and
velocity are both zero [Given: E of concrete = 400  103 k/ft2, Unit weight of water = 62.5 lb/ft3].

2. For beam AB loaded as shown below, use the CAA method to calculate rotation at A at time t = 0.10 sec
(starting with zero initial displacement and velocity) [Given: EI = 40  103 k-ft2, m = 0.0045 k-sec2/ft2].

w(t) w (k/ft) 2

A B
t (sec)
20 0 0.2 0.4

3. Calculate the natural frequency(s) and period(s) of the

(i) Axial rod (Example 4.1) and cantilever beam (Example 4.2) shown in class if each carries a weight of
3 kip at the free end A.
(ii) Lumped mass system shown below
m2 m1 = m2 = 1 k-sec2/ft
k1 = k2 = 10 k/ft
k2 c2

m1

k1 c1

(iii) Beam ABC shown below, if it made of a material with unit weight = 0.15 k/ft3 [considering flexural
deformations only with geometric nonlinearity].

A and B are Guided Rollers E = 400  103 k/ft2

B
A C 24
500 k 250 k 12

24 24 Beam Section

(iv) Frame abcde shown below, if it weighs 3.0 kN/m.

a b
c

EI = constant = 10  103 kN-m2

10m
a and c are Guided Rollers
d e
10m 10m

41
 Structures on Flexible Foundations
Rather than idealized support conditions (i.e., roller, hinged or fixed), it is more rational to assume structures
to be supported on flexible supports. In addition to real springs, foundations on flexible supports (e.g.,
columns) or soils can also be modeled by springs for horizontal, vertical displacements, as well as bending
and torsional rotations.

For example, if EI = 80 × 103 k-ft2, the Stiffness Matrix for the beam in Fig. 1.1 is S3 = 16 × 103 k-ft/rad
The stiffness formulation is, S3 A =  P0L/8  16 × 103 A = 25  A = 1.56 × 10-3 rad

10 k 10 k vB
A A B
uB

10′ 10′ 10′ 10′

Fig. 1.1 Fig. 1.2

If springs of stiffness Kh, Kv and K replace the ‘fixed’ support (Fig. 1.2), the stiffness formulation becomes

S3 0 S2 S4 A P0L/8
0 Sx + Kh 0 0 uB 0 ………………..(1.1)
= 
S2 0 S1 + Kv S2 vB P0/2
S4 0 S2 S3 + K B P0L/8

Stiffness of Circular Foundations and Long Pile Foundations

Motion KHalfspace KEmbed KPile
h 8GR/(2-) 4 GE 4 G0.75 (EpIp)0.25
v 4GR/(1-) 2.75 GE 1.5 G0.5 (EpAp)0.5
 8GR3/(3-3) [8+4(E/R)2] GER2/3 2 G0.25 (EpIp)0.75
t 16GR3/3 12 GER2 3 R G0.5 (EpJp)0.5
[h for horizontal, v for vertical ,  for bending and t for torsional motion]

If EA = 800 × 103 k, Sx = 40 × 103 k/ft, S1 = 120 k/ft, S2 = 1200 k-ft/ft, S4 = 8 × 103 k-ft/rad
G = Shear modulus of sub-soil, R = Radius of circular foundation,  = Poisson’s ratio
Kh = 8GR/(2), Kv = 4GR/(1), K = 8GR3/(33)
Assuming shear-wave velocity vs = 1000 ft/s, G = svs2 = (0.12/32.2) × (1000)2 = 3.73 × 103 k/ft2
R = 2 ft,  = 0.30  Kh = 8GR/(2-) = 35.08 × 103 k/ft, Kv = 4GR/(1-) = 42.59 × 103 k/ft,
K = 8GR3/(3-3) = 113.58 × 103 k-ft/rad

16.0 0 1.2 8.0 A 25 A = 1.721 × 10-3 rad

0 75.07 0 0 uB 0 uB = 0
103 =  
1.2 0 42.71 1.2 vB 5 vB = 0.133 × 10-3 ft
8.0 0 1.2 129.58 B 25 B = 0.298 × 10-3 rad

These values may vary significantly with the stiffness(es) of foundation, which are directly proportional to
the value of G (shear modulus of sub-soil), which in turn depends on the shear-wave velocity vs of sub-soil.
Table below shows the variation of A, vB and B with vs.

vs (ft/s) A (10-3 rad) vB (10-3 ft) B (10-3 rad)

 1.563 0 0
1000 1.721 0.133 0.298
300 2.476 1.138 1.657
100 3.372 7.325 2.520

42
 Practice Problems on Structures on Flexible Foundations
1. Use the Stiffness Method to calculate the horizontal and vertical deflections at joint b of the plane
trusses loaded as shown below, if the support at b is a
(i) roller support
(ii) circular foundation of radius 2-ft on the surface of sub-soil (half-space) with shear-wave velocity
(vs) equal to 300 ft/sec
[Given: Sx of truss members = 10  103 k/ft, Unit weight of soil = 0.12 k/ft3, Poisson’s ratio = 0.30].

c c
a a

50 50
10 k 10 k

b 10 k b 10 k
Spring Stiffness kvb
50 50

(i) (ii)

2. The figure below represents a cricket stump ab that has been hit by a horizontal force P = 10 lb at b.
Calculate the horizontal deflection and rotation at point b of the stump, considering the joint a to be
(i) fixed (i.e., no deflection and rotation at a) [Hint: May use  = PL3/3EI,  = PL2/2EI],
(ii) hinged (i.e., no deflection at a) and additionally supported by a circular foundation of 0.5-ft radius
on sub-soil (half-space) with shear-wave velocity (vs) equal to 1000 ft/sec
[Given: Unit weight of soil = 110 lb/ft3, Poisson’s ratio = 0.25, EI of the stump = 500 lb-ft2].

P = 10 lb b P = 10 lb b

2.5

ka

a a

43
 Short Questions and Explanations

Stiffness Method for 3D Trusses vs. 3D Frames

1. Unknowns: Deflections only vs. Deflections + Rotations
2. No. of Unknowns: doki = 3j vs. doki = 6j
3. Member Stiffness Matrix: (6  6) vs. (12  12)
4. Member Properties: E, A vs. G, E, A, J, Iy, Iz
5. Forces Calculated: Member Axial Forces vs. Member Axial, Shear Forces, Torsions, BM’s

Stiffness Method for

- 2D Trusses vs. 2D Frames
- 2D Trusses vs. 3D Trusses
- 2D Frames vs. 3D Frames
- Linear vs. Nonlinear Analysis
- Analysis for Geometric vs. Material Nonlinearity

Briefly explain why

- axial deformations are sometimes neglected for the structural analysis of frames but not trusses
- joint rotations are considered in calculating the doki of frames, but not trusses
- stiffness matrix of a 3D truss member is (6  6) while that of a 3D frame member is (12  12)
- the matrices K and G used for the nonlinear analysis of frames are only approximate
- the formulation of the geometric stiffness matrix G is a nonlinear problem
- a structure becomes unstable at buckling load (explain in terms of stiffness matrix)
- the terms material nonlinearity, plastic moment and collapse mechanism
- frames can be approximately modeled by lumped-mass systems
- the effect of foundation flexibility can be beneficial or harmful to the structure

* Sample questions from previous semester

Briefly explain (Set 1)
(i) why a truss member hinged at both ends is a zero-force member,
(ii) why the stiffness matrix and geometric stiffness matrix of a structure are both symmetric (use
equations with shape function ),
(iii) the main advantage and disadvantage of using the bending moment diagram method to obtain the
collapse load of a structure,
(iv) the main advantage and disadvantage of using consistent-mass matrix in structural dynamics,
(v) the effect of foundation flexibility on the structural response to seismic vibration.

Briefly explain (Set 2)

(i) how the effect of support settlement can be incorporated in the structural analysis of beams,
(ii) why the stiffness matrix and mass matrix of a structure are both symmetric (use equations with
shape function ),
(iii) the main advantage and disadvantage of using the energy method to obtain the collapse load of a
structure,
(iv) the main advantage and disadvantage of using lumped-mass matrix in structural dynamics,
(v) the effect of foundation flexibility on the structural response to wind vibration.

44