Chapter 1 Review of Matrix Method

List of Symbols
Some of general symbols used throughout this course are listed below. Other
symbols that are not listed here will be explained when they are covered in the
lecture.
F

= load vector of FE system model

= point load

= body force vector of an element

= traction force vector of an element

= displacement vector of FE system model

q, u, d =

displacement vectors of elements

= displacement in the x direction

= displacement in the y direction

= displacement in the z direction

= element stiffness matrix

= stiffness matrix of FE system model

= strain-displacement matrix

= interpolation / shape function (matrix) in , , and coordinate system

= Material property matrix

li

= Lagrange polynomial

= number of nodal displacements of FE system model (total number of degree

of freedoms of FE system model)

Ni
p()

= interpolation polynomial (element of matrix N)

= approximating polynomial

= traction or surface force vector

= surface

= body force vector

= volume

= displacement component in x direction

= displacement component in y direction

= displacement component in z direction

= point load

u = displacement vector

Redzuan Abdullah, FKA, UTM

Chapter1 page 1

Chapter 1 Review of Matrix Method

COURSE COVERAGE

This course emphasizes:

Basic concepts of Finite Element Method (FEM) covering linear static problems

Formulation of the Finite Elements (FE) for structural application spatially

discretized by displacement-formulated FEM

Analysis of structures using stiffness method

Computer applications (LUSAS)

Redzuan Abdullah, FKA, UTM

Chapter1 page 2

Chapter 1 Review of Matrix Method

REVIEW OF STIFFNESS MATRIX

METHOD
The stiffness matrix method underlies the basis for the computer analysis software.
In this topic, we will review the basic concepts of stiffness matrix analysis:

Basic terms and coordinate system

Formulation of stiffness matrix for 2D truss element by direct stiffness

method

Formation of system equilibrium equation

Linear elastic solution

Basic Terms, Terminology and Coordinate system

Element (member) Represent the mass of the structure.

Node / joint Where the elements are jointed.

Connectivity arbitrarily selection of near and far node of the element

determines the member connectivity.

Fig. 1 (a) Truss system, (b) Truss element

Global coordinate, x y : refer to the whole structure.

Local coordinate, x y: refer to member direction

Redzuan Abdullah, FKA, UTM

Chapter1 page 3

Chapter 1 Review
w of Matrix Method

m shall folloow the rightt hand

The orthoggonal axes of the Carteesian coordiinate system
rule. Posittive rotationn and positivve moment are definedd in terms off the right hand
h
rule.

Fig. 2 Coordin
nate system
m follows rigght hand ru
ule

Deegree of freeedom DO
OF. In mechaanics, degreees of freedoom (DOF) are
a the
sett of indepenndent displaacements an
nd/or rotations that speccify compleetely the
dissplaced or deformed
d
poosition and orientation
o
of the bodyy or system. Each
dissplacement componentt is a Degreee of Freedom
m. The num
mber of the DOF
D
is
thee function of
o the manneer in which the real struucture has bbeen idealizzed for
analysis.

Loocal directioon DOF deepends on element

e
orieentation behhavior.

Gllobal directiion DOF - Displaceme

D
ent componeents in globbal direction
n.

Freee DOF DOF

D
at unsuupported no
odes unknoown.

Fixxed DOF DOF at suppported nod

des knownn.

Ideentify the DOF

D
for the truss structu
ure in Fig. 1

Ideentify the DOF

D
for the beam and frame
f
structtures in the figures belo
ow

Redzuan Abdullah, FKA

A, UTM

Chapteer1 page 4

Chapter 1 Review
w of Matrix Method

Fig. 3 Beam structurre showing the DOF

me structurre showing the DOF

Fig. 4 Fram

Redzuan Abdullah, FKA

A, UTM

Chapteer1 page 5

Chapter 1 Review of Matrix Method

Stiffness method
The stiffness, k, of a body is a measure of the resistance offered by an elastic body to
deformation. For an elastic body with a single DOF (for example, stretching or
compression of a rod), the stiffness is defined as k =

F
.
d

Fig. 5 Definition of Stiffness

Stiffness is a property of a structure, which is an extensive property of the solid body
dependent on the material and the shape and boundary conditions. For example, for
an element in tension or compression, the axial stiffness is k =

AE
. For flexural
L

member such as a beam, the bending or rotational stiffness is k =

nEI
, where n is
L

integer whose value depends on the type of support or boundary conditions.

The stiffness method of analysis deals with analyzing the forces in structural
components by first, solving for the displacement from equilibrium equation. Once
the displacement is obtained, the internal forces in the structure can be calculated.
For example, from Fig. 5, the equilibrium of forces at node 2 is F = kd, where F is
the external force and kd is the internal force in the member resulting from a stretch,
d. This internal force resists the external force (or the applied load) so that the end
point is in equilibrium. F = kd is the equilibrium equation for a single DOF system.
For multiple DOF system, a systematic formulation to derive the equilibrium

Redzuan Abdullah, FKA, UTM

Chapter1 page 6

Chapter 1 Review of Matrix Method

equation which is presented in the form of matrix equation, and subsequently the
solution procedure, is required.

Truss Structure
Truss element formulation (Direct Method)
Determination of stiffness, k and equilibrium equation for the truss member by direct
method is given below.

Fig. 6 A truss structure under global coordinate system, displaced due to

loading

Fig. 7 Truss element, one DOF per node in local direction

N denotes near Node and F denotes Far node

Redzuan Abdullah, FKA, UTM

Chapter1 page 7

Chapter 1 Review of Matrix Method

AE
u 'N
L

Fig. 7(a)

f N' =

Fig. 7(b)

f N'' =

f F' =

AE
u 'F
L

f F'' =

AE
u 'N
L

AE
u 'F
L

Fig. 7(c) is the result of superposition of Fig. 7(a) and (b):

fN =

AE
AE
u 'N
u 'F
L
L

fF =

AE
AE
u 'F
u 'N
L
L

Rearrange the above into matrix form:

f N AE 1 1 u 'N
=

f F L 1 1 u 'F

(1)

(2)

f = k'q

Where k ' =

AE 1 1
is the truss member stiffness matrix in local axis and Eq.
L 1 1

(1) is the equilibrium equation for a two DOF truss element, in local direction.

Transformation Matrix
Each element in the structural system is orientated in local or arbitrary directions.
When they are assembled to form a truss system, their DOF and forces which are in
the element direction, are needed to be transformed into a common direction so that
equilibrium equation at their nodes can be written. The common direction is in the
global coordinate system.
Cosine directions of the member are used in the transformation:

c = cos x =

Redzuan Abdullah, FKA, UTM

xF x N
=
L

xF x N
( xF x N ) 2 + ( y F y N ) 2

Chapter1 page 8

Chapter 1 Review of Matrix Method

s = sin x = cos y =

yF y N
=
L

yF y N
( xF x N ) 2 + ( y F y N ) 2

(3)

Displacement Transformation Matrix

x'
QFy Cos

QFx Cos
F

QFy
x

QFx

x
N

(b)

Fig. 8 Resolving displacement into global coordinate

Fig. 8(a) and (b):

u ' N = QNx Cos x + QNy Cos y

Fig. 8(c) and (d):

u 'F = QFx Cos x + QFy Cos y

u 'N = QNx c + QNy s

u 'F = QFx c + QFy s

where c and s are as defined in Eqn. (3).

Rewrite the above equations into a matrix form,

Redzuan Abdullah, FKA, UTM

Chapter1 page 9

Chapter 1 Review of Matrix Method

QNx

u 'N c s 0 0 QNy
=
or

u 'F 0 0 c s QFx
QFy

(4)

q = TQ

where q = nodal displacement matrix in local direction, T = displacement

transformation matrix, Q = nodal displacement matrix in global direction

Force Transformation Matrix

Fig. 9 Resolving local forces in to global coordinates

Node N,

Node F,

FNx = f N cos x
FNy = f N cos y
FFx = f F cos x
FFy = f F cos y

FNx = f N c
FNy = f N s
FFx = f F c
FFy = f F s
FNx c
F
Ny s
=
FFx 0
FFy 0

Redzuan Abdullah, FKA, UTM

0
0 f N

c fF

Chapter1 page 10

Chapter 1 Review of Matrix Method

(5)

F = TT f

Where F is nodal forces in global direction, f is nodal force in local direction, TT is

force transformation matrix which is the transpose of the displacement
transformation matrix.

Element Global Stiffness Matrix

Eq. (2)

f = k'q

Eq. (4)

q = TQ

Eq. (2) into (4)

f = k'TQ

Eq. (5)

F = TT f

(6)

F = TTk'TQ

Eq. (6) into (5)

F = kQ

(7)

k = TTk'T

(8)

where

k = member stiffness matrix in global direction. It represents the force-displacement

relation for the member in global coordinate.

Substitute the matrices into Eq (8) yields,
c
s

k=
0

0
0 AE 1 1 c s 0 0
c L 1 1 0 0 c s

Expand the above matrices yields the stiffness matrix of a truss element in global
direction,

Redzuan Abdullah, FKA, UTM

Chapter1 page 11

Chapter 1 Review of Matrix Method

Nx

Ny

Fx

Fy

c2
cs c 2 cs N x

s 2 cs s 2 N y
AE cs
k=
L c 2 cs c 2
cs Fx

2
cs
s 2 Fy
cs s

(9)

The location of each element in k is referenced with each global DOF associated
with the near end N, followed by the far end F, as indicated by code number Nx,
Ny, Fx, Fy along the rows and columns.

Finally the equilibrium equation, Eq. (7) of a truss element in global direction is,

F = kQ
where,
FNx
c2
cs c 2 cs QNx
F

s 2 cs s 2 QNy
Ny AE cs
=

2
2
cs QFx
FFx L c cs c

2
FFy
cs
s 2 QFy
cs s

(10)

Truss Structure Stiffness Matrix and Equilibrium Equation

Once all the member stiffness matrices are formed in the global coordinate, it
becomes necessary to assemble them in the proper order so that the structure
stiffness matrix [K] for the entire truss system can be formed. We can write
equilibrium equations at the nodes by first expanding Eqn. (10).

Redzuan Abdullah, FKA, UTM

Chapter1 page 12

Chapter 1 Review of Matrix Method

AE 2
c ( QNx QFx ) +
L
AE
FNy =
cs ( QNx QFx ) +
L
AE 2
FFx =
c ( QFx QNx ) +
L
AE
FFy =
cs ( QFx QNx ) +
L

FNx =

AE
cs ( QNy QFy )
L
AE 2
s ( QNy QFy )
L
AE
cs ( QFy QNy )
L
AE 2
s ( QFy QNy )
L

(11)

These four equations relate the nodal forces and net nodal displacements (amount of
stretches in the element) in global direction resulting from the movement at near and
far nodes (see Fig. 10).

Fig. 10 Forces and displacements at the ends of truss member in global direction

We can write Eqn. (11) for each element in a truss system and assemble them at
each node to obtain the equilibrium at the node.

Redzuan Abdullah, FKA, UTM

Chapter1 page 13

Chapter 1 Review of Matrix Method

Example

Fig. 11 Truss system labeled with nodes and DOF

Fig. 12 Forces acting node 1

The equilibrium in x-direction at node 1 is,

F1 = F1x + F2 x
Expand it by using Eqn. (11),

AE 2
AE
AE 2
AE
F1 =
c1 ( Q1 Q5 ) +
c1s1 ( Q2 Q6 ) +
c2 ( Q1 Q3 ) +
c2 s2 ( Q2 Q4 )
L 1
L 1
L 2
L 2
AE 2 AE 2
AE

AE
AE 2
F1 =
c1 +
c2 Q1 +
c1s1 +
c2 s2 Q2
c2 Q3
L 2
L 2
L 2
L 1
L 1

AE
AE 2
AE

c2 s2Q4
c1 Q5
c1s1Q6
L 2
L 1
L 1

Redzuan Abdullah, FKA, UTM

Chapter1 page 14

(12)

Chapter 1 Review of Matrix Method

The equilibrium of forces in y-direction at node 1 is,

F2 = F1 y + F2 y

AE
AE 2
AE
AE 2
0=
c1s1 ( Q1 Q5 ) +
s1 ( Q2 Q6 ) +
c2 s2 ( Q1 Q3 ) +
s2 ( Q2 Q4 )
L 1
L 1
L 2
L 2
AE

AE 2 AE 2
AE
AE
0 =
c1s1 +
c2 s2 Q1 +
s1 +
s2 Q2
c2 s2Q3
L 2
L 2
L 2
L 1

L 1
AE 2
AE
AE 2

s2 Q4
c1s1Q5
s1 Q6
L 2
L 1
L 1

(13)

Fig. 13 Forces acting at node 2

The equilibrium of forces in x-direction at node 2 is,

F3 = F2 x + F3 x
AE 2
AE
AE 2
AE
F3 =
c2 ( Q3 Q1 ) +
c2 s2 ( Q4 Q2 ) +
c3 ( Q3 Q5 ) +
c3 s3 ( Q4 Q6 )
L 2
L 2
L 3
L 3
AE 2 AE 2
AE 2
AE
AE
AE 2
0 =
c2 Q1
c2 s2Q2 +
c2 +
c3 Q3 +
c2 s2Q4
c3 Q5
L 2
L 2
L 3
L 2
L 3
L 2
(14)
The equilibrium of forces in y-direction at node 2 is,
R4 = F2 y

AE
AE 2
=
c2 s2 ( Q3 Q1 ) +
s2 ( Q4 Q2 )
L 2
L 2
AE
AE 2
AE
AE 2
=
c2 s2Q1
s2 Q2 +
c2 s2Q3 +
s2 Q4
L 2
L 2
L 2
L 2

Redzuan Abdullah, FKA, UTM

(15)

Chapter1 page 15

Chapter 1 Review of Matrix Method

Fig. 14 Forces acting at node 3

The equilibrium of forces in x-direction at node is,

R5 = F1x + F3 x
AE 2
AE
AE 2
AE
R5 =
c1 ( Q5 Q1 ) +
c1s1 ( Q6 Q2 ) +
c3 ( Q5 Q3 ) +
c3 s3 ( Q6 Q4 )
L 1
L 1
L 3
L 3
AE 2 AE 2
AE 2
AE
AE 2
AE
=
c1 Q1
c1s1Q2
c3 Q3 +
c1 +
c3 Q5 +
c1s1Q6
L 1
L 1
L 3
L 3
L 1
L 1
(16)
The equilibrium of forces in y-direction at node 3 is,
R6 = F1 y

AE
AE 2
=
c1s1 ( Q5 Q1 ) +
s1 ( Q6 Q2 )
L 1
L 1
AE
AE 2
AE
AE 2
=
c1s1Q1
s1 Q2 +
c1s1Q5 +
s1 Q6
L 1
L 1
L 1
L 1

(17)

Eqn. (12) to (17) can be combined into matrix form as follows,

AE 2 AE 2
L c1 + L c2
1

AE
AE

c1s1 +
c2 s2
F1 L 1
L 2
0
AE c2 2
0 L 2
=
R4 AE c s
R5 L 2 2 2

R6 AE 2
c

L 1 1

AE c s
L 1 1 1

AE
AE

c1s1 +
c2 s2
L 1
L 2
AE 2 AE 2

s1 +
s2
L 1
L 2

AE 2

c2
L 2
AE

c2 s2
L 2

AE

c2 s2
L 2
AE 2

s2
L 2

AE 2

c1
L 1
AE

c1s1
L 1

AE

c2 s2
L 2
AE 2

s2
L 2

AE 2 AE 2

c2 +
c3
L 2
L 3
AE

c2 s2
L 2

AE

c2 s2
L 2
AE 2

s2
L 2

AE 2

c3
L 3

AE

c1s1
L 1
AE 2

s1
L 1

AE 2

c3
L 3

0
AE 2 AE 2

c1 +
c3
L 1
L 3
AE

c1s1
L 1

AE

c1s1
L 1

AE 2

s1
L 1
Q1
Q2
0

Q3
Q
4
0
Q5
Q
AE
6

c1s1
L

AE 2
s

1
L 1

(18)

Redzuan Abdullah, FKA, UTM

Chapter1 page 16

Chapter 1 Review of Matrix Method

The above is called equilibrium equation for the given truss, which can be shorten
symbolically as,
(19)

F = KQ

Eqn. (18) considers that all DOFs can move, hence the system is unstable where rigid
body movement may occur. In order to ensure that the truss is stable, we need to
provide sufficient supports such as Q4, Q5 and Q6 as shown in the figure. As can be
seen, the matrices of Eqn. (18) were arranged such that the first three DOFs are the
free DOF and the last three DOFs are the fixed DOF. Eqn. (18) can be partitioned to
separate between the stiffness coefficients associated with the free and fix DOF. The
partitioning of the matrices above can be represented symbolically by,
Fk K11
F = K
u 21

K12 Qu
K 22 Qk

(20)

{Fk } = [ K11 ]{Qu } + [ K12 ]{Qk }

(21)

{Fu } = [ K 21 ]{Qu } + [ K 22 ]{Qk }

(22)

where
{Fk} Force vector at unconstraint DOF (external loads).
{Fu} Force vector at constraint DOF (support reactions)
{Qu} Displacements vector at unconstraint DOF (The first variables that we want
to solve)
{Qk} Displacements vector at constraint DOF. For rigid supports, the
displacements are zero.

Expanding Eqn. (21) gives,

Redzuan Abdullah, FKA, UTM

Chapter1 page 17

Chapter 1 Review of Matrix Method

AE 2 AE 2

c1 +
c2
L 1
L 2

F1
AE
AE
0 =
c1s1 +
c2 s2
L 2
0 L 1

AE 2

c2

L 2
AE

c2 s2
L

AE 2

s2
L

AE

c2 s2
L 2

AE
AE

c1s1 +
c2 s2
L 1
L 2
AE 2 AE 2

s1 +
s2
L 1
L 2
AE

c2 s2
L 2

Q
1
Q2 +
Q
3
AE 2 AE 2

c2 +
c3
L 2
L 3
AE 2

c2
L 2
AE

c2 s2
L 2

AE 2
AE

c1
c1s1
L 1
L 1
Q
4
AE
AE 2

c
s
s
Q
11

1 5
L 1
L 1
Q
6
AE 2

c3

L 3

(23)

Expanding Eqn. (22) gives;

AE

c2 s2
L

2
R4
AE 2
R5 =
c1
L

1

R6
AE

c1s1
L 1

AE 2 AE

s2
c2 s2
L 2
L 2
Q
1
AE
AE

c
s
c
Q
11

3 2+
L 1
L 3
Q
3
AE 2

0
s1

L 1

AE 2
0

s2
L 2

AE 2 AE 2
0

c1 +
c3
L 1
L 3

AE

c1s1

L 1

Q
4
AE

c1s1 Q5
L 1
Q
6
AE 2

s1
L 1
0

(24)

The structure can be analyzed to obtain the free DOF using Eqn. (23) and the
reaction force can be determined using Eqn. (24). The individual member forces can
be found using Eqn. (6).

Redzuan Abdullah, FKA, UTM

Chapter1 page 18

Chapter 1 Review of Matrix Method

Example
Analyze the following truss to determine the displacements components at free DOF
nodes and the reaction forces at fixed DOF (supports).

6
AE 0.0001 200(10 )
=
= 4717

4.24
L 1
c1 = cos 1 = 0.707

s1 = sin 1 = 0.707
0.0001 200(106 )
AE
=
= 4000

5
L 2
c1 = cos 2 = 0.8
s1 = sin 2 = 0.6
6
AE 0.0001 200(10 )
= 2857

=
7
L 3
c1 = cos 3 = 1

s1 = sin 3 = 0
Substitute the above values into Eqn. (18)

Redzuan Abdullah, FKA, UTM

Chapter1 page 19

Chapter 1 Review of Matrix Method

20 4921
0 441

0 -2560
=
R4 1920
R5 -2361

R6 -2361

441

-2560

1920

-2361

3801
1920

1920
5417

-1440
-1920

-2361
-2857

-1440
-2361

-1920
-2857

1440
0

0
5219

-2361

2361

-2361 Q1
-2361 Q2
0 Q3

0 Q4 = 0
2361 Q5 = 0

2361 Q6 = 0

{Fk } = [ K11 ]{Qu } + [ K12 ]{Qk }

20 4921

0 = 441
0 -2560

441
3801
1920

-2560 Q1

1920 Q2 + {0}
5417 Q3

{Qu } = [ K11 ] {Fk }

1

Q1 4921

Q2 = 441
Q -2560
3

441
3801
1920

-2560
1920
5417

20 0.0064

0 = 0.0028 m
0 0.0040

{Fu } = [ K 21 ]{Qu } + [ K 22 ]{Qk }

R4 1920

R5 = -2361
R -2361
6

-1440 -1920 0.0064 8.57 kN

-2361 -2857 0.0028 = 20.00 kN

-2361
0 0.0040 8.57 kN

The equilibrium equation ( Eqn. (19)) for any structural system can be formed
systematically by adding together the stiffness coefficients of each element in the
same DOF;

K = k 1 + k 2 + ... + k n

(25)

Where k1, k2, kn are obtained using Eqn. (9)

Redzuan Abdullah, FKA, UTM

Chapter1 page 20

Chapter 1 Review of Matrix Method

Example

The general form of equilibrium equation for the truss element is Eqn. (10);
FNx
c2
cs c 2 cs QNx
F

s 2 cs s 2 QNy
Ny AE cs
=

2
2
cs QFx
FFx L c cs c

2
cs
s 2 QFy
cs s
FFy
Determine the global stiffness for each member using Eqn. (9).
Element 1
c1 = 1, s1 = 0, L1 = 3
1

0.333 0 0.333 0 1
0
0
0
0 2

AE
0.333 cs 0.333 0 3

0
0
0 4
0

Redzuan Abdullah, FKA, UTM

Chapter1 page 21

Chapter 1 Review of Matrix Method

Element 2
c1 = 0.6,

s1 = 0.8,

L1 = 5
5

0.072 0.096 0.072 0.096 1

0.096 0.128 0.096 0.128 2

AE
0.072 0.096 0.072 0.096 5

0.096 0.128 0.096 0.128 6

By proper ordering of the number of DOF, the element stiffness coefficient can be
assembled to form the equilibrium equation for the truss system,

0
0.405 0.096 0.333
2

0

0.096 0.128
R3
0.333
0
0.333
= AE
0
0
R4
0
R5
0.072 0.096
0

0
R6
0.096 0.128

0 0.072 0.096 Q1
0 0.096 0.128 Q2
0
0
0 Q3

0
0
0 Q4
0 0.072 0.096 Q5

0 0.096 0.128 Q6

{Fk } = [ K11 ]{Qu } + [ K12 ]{Qk }

0
0.405 0.096 Q1
= AE
+0
2
0.096 0.128 Q2
4.505
Q1 1 0.405 0.096 0 AE
=

=
Q2 AE 0.096 0.128 2 19.003
AE
1

0 4.505
Q3
0.333
1.5

0

0 AE
Q4
0

0
AE
=
+
=

kN

Q
0.072
0.096
19.003
1.5

Q6

AE

0.096
0.128
2

Redzuan Abdullah, FKA, UTM

Chapter1 page 22

Chapter 1 Review of Matrix Method

Frame Element

Fig. 15 Frame element in arbitrary direction with local forces and DOF

Fig. 16 2D frame element in arbitrary direction, with end forces in and DOF in
global direction

The end forces, fNx and fFx in the axial direction are similar to the end forces in the
truss element. The forces in other DOFs can be calculated by applying one unit

Redzuan Abdullah, FKA, UTM

Chapter1 page 23

Chapter 1 Review of Matrix Method

displacement at a time at each DOF while locking other DOF from moving. The
forces are calculated using classical method such as slope deflection method.1

Fig. 17 End forces in a frame due to movement of DOF 2, qNy

Fig. 18 End forces in the frame element due to movement of DOF 5, qFy

Slope deflection eqn. M N = 2

Redzuan Abdullah, FKA, UTM

EI

2 N + F 3 + FEM N
L
L
Chapter1 page 24

Chapter 1 Review of Matrix Method

Fig. 19 End forces in the frame element due to movement of DOF 3, qNx

Fig. 20 End forces in the frame element due to movement of DOF 6, qFz

All forces in Fig. 7(c), and Fig. 17 to Fig. 20 can be written in the matrix form as,

Redzuan Abdullah, FKA, UTM

Chapter1 page 25

Chapter 1 Review of Matrix Method

f Nx ' AE

f Ny ' 0

f
Nz ' 0

=
f Fx ' AE

f Fy ' 0

f Fz ' 0

12 EI
L3
6 EI
L2

6 EI
L2
4 EI
L

12 EI
L3
6 EI
L2

6 EI
L2
2 EI
L

AE
L
0
0

AE
L
0
0

0
12 EI
L3
6 EI
L2

0
12 EI
L3
6 EI
2
L

q
Nx '

6 EI q
Ny '
L2

2 EI
qNz '

0 qFx '

6 EI
2 qFy '
L

4 EI

q
L Fz '
0

(26)

Transformation of the displacement and forces into the global

coordinate system
For a member without loads between nodal points,

f = k 'q

(27)

This is an equilibrium equation for a member in local coordinate. We need to

transform this equation into global coordinate so that all member stiffness of the
frame system can be assembles in a similar manner as in Eqn. 20.

Displacement Transformation Matrix

Fig. 21 Transformation of displacement components from global to local

directions

(Note: D = Q = Global direction displacement, d = q = local direction displacement)

Redzuan Abdullah, FKA, UTM

Chapter1 page 26

Chapter 1 Review of Matrix Method

At the near node, (Fig. 21(a)):

qN x ' = QN x cos x ;

qN y ' = QN x cos y

(28)

qN y ' = QN y cos x

(29)

qN x ' = QN y cos y ;

qN z ' = QN z

(30)

At the far node, (Fig. 21 (b));

qFx ' = QFx cos x ;

qFy ' = QFx cos y

qFx ' = QFy cos y ;

qFy ' = QFy cos x

(31)

qFz ' = QFz

Hence, the local displacement is related to the global displacement by:
qNx ' c
q s
Ny '
qNz ' 0

=
q
Fx
'

0
qFy ' 0

qFz ' 0

s 0
c 0
0 1
0 0
0 0
0 0

0 0 QNx
0 0 0 QNy
0 0 0 QNz

c s 0 QFx
s c 0 QFy

0 0 1 QFz
0

(32)

q = TQ
where T = displacement transformation matrix. It transforms the global
displacements to the local displacements.

Redzuan Abdullah, FKA, UTM

Chapter1 page 27

Chapter 1 Review
w of Matrix Method

Force Trransformattion Matrixx

Fig. 22 Transforma
T
ation of elem
ment axial and shear forces at near node to
o global
directtion
(Note: Q = F = force in global direction,
d
q = f = force in
i local direection)

FN x = f N x ' cos x ;

FN y = f N x ' cos y

FN x = f N y ' cos y ;

(33)

FN y = f N y ' cos x

(34)

FN z = f N z '

(35)

mber loads in the locall

Similarlyy done for thhe far node,, we can obttain the mem
coordinaate related too the global direction by:
b

FNx c s
F s c
Ny
FNz 0 0
=
FFx 0 0
FFy 0 0

FFz 0 0

0 f Nx '
0 f Ny '
0 f Nz '

0 c s 0 f Fx '
0 s c 0 f Fy '

0 0 0 1 f Fz '

0 0
0 0
1 0

0
0
0

(36)

F = TT f
Now we want to relaate the globbal loading, F to the gloobal displaccement, Q.
Substitutte Eqn. (32)) into (43),

f = k'TTQ

(37)

Substitutte Eqn. (37)) into Eqn. (36),

(

F = TTk'TQ
k

Redzuan Abdullah,
A
FKA
A, UTM

Chapterr1 page 28

Chapter 1 Review of Matrix Method

or

F = kQ

where

(38)

k = TT k'T

Expanding Eqn. (38), we get

Nx

Ny

Nz

Fx

Fy

Fz

AE 2 12 EI 2
6 EI
6 EI N
AE 12 EI
AE 2 12 EI 2
AE 12 EI
c + 3 s
3 cs
2 s
3 cs 2 s x

L c + L3 s
L
L
L
L
L
L
L

AE 12 EI
N
AE 2 12 EI 2
AE
EI
AE
EI
EI
6 EI
12
12
6

s + 3 c
c
s2 + 3 c2
c y
3 cs

3 cs

2
2
L
L
L
L
L
L
L
L
L

6 EI
6 EI
2 EI N
6 EI
6 EI
4 EI
s
c
2 s
2 c

z
L
L2
L
L
L
L2
k=
6 EI
6 EI
AE 2 12 EI 2
AE 2 12 EI 2
AE 12 EI
AE 12 EI
s
c + 3 s
s F
3 cs

L c + L3 s L L3 cs
L2
L
L
L
L
L2 x

AE 12 EI
6 EI
6 EI
AE 2 12 EI 2
AE 2 12 EI 2
AE 12 EI
s + 3 c 2 c
s + 3 c
3 cs
3 cs
2 c Fy

L
L
L
L
L
L
L
L

L
L

2 EI
6 EI
6 EI
4 EI
6 EI
6 EI
c
s
2 c
2 s

L
L2
L
L Fz
L
L2

(39)

and

FNx
QNx
F
Q
Ny
Ny
F
Q
F = Nz and Q = Nz
FFx
QFx
FFy
QFy

FFz
QFz

Beam-member global stiffness matrix

Beam is a special case of frame where the member is horizontal. For beams, axial
displacement is negligible: Row and column 1 and 4 of Eqn. (39) that associate
with axial DOF are remove, and also the axial stiffness (AE/L) are removed.
Local and global coordinates are coincided (c = 1, s = 0). Therefore,

k = TT k'T

Redzuan Abdullah, FKA, UTM

Chapter1 page 29

Chapter 1 Review of Matrix Method

Ny

Nz

12 EI
L3

6 EI
L2
k=
12 EI
L3
6 EI

L2

6 EI
L2
4 EI
L
6 EI
L2
2 EI
L

Fy
12 EI
L3
6 EI
L2
12 EI
L3
6 EI
2
L

Fz
6 EI N y
L2

2 EI N
z
L
6 EI
2 Fy
L
4 EI

L Fz

(40)

where global k is the same as local k (refer Eqn. (26) with 1st and 4th rows and
columns removed).
For beams with rotational DOF only (no support settlement and overhanging), the
above equation can be further reduced to:
Nz

Fz

4 EI

k= L
2 EI
L

2 EI N
z
L

4 EI
F
L z

(41)

After member global stiffness matrices are obtained, we assemble it to form

system stiffness matrix,
n

K = ki

(42)

i=1

Matrix equation is written to obtain equilibrium at the nodes. For beam and frame,
the loads are applied along the length or somewhere between the nodes. In This
case, the beam is analyzed using equivalent nodal loads where the values are
similar to reverse values of the fix end forces.

Redzuan Abdullah, FKA, UTM

Chapter1 page 30

Chapter 1 Review of Matrix Method

Fig. 23 Equivalent loading for element with loads between nodes

Case (c) is solved using

Fa = KQ + F F
Fa F F = KQ

Q = K 1 Fa F F

(43)

where Fa is loads that are applied at nodes, and FF is internal forces is similar to the
reverse of the fixed end forces.
The internal forces of element a that are obtained from the analysis using Eq. (43)
must be added to the fixed end forces (case b) to yield final results for the element.

Redzuan Abdullah, FKA, UTM

Chapter1 page 31

Chapter 1 Review of Matrix Method

Example:
F1
(a)
F2F

F8F
(b)

RLF

RRF
F8F

F2F

F1

(c)

Fig. 24
Analysis:
3
4

(d)

Fig. 25

x11 x12 x13 x14 3

x 22 x 23 x 24 4

k1 =
sym
x33 x34 5

x 44 1

Redzuan Abdullah, FKA, UTM

5
y11

k2 =
sym

1
y12
y 22

y13 y14 5
y 23 y 24 1
y33 y34 6

y 44 2

Chapter1 page 32

Chapter 1 Review of Matrix Method

z11 z12 z13 z14 6

z 22 z 23 z 24 2

k3 =
sym
z 33 z 34 7

z 44 8

1
x 44 + y 22

K = k1 + k2 + k3 =

symm

y 24

0 1
0
0
y14
y34
z 23 z 24 2
x11 x12
x13
0
0
0 3

x 22
x 23
0
0
0 4
x33 + y11
y13
0
0 5

y33 + z11 z13 z14 6

z 33 z 34 7

z 44 8

x14 x 24

y 44 + z 22

5
x34

y 23

System equilibrium equation, (Eqn. (43))

KQ = F a F F
K11
K
21

K12
K 22

Qu Fk
=
Qk Fu

Displacement vector,

Q1
Q
2
Q3

Q u
Q4
Q =

k
Q5
Q6

Q7
Q
8

Redzuan Abdullah, FKA, UTM

Chapter1 page 33

Chapter 1 Review of Matrix Method

Force vector,

F1 0
0 F F
2
0 0

0 0
F
a
F + ( F ) = +

0 0
0 0

0 0
F
0 F2

[ K11 ]{Qu } = {Fk }

x 44 + y 22
y 24

y 24 Q1 F1
=

y 44 + z 22 Q2 F2F

Q1 x 44 + y 22
=
Q2 y 24

y 24 F1

y 44 + z 22 F2F

After nodal deflections are known, the member forces can be calculated using
element equilibrium equation (Eq. (27).

f = k'q
where local displacement is calculated from global displacement using Eqn. (32),

q = TQ .

F
For element 3 where subjected to force between nodes, f = k'q + f

7
8

2
3

f6
z11 z12 z13 z14
f

z 22 z 23 z 24
2 EI
3
f = =
z33 z 34
f 7 L sym

f8
z 44

Redzuan Abdullah, FKA, UTM

c
0

0
1
0
0

0
0
c
0

0 0 RL
0 Q2 F2F
+
0 0 RR

1 0 F8F

Chapter1 page 34

Chapter 1 Review of Matrix Method

Note that the transformation matrix is actually an identity matrix (c = 1) because the
local and global coordinate for beam is coincide.

Redzuan Abdullah, FKA, UTM

Chapter1 page 35

Chapter 1 Review of Matrix Method

Example
Determine the moment developed at support A and B (Hibbeler pg 546)

||

The problem of Fig. (a) is solved by solving Fig. (b) using stiffness matrix method.
The load applied to Fig. (b) is the reversed of the fixed end forces in Fig. (c). The
answer of the analysis of Fig. (b) will be superimposed with the fixed end forces of
Fig. (c) to obtain the final results.

Redzuan Abdullah, FKA, UTM

Chapter1 page 36

Chapter 1 Review of Matrix Method

Member 1:
12 EI 12 ( 200 )( 216 )
=
= 2400
L3
63
6 EI 6 ( 200 )( 216 )
=
= 7200
L2
62

4 EI 4 ( 200 )( 216 )
=
= 28800
L
6
2 EI 2 ( 200 )( 216 )
=
= 14400
L
6
4

2400 7200 2400 7200 4

7200 28800 7200 14400
3
k1 =
2400 7200 2400 7200 5

7200 14400 7200 28800 2

Member 2:
12 EI 12 ( 200 )( 216 )
=
= 64800
L3
23
6 EI 6 ( 200 )( 216 )
=
= 64800
L2
22

4 EI 4 ( 200 )( 216 )
=
= 86400
L
2
2 EI 2 ( 200 )( 216 )
=
= 43200
L
2
5

64800 64800 64800 64800 5

64800 86400 64800 43200
2
k2 =
64800 64800 64800 64800 6

64800 43200 64800 86400 1

-FF = KQ
Fk K11
F = K
u 21
Redzuan Abdullah, FKA, UTM

K12 Qu
K 22 Qk
Chapter1 page 37

Chapter 1 Review of Matrix Method

43200
0
0
64800 64800 Q1
12 86400
84 43200 115200 14400 7200 57600 64800 Q

2
F3 0
14400 28800 7200 7200
0 0
=

7200
7200 2400 2400
0 0
F4 0
F5 64800 57600 7200 2400 67200 64800 0


64800 64800 0
0
0
F6 64800 64800

Solving Fk = K11Qu
12 = 86400Q1 + 43200Q2
84 = 43200Q1 + 115200Q2

Q1 = 0.2778 (103 ) radian

Q2 = 0.8333 (103 ) radian

Thus
F3 = 0 + 14400 ( 0.8333 ) (10 3 ) m = 12kN.m

The actual moment at A must include the fixed-supported reaction of +96kN-m

shown in Fig. c, along with the calculated result for F3.
Thus

M AB = 12kN.m + 96kN.m = 108kN.m

The internal moment and shear at B can be determined by expanding f1 = k 1' d + f1F
of member 1;

f 4 2400 7200 2400 7200 0

96 102
f 7200 28800 7200 14400 0

3
96 108
3

(10 ) + 96 = 90
=
f 5 2400 7200 2400 7200 0

f 2 7200 14400 7200 28800 0.8333

96 72

Redzuan Abdullah, FKA, UTM

Chapter1 page 38

Chapter 1 Review of Matrix Method

And for member 2;

5

0
q5 64800 64800 64800 64800
24 60
q 64800 86400 64800 43200 0.8333 103 12 72

=
+
=
0
q6 64800 64800 64800 64800
24 12

3
q1 64800 43200 64800 86400 0.2778 10 12 0
The reaction forces are;
14400
Q3 0
96 108
Q 0

3
7200 0.2778 10 96 102
4
=
+ =

Q
64800
57600

0
8333
10
.
5

120 150

24 12
Q6 64800 64800

Example
Determine the loadings at the joints of the 2-member frame as shown.
Take for both members:
E = 200 GPa
6

I = 180(10 ) mm
4

A = 6000 mm

Redzuan Abdullah, FKA, UTM

Chapter1 page 39

Chapter 1 Review of Matrix Method

0 6
0 7
Dk =
0 8

0 9

20 1
0 2

Qk = 0 3

0 4
0 5

AE 6(103 )(200)(106 )
=
= 200(103 ) kN/m
L
6
12 EI 12(200)(106 )(180)(106 )
=
= 2(103 ) kN/m
L3
63
6 EI 6(200)(106 )(180)(106 )
=
= 6(103 ) kN
L2
62

4 EI 4(200)(106 )(180)(106 )
=
= 24(103 ) kNm
L
6
2 EI 2(200)(106 )(180)(106 )
=
= 12(103 ) kNm
L
6
Member 1

x =

60
00
= 1 y =
=0
6
6

Redzuan Abdullah, FKA, UTM

Chapter1 page 40

Chapter 1 Review of Matrix Method

We have:

Member 2

x =

66
=0
6

y =

6 0
= 1
6

We have:

The structure stiffness matrix is determined by assembling k1 and k2

The result is:

Redzuan Abdullah, FKA, UTM

Chapter1 page 41

Chapter 1 Review of Matrix Method

Expanding to determine the disp yields:

Solving we obtain:
3
D1 17.51(10 ) m
D 37.47(106 ) m

2
3

D3 = 2.505(10 ) rad

D
17.51(10
)
m
4

D
3
5 1.243(10 ) rad

Using these results, the support reactions are determined from eqn (1) as follows:
Q6 7.50 kN
Q

7 = 20 kN
Q8 7.50 kN

Q9 75 kNm
The internal loadings at node 2 can be determined by applying equation to member 1.
Thus,

Note the appropriate arrangement of the elements in the matrices as indicated by the
code numbers alongside the columns & rows.
Solving yields:

Redzuan Abdullah, FKA, UTM

Chapter1 page 42

Chapter 1 Review of Matrix Method

q4 0

6 7.50 kN
q5 0

q1 0

q 7.50 kN
2

q3 45 kNm

Redzuan Abdullah, FKA, UTM

Chapter1 page 43

Chapter 1 Review of Matrix Method

Problems
Q1.

Solve the truss in Fig. 1 using matrix stiffness method by hand calculation.
Determine the reaction and the forces in member 1, 2 and 3. Use Excel for
matrix arithmetic.

40 kN
2

20 kN

3m

3
1
2@3m= 6m
E = 200 GPa, A = 400 mm2

Fig. 1

Q2.

Using matrix stiffness analysis, analyze the beam in Fig. 2 by hand

calculation. Sketch the moment diagram for the beam. Assume EI as
constant.

1
1
8m

12 kN/m

2
12m

5m

Fig. 2

Redzuan Abdullah, FKA, UTM

Chapter1 page 44

Chapter 1 Review of Matrix Method

Q3.

Solve problem Q1 using LUSAS

a)

Plot the deflected shape and label the vertical and horizontal
displacement at each node. Note: Different plots are needed for
labeling the vertical and horizontal displacements.

b)

Plot the truss geometry and label axial forces in each truss member
and reaction forces at supports.

Q4.

Solve problem Q2 using LUSAS

a)

Plot the deflected shape and label the vertical displacement and
rotation at each node.

b)

Plot the bending moment diagram for the beam and label the values at
all nodes.

c)

Label the reaction forces at the supports.

Redzuan Abdullah, FKA, UTM

Chapter1 page 45