Thick Cylinders

February 13, 2024

February 13, 2024 1 / 46

Introduction
This problem is of considerable practical importance in pressure vessels
and gun barrels. A long hollow cylinder is subjected to uniformly
distributed internal and external pressure. There are two methods of
maintaining the pressure inside the cylinder:

1) By end caps;
2) By pistons in each end of the cylinder .

τr θ = 0 and σθ is constant at any particular radius due to the axial

X X
symmetry. p0 p 0

pi pi

X X
a) b)

February 13, 2024 2 / 46

Introduction
Difference in treatment between thin and thick cylinders

The theoretical treatment of thin cylinders assumes that the hoop stress is
constant across the thickness of the cylinder wall, and also that there is no
pressure gradient across the wall (σr = 0). Neither of these assumptions
can be used for thick cylinders for which the variation of hoop and radial
stresses shown below , their values being given by the Lamè equations:
B B
σθ = A + 2 and σr = A − 2
r ∂σr
r
σr + ∂r
dr

σθ σθ
σr

p0
ri
pi r0

February 13, 2024 3 / 46

Plane stress in cylindrical co-ordinate

There are certain cases such as cylinders, discs, curved bars, etc, in which
it is rather more convenient to use the cylindrical co-ordinates. In such a
case the coordinates are expressed in terms of r (radius), θ (angle) and z
(longitudinal coordinate, thickness, etc) coordinates. For simplicity we
assume that the stresses are constant along Z co-ordinate, which often is
the case.

Consider the element ABCD (see figure below), which is bounded by radial
lines OC and OD, subtending an angle dθ at the origin; and circular arcs
AB and CD at radii r and r + dr respectively. The element of thickness dz
is shown next.

February 13, 2024 4 / 46

Plane stress in cylindrical co-ordinate
∂σθ C Θ
σθ + dθ R
∂θ
∂τθr
τθr + dθ
∂θ ∂σr
σr + dr
B ∂r
dr
∂τrθ
τrθ + dr
∂r
r τrθ
σr
A D
dθ τθr
θ
O
dz

σθ
February 13, 2024 5 / 46
Plane stress in cylindrical co-ordinate
Considering equilibrium along the radial centre-line of the element, there
will be, in addition to forces from the radial stresses, the resolved
components of force from the hoop and shear stresses (see the figure).

∂τθr
τθr + dθ
∂θ
dθ/2
∂σθ
σθ τθr σθ + dθ
dθ/2 ∂θ

dθ

O
February 13, 2024 6 / 46
Plane stress in cylindrical co-ordinate

 
∂σr
σr + dr (r + dr )dθ dz − σr rdθ dz
∂r
 
∂σθ dθ dθ
− σθ + dθ dr dz sin − σθ dr dz sin
∂θ 2 2
 
∂τθr dθ dθ
+ τθr + dθ dr dz cos − τθr dr dz cos + R r dθ dr dz = 0
∂θ 2 2
1 1 1
As dθ −→ 0, sin dθ −→ dθ and cos dθ −→ 1, also neglecting second
2 2 2
and higher-order terms and dividing by r dr dθ dz, the equation reduces to
σr ∂σr σθ 1 ∂τθr
+ − + +R=0
r ∂r r r ∂θ
or
∂σr 1 ∂τθr σr − σθ
+ + +R=0
∂r r ∂θ r
February 13, 2024 7 / 46
Plane stress in cylindrical co-ordinate

Resolving in the tangential direction,

 
∂σθ dθ dθ
σθ + dθ dr dz cos − σθ dr dz cos
∂θ 2 2
 
∂τθr dθ dθ
+ τθr + dθ dr dz sin + τθr dr dz sin
∂θ 2 2
 
∂τr θ
+ τr θ + dr (r + dr )dθ dz − τr θ rdθ dz + Θ r dθ dr dz = 0
∂r

Applying the same simplification techniques as for the previous equation

we arrive at:
1 ∂σθ ∂τr θ τθr
+ +2 +Θ=0
r ∂θ ∂r r
Similarly, R, and Θ are volumetric forces arising from gravity, inertia, etc.

February 13, 2024 8 / 46

Plane stress in cylindrical co-ordinate

Axial symmetry. In certain cases, such as ring, disc or cylinder, the body
is symmetrical about a central axis z. Then the stress components depend
only on radius r , and σθ at any particular radius is constant. Also τθr = 0.
Only one equation remains:
dσr σr − σθ
+ +R=0
dr r
Where R is the body force and in most cases is zero.

February 13, 2024 9 / 46

Development of the Lamè theory
The equation of equilibrium for an element of material in a cylindrical
system of coordinates wit a symmetrical distribution of the stresses is
given as
dσr σr − σθ
+ =0 (1)
dr r
The above equation can also be written as
dσr
σθ − σr = r (2)
dr
Assuming now that plane sections remain plane, i.e. the longitudinal stress
εz is constant across the wall of the cylinder, viz.,
1
εz = [σz − ν(σr + σθ )] = constant (3)
E
It is also assumed that the longitudinal stress σz is constant across the
cylinder walls at points remote from the ends. Therefore, we can also say
that the sum of σr and σθ is also a constant.
February 13, 2024 10 / 46
Development of the Lamè theory
Based on this, let us make the following assumption:
σr + σθ = constant = 2A (4)
Substituting in (2) for σθ ,
dσr
2A − σr − σr = r
dr
Multiplying through by r and rearranging,
dσr
2σr r + r 2 − 2Ar = 0
dr
Taking into account that
dσr d
2σr r + r 2 = (r 2 σr )
dr dr
we get
d
(σr r 2 − Ar 2 ) = 0
dr
February 13, 2024 11 / 46
Development of the Lamè theory

After integrating,

σr r 2 − Ar 2 = constant = −B (we assume)

Therefore, we arrive at
B
σr = A − (5)
r2
and from the equation (4)
B
σθ = A + (6)
r2
The equations (5) and (6) are called Lamè equations.

February 13, 2024 12 / 46

Development of the Lamè theory
Internal pressure only

In this case there are two known conditions of stress which enable the
Lameè constants A and B to be determined. by denoting ri as an internal
and ro external radii, respectively, we have

at r = ri , σr = −p; at r = ro , σr = 0

where p is the internal pressure.

The internal pressure is considered as a negative radial stress since it

will produce a compression of the cylinder walls.

Substituting the above conditions into equation (5),

B B
−p = A − and 0=A−
ri2 ro2

February 13, 2024 13 / 46

Development of the Lamè theory
Internal pressure only

We obtain the constants A and B:

pri2 pri2 ro2
A= and B =
r02 − ri2 ro2 − ri2

Finally,
pr 2 r 2 − ro2
 
B
σr = A − 2 = 2 i 2
r r0 − ri r2
pr 2 r 2 + ro2
 
B
σθ = A + 2 = 2 i 2
r r0 − ri r2
According to these equations the maximum stresses σr and σθ occur at
the inside radius.

February 13, 2024 14 / 46

Development of the Lamè theory
Longitudinal stress

Consider now the cross-section of a thick cylinder with closed ends

subjected to an internal pressure pi and an external pressure po . For
longitudinal equilibrium:

pi × πri2 − po × πro2 = σz × π(ro2 − ri2 )

where σz is the longitudinal stress set up in the cylinder walls, thus we get

pi ri2 − po ro2
σz = (7)
r02 − ri2

which is a constant. It can be shown that the constant (σz ) has the same
value as the constant A of the Lamè equations. For combined internal and
external pressures, the relationship σz = A is also valid.

February 13, 2024 15 / 46

Development of the Lamè theory
Change of cylinder dimensions

Change of diameter = circumferential strain × original diameter. With

the principal stress system of circumferential, radial and longitudinal
stresses, all assumed tensile, the circumferential strain is given by (the
Hooke’s law):
1
εθ = (σθ − ν(σr + σz ))
E
Thus the change of diameter at any radius of the cylinder is given by
2r
∆D = [σθ − ν(σr + σz )] (8)
E
Change of length. Similarly, the change of length of the cylinder is given
by
L
∆L = [σz − ν(σr + σθ )] (9)
E

February 13, 2024 16 / 46

Development of the Lamè theory
Change of cylinder dimensions

Example 1. The cylinder of a hydraulic jack has a bore (internal

diameter) of 150mm and is required to operate up to 13.8MN/m2 .
Determine the required wall thickness for a limiting tensile stress in the
material of 41.4 MN/m2 .

Solution. The given boundary conditions are that at r = 75 × 10−3 m,

σr = −13.8 × 106 Pa and σθ = 41.4 × 106 , since the maximum hoop stress
occurs at the inner surface.
B
−13.8 × 106 = A −
5625 × 10−6
And
B
41.4 × 106 = A +
5625 × 10−6

February 13, 2024 17 / 46

Development of the Lamè theory
Change of cylinder dimensions

Solving the system of two equations

A = 13.8 × 103 ; B = 154

At the outside surface, σr = 0, therefore

B
0=A− =⇒ r = 0.106
r2
And the wall thickness must be at least

thickness = 0.106 − 0.075 = 0.031m = 31mm

February 13, 2024 18 / 46

 Shrinkage or interference allowance
In the design of compound cylinders it is important to relate the difference
in diameter of the mating cylinders to the stresses this will produce. This
difference in diameter at the “common” surface is normally termed the
shrinkage or interference allowance whether the compound cylinder is
formed by a shrinking or a force fit procedure respectively. Normally,
however, the shrinking process is used, the outer cylinder being heated
8 until it will freely slide over the innerofcylinder
Mechanics thus exerting the required g10.11
Materials
junction or shrinkage pressure on cooling.

Final common radius

Originol I D of
outer cylinder

/ February 13, 2024 19 / 46

Shrinkage or interference allowance
Stresses set up by a shrink-fit assembly. A shrink fit between two
components is a very important and secure method of assembly. It makes
the distribution of stresses through the thickness in more uniform matter.
We consider the simplest case, – two cylindrical components. In this case
the equations for the stresses can be written as
B B

σθ = A + 2 ; σr = A − 2 (inner component)  
r r (10)
D D
σθ = C + 2 ; σr = C − 2 (outer component) 

r r
where the constants A, B, C , and D are determined from the boundary
conditions, which are as follows: σr = 0 at the inside of the inner cylinder
and outside of the outer cylinder,and at the mating surface rm the radial
stress in each vessel must be the same:
B D
A− 2 =C − 2 (11)
rm rm
February 13, 2024 20 / 46
Shrinkage or interference allowance

Finally, at the mating surface the radial interference δ is the sum of the
displacement of the inner cylinder inwards, −u ′ , and the outer cylinder
outwards, +u ′′ , thus (ε = u/r )

δ = −u ′ + u ′′ = rm (ε′′θ − ε′θ ) (12)

Next we substitute into equation (12) the expressions for ε′θ and ε′′θ

σθ′ ν′ ′
ε′θ = − σ
E′ E′ r
σθ′′ ν ′′ ′′
ε′′θ = − σ
E ′′ E ′′ r
The relationship for the stresses are given above in (10). We now have
sufficient equations to solve for the constants A, B, C , and D.

February 13, 2024 21 / 46

Shrinkage or interference allowance
Example 2

A bronze bush of 25mm wall thickness is to be shrunk on to a steel shaft

of 100mm in diameter. If an interface pressure of 69MN/m2 is required,
determine the interference between bush and shaft. Steel E = 207GN/m2 ,
ν = 0.28; bronze E = 100GN/m2 , ν = 0.29.

Solution: Using constants A and B for the shaft and C and D for the
bush, then the radial stress for the shaft is
B
σrs = A −
r2
At r = 0 σrs = ∞, which is impossible, therefore B = 0 and σrs = A = σθs
at all points in the shaft. The boundary conditions are:

At the interface σrs = −69 × 106 = A

February 13, 2024 22 / 46

Shrinkage or interference allowance
Example 2

At rm = 50mm
D
σrb = C − = −69 × 106
0.0025
At r0 = 75mm
D
σrb = C − =0
0.0056
From which
D = 3.12 × 105 and C = 5.55 × 107
So at rm = 50mm σrb = −69MN/m2 and σθb = 180MN/m2 .

Now the interference is

δ = −us + ub = rm (εθb − εθs ) where rm = 50mm.
Substituting values for σθb , σrb , σθs and σrs we get δ = 0.112mm.
February 13, 2024 23 / 46
Shrinkage or interference allowance
Example 3

A thick cylinder of 100 mm internal radius and 150 mm external radius is

subjected to an internal pressure of 60 MN/m2 and an external pressure of
30 MN/m2 . Determine the hoop and radial stresses at the inside and
outside of the cylinder together with the longitudinal stress if the cylinder
is assumed to have closed ends.

Solution: the internal and external pressures both have the effect of
decreasing the thickness of the cylinder; the radial stresses at both inside
and outside radii are thus compressive, i.e. negative
at r = 0.1 m, σr = −60 MN/m2
at r = 0.15 m, σr = −30 MN/m2
Therefore,
−60 = A − 100B and − 30 = A − 44.5B
February 13, 2024 24 / 46
Shrinkage or interference allowance
Example 3

Solving the two equations yields

A = −6; B = 0.54

Therefore, at r = ri = 0.1 m we have (equation (6))

B
σθ = A + = −6 + 0.54 × 100 = 48 MN/m2 ←
r2
And at r = ro = 0.15 m we have

σθ = −6 + 0.54 × 44.5 = 18 MN/m2 ←

From equation (7) the longitudinal stress is given by

pi ri2 − p0 ro2 60 × 0.12 − 30 × 0.152

σz = = = −6 MN/m2 ←
ro2 − ri2 0.152 − 0.12

February 13, 2024 25 / 46

Shrinkage or interference allowance
Example 4

An external pressure of 10 MPa is applied to a thick cylinder of internal

diameter of 160 mm and external diameter 320 mm. If the maximum
circumferential stress is permitted on the inside wall of the cylinder is
limited to 30 MPa, what maximum internal pressure can be applied
assuming the cylinder has closed ends? What will be the change in outside
diameter when this pressure is applied? E = 207 GPa, ν = 0.29.

The conditions for the cylinder are:

1
r = 0.08 m; σr = −p; = 156
r2
1
r = 0.16 m; σr = −10 MPa; = 39
r2
And, also (maximum hoop stress inside the cylinder)
r = 0.08 m σθ = 30 MPa
February 13, 2024 26 / 46
Shrinkage or interference allowance
Example 4

Using the last two conditions:

−10 = A − 39B
30 = A + 156B
We get
A = −2 and B = 0.205
Therefore, at r = 0.08 m
σr = −p = A − 156B = −2 − 156 × 0.205 = −34 MPa ←
That is, the allowable internal pressure is 34 MPa.

From the equation (8) the change in diameter is given by

2r
∆D = [σθ − ν(σr + σz )]
E
February 13, 2024 27 / 46
Shrinkage or interference allowance
Example 4

Now at the outside surface

σr = −10 Mpa

and
B
σθ = A + = −2 + (39 × 0.205) = 6 MPa ←
r2
pi ri2 − p0 ro2 34 × 0.082 − 10 × 0.162
σz = = = −1.98 MPa
ro2 − ri2 0.162 − 0.082
Finally, the change in diameter
2 × 0.16
∆D = [6−0.29(−10+(−1.98))]×106 = 1.47×10−5 m = 14.7 µm
207 × 109

February 13, 2024 28 / 46

Shrinkage or interference allowance
Example 5

(a) In an experiment on a thick cylinder of 100 mm external diameter and

50 mm internal diameter the hoop and longitudinal strains gauges applied
to the outer surface of the cylinder were 240 × 10−6 and 60 × 10−6 ,
respectively, for an internal pressure of 90 MPa, the external pressure
being zero.

Determine the actual hoop and longitudinal stresses present in the cylinder
if E = 208 GPa and ν = 0.29.

(b) Assuming that the above strain readings were obtained for a thick
cylinder of 100 mm external diameter but unknown internal diameter
calculate this internal diameter.

February 13, 2024 29 / 46

Shrinkage or interference allowance
Example 5

Solution: Part (a).

1
εθ = (σθ − νσz )
E
1
εz = (σz − νσθ )
E
Since σr = 0 at the outside surface of the cylinder for zero external
pressure

240 × 10−6 × 208 × 109 = σθ − 0.29σz = 50 × 106

60 × 10−6 × 208 × 109 = σz − 0.29σθ = 12.5 × 106

Solving we obtain

σθ = 58.7 MPa σz = 29.5 MPa ←

February 13, 2024 30 / 46

Shrinkage or interference allowance
Example 5

Part (b).

From part (a) σθ = 58.7 MPa at r = 0.05 m and σr = 0 at r = 0.05 m.

Therefore,

58.7 = A + 400B and 0 = A − 400B

A = 29.35; B = 0.0734
Therefore, for the internal radius ri where σr = 90 MPa
0.0734
−90 = 29.35 − −→ ri = 0.00248 m = 24.8 mm ←
ri2

Internal diameter = 49.6 mm.

February 13, 2024 31 / 46

Shrinkage or interference allowance
Example 6

A compound cylinder is formed by shrinking a tube of 250 mm internal

diameter and 25 mm wall thickness onto another tube of 250 mm external
diameter and 25 mm wall thickness, both tubes being made of the same
material. The stress set up at the junction owing to shrinkage is 10 MPa.
The compound tube is then subjected to an internal pressure of 80 MPa.
Compare the hoop stress distribution now obtained with that of a single
cylinder of 300 mm external diameter and 50 mm thickness subjected to
the same internal pressure.

Solution: A solution is obtained by considering the effects of shrinkage

and internal pressure separately and combining the results algebraically.

February 13, 2024 32 / 46

Shrinkage or interference allowance
Example 6

Shrinkage only - outer tube:

r = 0.15, σr = 0 and r = 0.125, σr = −10 MPa

B
0=A− = A − 44.5B
0.152
B
−10 = A − = A − 64B
0.1252
After solving,
A = 22.85; B = 0.514
Thus,
r = 0.15 m : σθ = A + 44.5B = 45.7 MPa
r = 0.125 m : σθ = A + 64B = 55.75 MPa
February 13, 2024 33 / 46
Shrinkage or interference allowance
Example 6

Shrinkage only - inner tube:

r = 0.10, σr = 0 and r = 0.125, σr = −10 MPa

B
0=A− = A − 100B
0.12
B
−10 = A − = A − 64B
0.1252
After solving,
A = −27.8; B = −0.278
Thus,
r = 0.125 m : σθ = A + 64B = −45.6 MPa
r = 0.1 m : σθ = A + 100B = −55.6 MPa
February 13, 2024 34 / 46
Shrinkage or interference allowance
Example 6

Considering internal pressure only (on complete cylinder)

r = 0.15, σr = 0 and r = 0.10, σr = −80 MPa

0 = A − 44.5B
−80 = A − 100B

After solving,
A = 64.2; B = 1.44
Thus,
r = 0.15 m : σθ = A + 44.5B = 128.4 MPa
r = 0.125 m : σθ = A + 64B = 156.4 MPa
r = 0.1 m : σθ = A + 100B = 208.2 MPa
February 13, 2024 35 / 46
Shrinkage or interference allowance
Example 6

The resultant stress for combined shrinkage and internal pressure are then:

Outer tube:

r = 0.15 σθ = 128.4 + 45.7 = 174.1 MPa ←

r = 0.125 σθ = 156.4 + 55.75 = 212.15 MPa ←

Inner tube:

r = 0.125 σθ = 156.4 − 45.6 = 110.8 MPa ←

r = 0.10 σθ = 208.2 − 55.6 = 152.6 MPa ←

February 13, 2024 36 / 46

Shrinkage or interference allowance
Example 7

A compound tube is made by shrinking one tube of 100 mm internal

diameter and 25 mm wall thickness on to another tube of 100 mm external
diameter and 25 mm wall thickness. The shrinkage allowance, based on
radius, is 0.01 mm. If both tubes are of steel (with E = 208 GPa),
calculate the radial pressure set up at the junction owing to shrinkage.

Solution: Let p be the required shrinkage pressure, then for the inner
tube:
r = 0.025, σr = 0 and r = 0.05, σr = −p

B
0=A− = A − 1600B
0.0252
B
−p = A − = A − 400B
0.052

February 13, 2024 37 / 46

Shrinkage or interference allowance
Example 7

which leads to
4p p
A=− ; B=−
3 1200
Therefore, at the common radius the hoop stress is given by the equation
(6),
(i) B 4p  p  5p
σθ = A + = − + 400 − =−
0.052 3 1200 3
For the outer tube:

r = 0.05, σr = −p and r = 0.075, σr = 0

B
−p = A − = A − 400B
0.052
B
0=A− = A − 178B
0.0752
February 13, 2024 38 / 46
Shrinkage or interference allowance
Example 7

which leads to
178p p
A= ; B=
222 222
Therefore, at the common radius the hoop stress is given by
(o) B 178p p 578p
σθ =A+ = + 400 =
0.052 222 222 222
Now the shrinking allowance for the same material is
rm (o) (i)
δ= (σθ − σθ )
E
50 × 10−3 578p
  
−3 5p
0.01 × 10 = − − × 106
208 × 109 222 3
where p has units of MPa (×106 in the above equation)
p = 9.74 MPa ←
February 13, 2024 39 / 46
Shrinkage or interference allowance
Example 8

Two steel rings of radial thickness 30 mm, common radius 70 mm and

length 40 mm are shrunk together to form a compound ring. It is found
that the axial force required to separate the rings out, is 150 kN.
Determine the shrinkage pressure at the mating surfaces and the shrinkage
allowance. E = 208 GPa. The coefficient of friction between the junction
surfaces of the two rings is 0.15.

Solution: Let the pressure between the rings be p MPa. Then, normal
force between rings = p × 2πrL = N

= p × 106 × 2π × 70 × 10−3 × 40 × 10−3 = 5600πp newtons

Friction force between rings = µN

= 0.15 × 5600πp = 150 × 103 −→ p = 57 MPa ←

February 13, 2024 40 / 46

Shrinkage or interference allowance
Example 8

Now, for the inner tube:

r = 0.07, σr = −57 and r = 0.04, σr = 0

B
−57 = A − = A − 204B
0.072
B
0=A− = A − 625B
0.042
Solving the equations gives
A = −84.5; B = −0.135
Therefore, at the common radius the hoop stress in the inner tube is given
by
(i) B
σθ = A + = A + 204B = −112.1 MPa
0.072
February 13, 2024 41 / 46
Shrinkage or interference allowance
Example 8

For the outer tube:

r = 0.07, σr = −57 and r = 0.01, σr = 0

B
−57 = A − = A − 204B
0.072
B
0=A− = A − 100B
0.12
Solving the equations gives
A = 54.8; B = 0.548
Therefore, at the common radius the hoop stress in the outer tube is given
by
(o) B
σθ = A + = A + 204B = 166.8 MPa
0.072
February 13, 2024 42 / 46
Shrinkage or interference allowance
Example 8

Shrinkage allowance:
rm (o) (i) 0.07
(σθ − σθ ) = (166.8 − (−112.1)) × 106
E 208 × 109
= 93.8 × 10−6 = 0.094 mm ←

February 13, 2024 43 / 46

Shrinkage or interference allowance
Example 9

(a) A steel sleeve of 150 mm outside diameter is to be shrunk on a solid

steel shaft of 100 mm diameter. If the shrinkage pressure set up is 15
MPa, find the initial difference between the inside diameter of the sleeve
and the outside diameter of the shaft.

(b) What percentage error would be involved if the shaft were assumed to
be incompressible? For steel, E = 208 GPa, ν = 0.3.

Solution: Treating the sleeve as a thick cylinder with internal pressure 15

MPa,
r = 0.05, σr = −15 and r = 0.075, σr = 0
B
−15 = A − = A − 400B
0.052
B
0=A− = A − 178B
0.0752
February 13, 2024 44 / 46
Shrinkage or interference allowance
Example 9

Solving the equations gives

A = 12.05; B = 0.0676
Therefore, the circumferential stress in the sleeve at r = 0.05 m is given by
(o) B
σθ =A+ = A + 400B = 39 MPa
0.052
The shaft will be subjected to a hoop stress which will be compressive and
equal in value to the shrinkage pressure, i.e
(i)
σθ = −15 MPa
Thus the difference in radii or shrinkage allowance
rm (o) (i) 0.05
(σθ −σθ ) = (39−(−15))×106 = 13×10−6 m = 0.026 mm ←
E 208 × 109
February 13, 2024 45 / 46
Shrinkage or interference allowance
Example 9

(b) If the shaft is assumed incompressible, then the difference in diameter

will equal the necessary change in diameter of the sleeve to fit the shaft.
This can be found from the diameteral strain (assuming σz = 0),

2r 0.1
∆D = (σθ − νσr ) = (39 − 0.3 × (−15)) × 106
E 208 × 109
= 20.9 × 10−6 = 0.0209 mm

Percentage error:
0.026 − 0.0209
× 100 = 19.6%
0.026

February 13, 2024 46 / 46

Elementary Plasticity

January 17, 2024

January 17, 2024 1 / 53

Introduction
Engineering design is primarily concerned with maintaining machines and
structures working within their elastic range. However, it would be foolish
if designers knew nothing of what would happen to components that were
grossly overloaded where yielding and plastic deformation occurred.

Let analyse a stress-strain diagram for a structural steel in tension:

σ
ultimate stress

yield stress B C
E
proportional A
limit

II III IV

I
ε
linear perfect plasticity strain hardening necking
region (yielding)

January 17, 2024 2 / 53

Introduction

I Stress and strain are proportional. Beyond point A the proportionality

between stress and strain no longerexists. A is the proportional limit.

II In the region from B to C the material becomes perfectly plastic, which

means that it can deform without an increase in the applied load.

III After undergoing the large strains that occur during yielding in the
region BC , the steel begins to strain harden. During strain hardening the
material undergoes changes in its atomic crystalline structure, resulting in
increased resistance of the material to further deformation.

IV “Necking” and fracture (break).

January 17, 2024 3 / 53

Plastic bending of beams: plastic moment
In considering the behaviour of beams subjected to pure bending which
results in fibres being stressed beyond the limit of proportionality, the
following assumptions will be made:

1 The fibres are in condition of simple tension or compression.

2 Any cross-section of the beam will remain plane during bending as in
elastic bending.

In elastic bending of a beam there is a linear stress distribution over the

cross-section and when the extreme fibres reach the yield stress, the
bending moment is given by
I
MY = σY (1)
y
where I is the second moment of area about the neutral axis; y is the
distance from the neutral axis the extreme fibres; MY is a yielding bending
moment.
January 17, 2024 4 / 53
Plastic bending of beams: plastic moment

elastic
Z core
d

b σY σY σY

a) b) c)

January 17, 2024 5 / 53

Plastic bending of beams: plastic moment
Rectangular section

bd 2
MY = σY
6
and the stress distribution is shown in figure ( a). The state of the
cross-section in (b) is called elastic-plastic, whereas, (c) is fully plastic.

When the bending moment is increased above the value MY , some of the
fibres near the top and bottom surfaces of the beam begin to yield. With
further increase in bending moment, plastic deformation penetrates deeper
into the beam. The total bending moment is obtained by consideration of
both the plastic stress near the top and bottom of the beam and the
elastic stress in the core of the beam. This moment is called the
elastic-plastic bending moment. General expression for the moment:
Z
M = σydA where A is area

January 17, 2024 6 / 53

Plastic bending of beams: plastic moment
Rectangular section

a) elastic component: depth is (d − 2h), so

b(d − 2h)2
M = σY
6
b) plastic component of the moment:
 
d h
M = σY bh(d − h) {Second moment of area = bh × − × 2}
2 2

Hence
b(d − 2h)2 σy bd 2
  
h h
M = σY bh(d − h) + σY = 1+2 1− (2)
6 6 d d

January 17, 2024 7 / 53

Plastic bending of beams: plastic moment
Rectangular section

At a distance ( 12 d − h) from the neutral axis, the stress in the fibres has
just reached the value σY , then if R is the radius of curvature, we have
E ( 21 d − h) 1 σY
σY = or = 1
R R E ( 2 d − h)

The fully plastic moment (c ) is given by eqn. (2), putting h = 12 d, as

3 bd 2 bd 2 3
Mp = σY = σY = MY (3)
2 6 4 2
The ratio of Mp /MY is termed the plastic section modulus Zpl , for the
2
rectangular section it is given by bd4 (which is a geometric characteristic
of the cross-section and also called shape factor)

Mp = σY Zpl

January 17, 2024 8 / 53

Plastic bending of beams: plastic moment
I-section

Y σY

1/2b1 1/2b1

Z
d

d1
Stress
distribution

b σY

bd 3 b1 d13
 
2
MY = σY − (4)
12 12 d
In the fully plastic condition
bd 2 b1 d12
 
Mp = σY − (5)
4 4
January 17, 2024 9 / 53
Plastic bending of beams: plastic moment
I-section

And the shape factor is

3 1 − b1 d12 /bd 2
 
Mp
= ≈ 1.16 for standard rolled I section beam
MY 2 1 − b1 d13 /bd 3

January 17, 2024 10 / 53

Plastic bending of beams: plastic moment
Asymmetrical section

In the two previous cases the neutral axis in bending of the section
coincided with an axis of symmetry. If the cross-section is asymmetrical
about the axis of bending, then the position of the neutral axis must be
determined.
σY

C1

N N
h

C2

σY

January 17, 2024 11 / 53

Plastic bending of beams: plastic moment
Asymmetrical section

N—N neutral axis must divide the cross-section into two equal areas. If
C1 is the centroid of the area A1 , C2 the centroid of the area A2
(A = A1 + A2 ), h is the distance between C1 and C2 , then the fully plastic
moment is given by
1
Mp = σY Ah
2
This equation applies for any shape of cross-section.

January 17, 2024 12 / 53

Plastic bending of beams: plastic moment
Asymmetrical section: Example

The flange and web of the T-bar section in shown in the figure are each
12mm thick, the flange width is 100mm,and the overall depth of the
section is 100mm. The centroid of the section is at distance of 70.6mm
from the bottom of the web, and the second moment area Iz of the
section about a line through the centroid and parallel to the flange is
2.03 × 106 mm4 . determine the value of the shape factor.

Solution: Let m be the distance of the neutral axis NN from the top of
the flange, then

A1 = 100m and A2 = 100(12 − m) + (88 × 12)

A1 = A2 =⇒ m = 11.3

January 17, 2024 13 / 53

Plastic bending of beams: plastic moment
Asymmetrical section: Example

If n is the distance of the centroid of area A2 from the bottom of the web,
then n = 46.8mm. Therefore h = 47.55mm, so that
1
Mp = σY (12 × 188) × 47.55 = 53636σY
2
2.03 × 106
MY = σY = 28754σY
70.6
and the shape factor
Mp
= 1.87
MY

January 17, 2024 14 / 53

Plastic collapse of beams

The fully plastic bending moment developed in the preceding section was
due to the application of pure bending. A beam would therefore become
fully plastic at all cross-sections along the whole length once Mp was
reached. however, in practice pure bending rarely occurs and the bending
moment distribution varies depending on the loading conditions.

P
q

√
L/3 L/ 3

L L

January 17, 2024 15 / 53

Plastic collapse of beams

When a cross-section, such as those shown, reaches the fully plastic state,
it cannot carry any higher loading and the beam forms a hinge at that
cross-section. This is termed a plastic hinge. When or more plastic
hindhes occur such that the beam or structure becomes a mechanism then
this situation is described as plastic collapse..

In the first example in the figure the maximum bending moment is at the
centre and is PL/4. Therefore plastic collapse occurs for the single hinge
formation and
Pp L 4Mp
Mp = or Pp =
4 L

January 17, 2024 16 / 53

Plastic collapse of beams

Consider a cantilever propped at the free end and carrying a concentrated

load at mid-span, as shown below. P

C B A

L/2 L/2

Mp
Mp

In situation like this, plastic collapse will only occur if multiple hinges form.

January 17, 2024 17 / 53

Plastic collapse of beams
The sequence of events as follows:

1 Elastic behaviour occurs in a structural member until a plastic hinge

is formed at a section.
2 If rotation at this hinge results in diffusion of the load to other parts
of the structure or supports then additional load may be carried until
another plastic hinge is formed.
3 As each hinge forms the moment remains constant at the fully plastic
value irrespective of additional load or deformation.
4 When there is no remaining stable portion able to carry additional
load then collapse will occur.
5 The as whole, or in part, will form a simple mechanism at collapse.
6 The collapse load may be calculated by statical equilibrium if the
locations of the hinges can be identified.
January 17, 2024 18 / 53
Plastic collapse of beams

That is:

1 Beam will yield at A at some value of Ppl ;

2 As the load is increased beyond Ppl , yielding will also start to occur at
B.
3 Then plastic hinge forms at A, then at B. At this point the beam will
collapse.

The value of load Pp which causes plastic collapse may be determined

from a static equilibrium analysis of the beam because the moments at A
and B are known to be equal.

January 17, 2024 19 / 53

Plastic collapse of beams

Calculations:
X PL Pp Mp
MA = RC L − + Mp =⇒ RC = −
2 2 L
X RC L 6Mp
MB = 0; Mp − = 0 =⇒ Pp =
2 L
Elastic solution moment at A:
3 MA 16
MA = PL =⇒ P =
16 L 3
At the yield condition P = PY and M = MY . Using the above expressions
we get
Pp 9 Mp
=
PY 8 MY

January 17, 2024 20 / 53

Plastic collapse of beams

Example: Next consider the case of a beam fixed at both ends carrying a
uniformly distributed load.

If A is the left clamped end, then considering a half of the beam (L/2)
X qp L L 16Mp
MA = Mp + Mp − = 0 =⇒ qp =
2 4 L2
qL2 qp 4 Mp
Elastic moment = =⇒ =
12 qY 3 MY

January 17, 2024 21 / 53

Plastic collapse of simple beams
Alternative approach

There is a direct relationship between bending moment diagrams in

statically indeterminate beams and statically determinate beams. Study
these examples:

(3/16)PL

a P P
PL/4

x a

PL/4
x
(5/32)PL
PL/4

January 17, 2024 22 / 53

Plastic collapse of simple beams
Alternative approach

Let us now calculate the load Ppl which will cause the collapse this beam.

Mpl
0.5Mpl

P L/4
Mpl
Thus,
PL 3 6Mpl
= Mpl =⇒ Ppl =
4 2 L

January 17, 2024 23 / 53

Plastic collapse of simple beams
Alternative approach

Next consider the case of a beam fixed at both ends carrying a uniformly
distributed load q (see next page).

qL2 16Mpl
= 2Mpl =⇒ qpl =
8 L2
qL2
Where is the bending moment in a simply supported beam.
8
Mpl Mpl Mpl

qL2
8
Mpl

Mpl

January 17, 2024 24 / 53

Plasticity. Example 1

a) A rectangular-section steel beam, 50 mm wide by 20 mm deep, is used

as a simply supported beam over a span of 2 m with the 20 mm dimension
vertical. Determine the value of the central concentrated load which will
produce initiation of yield at the outer fibres of the beam.

b) If the central load is then increased by 10% find the depth to which
yielding will take place at the centre of the beam span.

c) Over what length of the beam will yielding then have taken place?

d) What are the maximum deflections for each load case?

For steel in simple tension and compression σY = 225 MPa and

E = 206.8 GPa.

January 17, 2024 25 / 53

Plasticity. Example 1
Solution: (a)The bending moment required to initiate yielding is
bd 2 50 × 202 × 10−9
MY = σY = × 225 × 106 = 750 Nm
6 6
But the maximum bending moment on a beam with a central point load is
PL/4, at the centre, viz.,
2P
= 750 −→ P = 1500 N
4
The load required to initiate yielding is 1500 N.

b) If the load is increased by 10% the new load is

P ′ = 1500 + 150 = 1650 N
The maximum bending moment is therefore increased to
2P ′ 2 × 1650
M′ = = = 825 Nm
4 4
January 17, 2024 26 / 53
Plasticity. Example 1

And this is sufficient to produce yielding to a depth of h. Elastic-plastic

moment for a rectangular cross-section :

b(d − 2h)2 σy bd 2
  
h h
Mep = σY bh(d − h) + σY = 1+2 1−
6 6 d d

By introducing a notation de = d − 2h, the above expression can be

rewritten as:
b  2
3d − de2 σY

Mep =
12

January 17, 2024 27 / 53

Plasticity. Example 1

After substituting numerical values we get

50 × 10−3 × 225 × 106 

3 × 0.022 − de2 = 825 Nm

12
where de is the depth of the elastic core.

8.8 × 10−4 = 1.2 × 10−3 − de2 −→ de = 0.01789m = 17.9 mm

Thus, the depth of yielding

h = 0.5(20 − 17.9) = 1.05 mm ←

January 17, 2024 28 / 53

Plasticity. Example 1

c) With the central load at 1650 N the yielding will have spread from the
centre as shown below in the figure. At the extremity of the yield region, a
distance x from each end of the beam, the section will just have yielded at
the extreme surface fibres, i.e. the moment carried at this section will be
maximum elastic moment.

1m 1650N
x yielded area
11111
00000
00000
11111
00000
11111
00000
11111
180mm
1650/2 2m
1650/2

January 17, 2024 29 / 53

Plasticity. Example 1

Now the bending moment at the distance x from the support is

1650 bd 2 2 × 750
x= σY = 750 −→ x = = 0.91 m
2 6 1650
Therefore length of beam over which yielding has occurred

2 − 2 × 0.91 = 0.18 m = 180 mm ←

January 17, 2024 30 / 53

Plasticity. Example 1

d) For P = 1500 N the beam is completely elastic and maximum

deflection, at the centre, is given by the standard form of

PL3 1500 × 23 × 12
∆= = = 0.0363 m = 36.3 mm
48EI 48 × 206.8 × 109 × 50 × 203 × 10−12
With P ′ = 1650 N and the beam partially plastic, deflections are
calculated on the basis of the elastic core only,

P ′ L3 1650 × 23 × 12
∆′ = ′
= = 0.0556 m = 55.6 mm
48EI 48 × 206.8 × 109 × 50 × 17.93 × 10−12

January 17, 2024 31 / 53

Plasticity. Example 2

a) Determine the shape factor of a T-section beam of dimensions

100 × 150 × 12 mm as shown below.

b) A cantilever is to be constructed from a beam with the above section

and is designed to carry a uniformly distributed load over its complete
length of 2 m. Determine the maximum u.d.l. that the cantilever can carry
if yielding is permitted over the lower part of the web to a depth of
25 mm. The yield stress of the material of the cantilever is 225 MPa.

January 17, 2024 32 / 53

Plasticity. Example 2
Solution: Shape factor = (Fully Plastic Moment)/(Maximum Elastic
Moment)

100

111111
000000
8.475 σY

000000
111111
ȳ = 49.5

31
000000
111111
000000
111111
19
12
N.A.(pl)
000000
111111
000000
111111
000000
111111
h=82.025

N.A.(el)
000000
111111
[mm] 000000
111111
000000
111111
Stress distribution
150

(plastic)

000000
111111
119

000000
111111
12
000000
111111
000000
111111
59.5

000000
111111
000000
111111
000000
111111
σY

January 17, 2024 33 / 53

Plasticity. Example 2

To determine the maximum moment carried by the beam while completely

elastic we must first determine the position of the neutral axis (N.A.(el)).

Take moments of area about the top edge (see figure):

(100 × 12 × 6) + (138 × 12 × 81) = [(100 × 12) + (138 × 12)] ȳ

7200 + 134136 = (1200 + 1656)ȳ −→ ȳ = 49.5 mm

Therefore, the second moment of area:

12 × 1383 100 × 123

 
IN.A.(el) = + 138 × 12 × 31.52 + + 100 × 12 × 43.52
12 12
× 10−12 = 6.55634 × 10−6 m4

January 17, 2024 34 / 53

Plasticity. Example 2
Now from the simple bending theory the moment required to produce the
yield stress at the edge of the section (in this case the lower edge), i.e. the
maximum elastic moment, is
σY IN.A.(el) 6.55634 × 10−6
Mel = = σY = 0.065 × 10−3 σY
ymax 0.1005
When the section becomes fully plastic the neutral axis (N.A.(pl)) divides
the area of the cross section into the two equal parts. That is, if the
plastic N.A. is a distance ȳp above the base, then
1
12ȳp = (1200 + 1656) −→ ȳp = 119 mm
2
Next determine the centroids of the top and bottom sections. Top section
(measuring from the top edge):
Atop = 119 × 12 = 1428
January 17, 2024 35 / 53
Plasticity. Example 2
X
Atop ȳ = 100 × 12 × 6 + 19 × 12 × (12 + 19/2) = 12102
12102
ȳ = = 8.475 mm
1428
Bottom section (from the bottom edge):
119
ȳ = = 59.5 mm
2
The distance between the centroids:
h = 150 − 8.475 − 59.5 = 82.025 mm
Then the fully plastic moment is calculated as
1 1
Mpl = AhσY = 2856 × 10−6 × 82.025 × 10−3 σY = 1.171 × 10−4 σY
2 2
1.171 × 10−4
Shape factor = = 1.802 ←
0.065 × 10−3
January 17, 2024 36 / 53
Plasticity. Example 2
b) For this part of the question the load on the cantilever is such that
yielding has progressed to a depth of 25 mm over the lower part of the
web. First we need to determine the position of the neutral axis.
100 125 − y
F4 σY
y

F3

125−y
113 − y
σY

113−y
12 y

N.A.

12 Stress distribution
150

(elastic − plastic)
y

[mm] F2

1
0
0
1 F1
25

0
1
0
1
0
1 σY

Tensile (compressive) force above N.A. = Compressive (tensile) force below.

January 17, 2024 37 / 53
Plasticity. Example 2
First we calculate the forces acting on different areas. Force on yielded
area F1 = stress × area:
F1 = 225 × 106 × (12 × 25 × 10−6 ) = 67.5 kN
Force on elastic portion below N.A.:
225 × 106
F2 = × (12 × y × 10−6 ) = 1.35y kN
2
where y in millimetres. Force in the web above N.A.:
113 − y (113 − y )2
F3 = × (225 × 106 )(113 − y ) × 12 × 10−6 = 1.35 kN
2y y
Force in flange:
1 113 − y 125 − y 238 − 2y
 
F4 = + 225×106 ×100×12×10−6 = 135 kN
2 y y y
January 17, 2024 38 / 53
Plasticity. Example 2

Now for the resultant force across the section to be zero,

F 1 + F2 = F3 + F4

(113 − y )2 238 − 2y
67.5 + 1.35y = 1.35 + 135
y y
642.5y = 49370 −→ y = 76.8 mm
Substituting back,

F1 = 67.5; F2 = 103.7; F3 = 23; F4 = 148.1 kN

January 17, 2024 39 / 53

Plasticity. Example 2
The moment of resistance of the beam can now be obtained by taking the
moments of these forces about the N.A. Here, for ease of calculations, it is
assumed that F4 acts at the mid-point of the web. This, in most cases,
sufficiently accurate for practical purposes.
 
2y 2
Mep = F1 (y + 12.5) + F2 + F3 (113 − y ) + F4 {(113 − y ) + 6} × 10−3
3 3
= (6030 + 5312 + 554 + 6243) × 10−3 = 18.14 kNm

The maximum bending moment present on cantilever carrying a u.d.l. is

qL2
= 18.15 × 103 . The maximum uniformly distributed load can be
2
carried by the cantilever is then

18.15 × 103 × 2
q= = 9.1 kN/m ←
4
January 17, 2024 40 / 53
Plasticity. Example 3

A horizontal cantilever of length L is simply–supported at the same level at

the free end (propped beam) and is subjected to a uniformly–distributed
load q over the full span. Determine the location of plastic hinge and the
magnitude of the collapse load qp .

January 17, 2024 41 / 53

Plasticity. Example 3
Solution: The bending-moment diagram is shown below. As the uniform
load is increased in magnitude,
 a plastic hinge will first form at the fixed
3 2 9 2
end ql > gl .
8 128

A B
L
1 2
qL
3 8
L
8 9
qL2
128

January 17, 2024 42 / 53

Plasticity. Example 3

The beam becomes a simply–supported beam with a plastic hinge on one

end and a real hinge on the other. Subsequent increase in the load will
cause the bending–moment to grow larger. The process will continue until
the moment at some other point (e.g. xp from the left support) reaches
Mp and create another plastic hinge.

Because the end conditions in the beam have changed the location of the
plastic hinge will not coincide with the location of the maximum moment
3
at L (see the figure). Also, there is no symmetry (due to different nature
8
of the hinges at the ends) and the location of the plastic hinge cannot be
at the midspan.

January 17, 2024 43 / 53

Plasticity. Example 3
It is convenient to use here the virtual–work approach here.
L − xp
 
qp L  
θ(L − xp ) = Mp θ + θ +
2 xp
1
Note a factor of in the equation on the left. Loads that act on sections
2
xp and L − xp are respectively qp xp and qp (L − xp ). They can be regarded
as two concentrated forces.
1
qp L
2
xp L − xp
A B
θ(L − xp )

α1 θ
Real hinge Plastic hinge

α2
Plastic hinge

January 17, 2024 44 / 53

Plasticity. Example 3

In the above figure the angles are evaluated in terms of θ as

L − xp L − xp
α1 = θ; α2 = θ + α1 = θ + θ
xp xp

Therefore, the plastic moment can be obtained from the virtual–work

equation (see above):
 
1
(qp L) θ(L − xp )
2
Mp = 
L − xp

θ+θ+
xp

January 17, 2024 45 / 53

Plasticity. Example 3
Solving for xp by taking the derivative of Mp with respect to coordinate xp
and equating it to zero,
dMp
=0
dxp
we obtain the location of the plastic hinge (this is the point where the
shear force is equal to zero):

xp = 0.414L ←−

Substituting the value of xp into the equation for the virtual–work and
solving it for qp we obtain the magnitude of the collapse load:

11.657Mp
qp = ←−
L2
This value also can be obtained graphically using the bending moment in a
equivalent simply–supported beam.
January 17, 2024 46 / 53
Plasticity. Example 3

NB: This problem

X can be solved
X in a simple way by using the equilibrium
conditions MA = 0 and MB = 0:

qp xp2
Mp =
2
qp (L − xp )2
2Mp =
2
The solution of the above two equations is xp = 0.414L, which located the
plastic hinge, and Mp = 0.0858qp L2 .

January 17, 2024 47 / 53

Plasticity. Example 4

Find the collapse load P = Pp for the beam shown whose bending
moment capacity is Mp throughout.
Note: The principal of superposition does not apply to plastic problems!

P 2.5P

A
B
C D

L/2 L/4 L/4

Solution: Theoretically there are four possible different collapse

mechanisms with the plastic hinges formed at: ACB, ADB, ACD and
CDB. However, the last two scenarios seem to be unrealistic (although it
might be a good exercise to check them as well).

January 17, 2024 48 / 53

Plasticity. Example 4
We consider only the first two. Deflections sketched are greatly
exaggerated.

a) The collapse mechanism looks as follows

2.5P
θ P θ
∆2
∆1

2θ

θL θL
∆1 = ; ∆2 =
2 4
The virtual–work equation (assuming that P = Pp ):
θL θL 32 Mp
P + 2P.5 = Mp (θ + θ + 2θ) −→ P =
2 4 9 L
January 17, 2024 49 / 53
Plasticity. Example 4
In order to check if this collapse mechanism is possible we will plot the
bending moment diagram. None of the moments must exceed the
value of Mp . For this we consider the equilibrium of the beam section
CDB: 80 Mp
9 L
Mp

D Mp
L/4 L/4
QC RB

X L 80 Mp L 76 Mp
MC = −RB + 2Mp + = 0 −→ RB =
2 9 L 4 9 L
Thus,
76 Mp L 10
MD = − Mp = Mp
9 L 4 9
Since MD > Mp , the above collapse mechanism is impossible.

January 17, 2024 50 / 53

Plasticity. Example 4
b) The second possible scenario is shown next

P
θ 2.5P 3θ
∆2 ∆1

4θ

3 1
∆1 = θL; ∆2 = θL
4 2
The virtual–work equation (assuming that P = Pp ):

3 1 64 Mp
2.5P θL + P θL = Mp (θ + 3θ + 4θ) −→ P =
4 2 19 L
January 17, 2024 51 / 53
Plasticity. Example 4
Now consider equilibrium of the section DB:

Mp Mp

L/4
QD RB

L 8Mp
−RB + 2Mp = 0 −→ RB =
4 L
8Mp L 64 Mp L 17
MC = − Mp − 2.5 = Mp
L 2 19 L 4 19

January 17, 2024 52 / 53

Plasticity. Example 4
11111111111
00000000000 00000
11111
00000000000
11111111111
Mp

00000000000
11111111111 00000
11111
00000
11111 Mp

00000000000 [BMD]
11111111111 00000
11111
00000000000
11111111111
00000000000111111111
1111111111100000000011111111111
00000000000
000000
11111100000
11111
00000
11111
D
000000000
111111111
00000000000
11111111111
000000
111111
C B
A
000000000
111111111
00000000000
11111111111
000000
111111
000000000
111111111
00000000000
11111111111
000000
111111
00000000011111111111
111111111
00000000000
00000000000111111
000000
11111111111
17
19
Mp 000000
111111 Mp

 
64 32
Obviously, the smaller collapse load governs here < . The
19 9
second collapse mechanism is valid and can take place if the load P
64 Mp
reaches its collapse value of .
19 L
January 17, 2024 53 / 53
Strain Energy

January 18, 2024

January 18, 2024 1 / 66

Introduction
Method of strain energy

As an example we consider a structural member subjected to a known

tensile load P.

L
P P

L+∆

In determining the amount of work done by the load P, it must be

understood that the load is considered to be applied gradually and that
the deflection commences as soon as the load starts to come on the
member. For a linear elastic structure, the load–deflection diagram would
be the straight line shown in the figure – at the maximum load P the
deflection is ∆.
January 18, 2024 2 / 66
Introduction
Method of strain energy

dp

dδ

0 δ ∆

Suppose at an intermediate stage when the load on the structure has

reached a value of p, the deflection has reached δ.
January 18, 2024 3 / 66
Introduction
Method of strain energy

Suppose also that a further small increment of load dp causes the

displacement to increase by dδ. The work W done by the load during this
further increment of load is, to a first approximation, given by p × dδ,
which will be seen to equal the shaded area. It follows that the load
increases from zero to P, the work done by the load is equal to the area of
the triangle under the complete line, thus work done by load P
1
W = P∆
2

January 18, 2024 4 / 66

Introduction
Method of strain energy

By the law of the conservation of energy this work is stored in the

structure. From the earlier courses of the Strength of Materials we know
that
PL
∆=
AE
The strain energy U stored in a member subjected to an axial force P is

P 2L
U=
2AE
and it is clearly irrelevant whether P is a positive or negative value.

January 18, 2024 5 / 66

Introduction
Displacements using the method of strain energy

If a moment M moves through a small angle dθ, the moment remaining

constant during this rotation, the work done by the moment is Mdθ. If a
moment increases gradually from zero to M and during this application a
total rotation of θ occurs, the rotation at any stage being proportional to
the applied moment. P P
x

z dx
y dθ

M M
x

dx

January 18, 2024 6 / 66

Introduction
Displacements using the method of strain energy

The work done on this section by the bending moment (internal forces) is
1
dW = Mdθ
2
The following ratio is valid here
dθ M
=
dx EI
Thus,we can write
M
dθ = dx
EI
Substituting it into expression for the work done, we arrive at

1 M2
dW = dx
2 EI
January 18, 2024 7 / 66
Introduction
Displacements using the method of strain energy

The work done for the entire beam can be obtained by integration of this
expression. Obviously, the work done equals to the accumulated strain
energy. Thus, the strain energy stored in the element:
M2
Z
U=W = dx
2EI
L
(In contrast, the strain energy stored in a single member subjected to an
P 2L
axial load is simply ).
2AE
Often, both E and I are constant along the length of the beam and the
expression would become:
Z
1
U= M 2 dx
2EI
L

January 18, 2024 8 / 66

Introduction
Displacements using the method of strain energy

Example: Find the strain energy in the beam and the deflection and slope
at the point of the application of the concentrated load P.
P

x
A
B
L

P L3
3EI

P L2
2EI

Solution: The bending moment at a distance x from the end of the beam:
M = −Px
Z L
1 P 2 L3
U= P 2 x 2 dx =
2EI 0 6EI
January 18, 2024 9 / 66
Introduction
Displacements using the method of strain energy

Work done by the load

1
W = P∆
2
As the work done by the load equals the strain energy stored in the beam:

1 P 2 L3 PL3
P∆ = −→ ∆=
2 6EI 3EI
This deflection being in the direction of P. Since
M Px
dθ = dx = dx
EI EI
we have
L
PL2
Z
1
θ= Pxdx =
EI 0 2EI
January 18, 2024 10 / 66
Castigliano and Lagrange theorems
Consider an elastic system (a simply supported beam) which is subjected to
an arbitrary system of loads Q and some generalized concentrated load P.
Let us now calculate the strain energy of such a system. For this purpose,
as well as, for convenience, the following order of the loading is adopted.

P1 P
q
A B

M
∆P Q ∆P P

January 18, 2024 11 / 66

Castigliano and Lagrange theorems
First we apply the load P to the system. The displacement directly under
the load and in the same direction is denoted ∆PP . Hereinafter, the first
subscript denotes a location (under load P) while the second one denotes
the cause of this displacement (due to load P). Afterwards, we apply the
system of loads Q. As a result of an additional deformation the load P will
move further more by a value of ∆PQ . Thus, the total displacement of the
point directly under the load P is
∆P = ∆PP + ∆PQ (1)
Obviously, the total accumulated strain energy is equal to the work of the
external forces:
1
U = P∆PP + P∆PQ + UQQ (2)
2
where UQQ is the strain energy accumulated only due to the application of
the system of loads Q, which numerically equals the sum of individual
forces Q along their corresponding displacements.
January 18, 2024 12 / 66
Castigliano and Lagrange theorems
The second item in the expression (P∆PQ ) does not contain 1/2 because
the force P was doing work without changing its magnitude, it was already
there. Putting ∆PP = PδPP the expression for U can be rewritten as
1
U = P 2 δPP + P∆PQ + UQQ (3)
2
Here δPP is the displacement at the point of the application of P due to
the unit load (P = 1).

Differentiating (3) with respect to P, taking also into account that

∆P = ∆PP + ∆PQ , we arrive at
∂U
= PδPP + ∆PQ = ∆PP + ∆PQ = ∆P (4)
∂P
Therefore
∂U
∆P = (5)
∂P
January 18, 2024 13 / 66
Castigliano and Lagrange theorems
Castigliano (1875) theorem: Partial derivative of the strain energy
with respect to a force gives the displacement corresponding to that
force

The second derivative of the strain energy is

∂2U ∂∆P
2
= = δPP (6)
∂P ∂P
and always positive.

To determine the linear or angular displacement at the point, where,

according to the problem requirements there is no force, it is necessary to
apply an imaginary generalized force. Then, writing down the expression
for the strain energy due to the system of forces, including the imaginary
force, differentiate the expression with respect to this imaginary force.
Finally, putting the imaginary force zero gives us the required displacement.
January 18, 2024 14 / 66
Castigliano and Lagrange theorems

By expressing the strain energy in terms of the independent displacements

∆1 , ∆2 , . . . , ∆n , it can be shown that a partial derivative with respect to
any of the displacements is equal to the force acting in the direction of
this displacement.
∂U
= Pi (7)
∂∆i
This theorem was proposed by Lagrange.

January 18, 2024 15 / 66

Modified Castigliano Theorem

We can solve for the deflection in one of two ways, either

Z L 2 
∂U ∂ M
∆= = dx (8)
∂P ∂P 0 2EI
or Z L
∂U M ∂M
∆= = dx (9)
∂P 0 EI ∂P
The last equation is called the Modified Castigliano theorem.

January 18, 2024 16 / 66

Displacements using the unit load method
The Modified Castigliano theorem can interpreted slightly different.
Z L
∂U M ∂M
∆= = dx
∂P 0 EI ∂P

After close examination we can see that

∂M
=M
∂P
Where M is the bending moment due to the unit load (P = 1). Thus, the
Modified Castigliano theorem can be given as
Z
Mp M
∆= ds
EI
Where Mp is the bending moment due to the given external load. By
using this method there is no need for dummy load.
January 18, 2024 17 / 66
Displacements using the unit load method

To determine a linear displacement at a particular point, a unit load

(P = 1) is applied at this point in the required direction of the
displacement (vertical, horizontal or at an angle).

To determine a slope, a unit bending moment (M = 1) is applied at

the point of interest.

January 18, 2024 18 / 66

Displacements using the unit load method
Example

Determine the deflection and slope at the free end of the cantilever.

q EI = const

A B
L
qL2
2
Mp

L
1 M1

1
M2
1

January 18, 2024 19 / 66

Displacements using the unit load method
Example

Solution: For the original loading, the bending moment Mp at the

distance x from the free-end is
qx 2
Mp = −
2
For the unit load placed at the position where the deflection is required
the bending moment M 1 on the same element is

M 1 = −1 · x = −x

Thus
ZL ZL
1 1 qx 3 qL4
∆A = Mp M 1 dx = dx =
EI EI 2 8EI
0 0

The positive value indicates that it is in the same direction as the unit load.
January 18, 2024 20 / 66
Displacements using the unit load method
Example

To determine the slope, we apply a bending moment equalled to one (unit

moment) at the free end, M 2 = 1.

ZL ZL
1 1 qx 2 qL3
θA = Mp M 2 dx = − dx = −
EI EI 2 6EI
0 0

The negative sign tell us that the direction of slope θA is opposite to the
unit moment, i.e. counter-clockwise.

January 18, 2024 21 / 66

Deflection of curved bars

It should be noted that there will be some strain energy in most cases due
to shear and axial forces but this is usually negligible in comparison with
that due to bending. In cases of twisting (torque), however, the strain
energy may be comparable with that due to bending.

If the torque at any point is T , likewise it is due to bending, the total

strain energy is calculated as

ZL
T2
U= dx
2GJ
0

E
where G = is the shear force modulus, and J = Ix + Iy is the
2(1 + ν)
polar moment of inertia. Next, we consider a few practical results.

January 18, 2024 22 / 66

Statically determinate structures
Example 1

A steel rod of uniform stiffness EI is bent into the shape of a semicircle

(arch), see the figure. One end is rigidly clamped, the other is free.
Determine the total deflection, and its direction, due to the horizontal load
applied to the free end.

R
θ
F

January 18, 2024 23 / 66

Statically determinate structures
Example 1

Solution: To do this, we should determine separately horizontal and

vertical components of the displacement. At first, we employ the
Castigliano theorem in its original form.

R sin θ
R
θ
F
R(1 − cos θ)
P0

January 18, 2024 24 / 66

Statically determinate structures
Example 1

Horizontal displacement: The expression for the bending moment at any

point C of the arch at the angle θ from the free end is given as
MC = FR sin θ
The total strain energy is
Zπ Zπ
1 1
U= MC2 ds = (FR sin θ)2 Rdθ
2EI 2EI
0 0

It must be noted that we integrate along the entire length of the arch and,
therefore, ds = Rdθ.
Zπ Zπ
1 2 3 2 F 2R 3 πF 2 R 3
U= F R sin θ dθ = sin2 dθ =
2EI 2EI 4EI
0 0

January 18, 2024 25 / 66

Statically determinate structures
Example 1

Finally, horizontal displacement will be

∂U πFR 3
∆h = = ←
∂F 2EI
Vertical displacement: To do this, we apply a dummy load P0 in the
vertical direction. In this case, the expression for the bending moment will
take the following form
MC = FR sin θ + P0 R(1 − cos θ)
The total strain energy is
Zπ
1
2
U = FR sin θ + P0 R(1 − cos θ) Rdθ
2EI
0
R3
3πP02 + 8FP0 + πF 2


=
4EI
January 18, 2024 26 / 66
Statically determinate structures
Example 1

The vertical component of the displacement is

∂U R3

∆v = = 6πP0 + 8F
∂P0 4EI
However, P0 = 0, thus the final expression for the vertical displacement
will be
2FR 3
∆v = ←
EI
The total displacement of the free end can be found as
s r
πFR 3 2 2FR 3 2 FR 3 π 2 2.543FR 3
  
∆= + = +4= ←
2EI EI EI 4 EI

January 18, 2024 27 / 66

Statically determinate structures
Example 1

2.543F R3
EI
2F R3
EI 4
arctan = 51.8◦
π

51.8◦

πF R3
2EI

NB: The evaluation of the above integrals can be quite cumbersome and
time consuming, which means that
this approach is probably not a best one.

January 18, 2024 28 / 66

Statically determinate structures
Example 1

Next, we repeat the calculations using the unit force method.

Horizontal displacement: Apply a unit force (=1) in the same point and
direction as the given force F . Then, the bending moment due to the unit
force can be written as
M C = R sin θ
and the horizontal component of the displacement is
Zπ Zπ
1 1 πFR 3
∆h = MC M C ds = FR 3 sin2 θ dθ = ←
EI EI 2EI
0 0

January 18, 2024 29 / 66

Statically determinate structures
Example 1

Vertical displacement: Apply a unit force in the vertical direction at the

free end:
M C = R(1 − cos θ)
Zπ Zπ
1
FR 3 2FR 3
∆v = (FR sin θ) R(1−cos θ) R dθ = sin θ(1−cos θ)dθ =
EI EI EI
0 0

Clearly the unit method is much more efficient here.

It is always advisable to estimate which method will be easier to

implement. It saves time and reduces the probability of getting an error in
the end.

January 18, 2024 30 / 66

Statically determinate structures
Example 2

A spring plate made of steel strip of rectangular section 20mm × 3mm is

curved to the shape shown in the figure. Determine the deflection at the
centre when P = 200N. Neglect friction. E = 200GN/m2 .

30◦ 30◦
R=50mm R=50mm

3mm
R=50mm

0000000000000000000
1111111111111111111
30◦ 30◦

January 18, 2024 31 / 66

Statically determinate structures
Example 2

Solution: For AB, taking the origin at A (see figure below):

R/2
C θ

P/2 B
R
θ R A

R/2
P/2

PR
M= sin θ = 0.025P sin θ
2
January 18, 2024 32 / 66
Statically determinate structures
Example 2

Zπ/6
1 P2
UAB = (0.025P sin θ)2 0.05dθ = 0.708 × 10−6
2EI EI
0
For BC, taking the origin at C
PR
M= (1 − sin θ) = 0.025P(1 − sin θ)
2
Zπ/6
1 P2
UBC = (0.025P(1 − sin θ))2 0.05dθ = 4.7 × 10−6
2EI EI
0
Total strain energy:
P2 P2
U = (0.708 + 4.7) × 10−6 × 2 = 10.816 × 10−6
EI EI
January 18, 2024 33 / 66
Statically determinate structures
Example 2

Finally, the displacement is

∂U 2 · 10.816 × 10−6 · 200

∆= = 3
= 4.8071 × 10−4 m = 0.481mm ←
∂P 0.02 · 0.003
200 × 109 ×
12

January 18, 2024 34 / 66

Statically determinate structures
Example 3

A thin bar is formed into an S-shape consisting of two semi-circles, as

shown in the figure. One end is (B) is firmly fixed to the wall while
another one (A) is free and subjected to both bending (MA ) and twisting
(TA ) moments. Taking into account that the bar has a uniform flexural
(EI ) and torsional (GJ) stiffnesses, estimate the angle of rotation (αA )
about x-axis (i.e calculate the slope) at A.

Top view
R x
B A MA x

R
y z
TA

January 18, 2024 35 / 66

Statically determinate structures
Example 3

Solution: It is enough to consider only one section (semi-circle) at one

time, then multiply the result by two. Let use here the Modified
Castigliano theorem. The notations shown in the figure for MA and TA are
in accordance with a right-hand thump rule.

Bending moment at any point,

∂M
M = MA cos θ − TA sin θ; = cos θ
∂MA
Torque,
∂T
T = −MA sin θ − TA cos θ; = − sin θ
∂MA

January 18, 2024 36 / 66

Statically determinate structures
Example 3

The first integral,

Zπ Zπ
M ∂M 2R πMA R
2× ds = (MA cos θ − TA sin θ) cos θ dθ = ←
EI ∂MA EI EI
0 0

The second integral,

Zπ Zπ
T ∂T 2R πMA R
2× ds = (−MA sin θ − TA cos θ)(− sin θ) dθ = ←
GJ ∂MA GJ GJ
0 0

The final answer will be

 
πMA R πMA R 1 1
αA = + = πMA R + ←
EI GJ EI GJ

January 18, 2024 37 / 66

Statically indeterminate structures

A structure is statically determinate if its reactions can be computed

from the equations of static equilibrium. In two-dimensional structures,
there are three equations of equilibrium:
X X X
Fx = 0; Fy = 0; M=0

A structure is statically indeterminate if its reactions cannot be

computed from the equations of static equilibrium alone (the number of
support reactions exceeds the number of independent equilibrium
equations), and the deformation conditions must be considered. The
superfluous reactions, those which are not necessary for static equilibrium,
are called redundant and the degree of statical indeterminacy is
defined by the number of redundant support reactions.

January 18, 2024 38 / 66

Statically indeterminate structures
Statically indeterminate structures are more common than statically
determinate. However, the analysis is more complex and internal forces
and stresses also depend on the stiffnesses (e.g. EA, EI and GJ) of their
members.

It is convenient to use the second Castigliano theorem for the analysis of

simple statically indeterminate structures, which states that
∂U
=0
∂R
if R is a support reaction or an internal force at any other point where the
displacement is known and equals zero.

In the case when a structure has more than one redundant, i.e., two,
three, or more times, statically indeterminate, it is convenient to obtain
the solution using the Force (Flexibility) method.
January 18, 2024 39 / 66
Statically indeterminate structures

In this method the redundant forces are removed and replaced by unknown
values of xi where i = 1, 2, 3, . . . , n and n is the number of redundant
forces.
δ11 x1 + δ12 x2 + δ13 x3 + · · · + δ1n xn + ∆1p = 0
δ21 x1 + δ22 x2 + δ23 x3 + · · · + δ2n xn + ∆2p = 0
δ31 x1 + δ32 x2 + δ33 x3 + · · · + δ3n xn + ∆3p = 0
............................................
δn1 x1 + δn2 x2 + δn3 x3 + · · · + δnn xn + ∆nn = 0

The above system of equations is symmetrical, i.e. δij = δji , and the
coefficients on the main diagonal δii are always positive. These coefficients
represent various displacements.

January 18, 2024 40 / 66

Statically indeterminate structures

X Z M2 X Z Mi Mj
i
δii = ds, δij = ds, i, j = 1, 2, 3, . . . , n
EI EI
l l
XZ M i Mp
∆ip = ds, i = 1, 2, 3, . . . , n
EI
l

where M i , M j , . . ., are bending moments due to the corresponding unit

forces xi = 1. Mp is the bending moment due to the external load.

After solving the above system the values of the sought redundant forces
are obtained.

January 18, 2024 41 / 66

Statically indeterminate structures

Finally, the following formulae might of good use when you calculate the
final values of the bending moment

MF = MP + M 1 x1 + M 2 x2 + . . . + M n xn

The final values of shear forces and torque can be calculated the same way.

January 18, 2024 42 / 66

Statically indeterminate structures
Example 4

The thin strip is fixed at C and free to slide without friction at A (see
figure below). Calculate the horizontal reaction Ax at A, when the strip
supports a vertical force FB at point B, as shown in the figure. EI is
constant.
C

B 45◦

FB
A Ax

January 18, 2024 43 / 66

Statically indeterminate structures
Example 4

Solution: This structure is one time statically indeterminate. The second

Castigliano theorem (modified form) is used for the analysis.

Consider separately sections AB and BC .

C
3π/4

θ
0 B 45◦
π/2

R
θ
FB
A Ax
0

January 18, 2024 44 / 66

Statically indeterminate structures
Example 4

∂MAB
MAB = Ax R(1 − cosθ), = R(1 − cos θ)
∂Ax
 
MBC = Ax R +Ax R sin θ−FB R(1−cos θ) = R Ax (1+sin θ)−FB (1−cos θ)

∂MBC
= R(1 + sin θ)
∂Ax
Using the second Castigliano theorem we have

Zπ/2 3π/4
Z
MAB ∂MAB MBC ∂MBC
∆A x = ds + ds = 0
EI ∂Ax EI ∂Ax
0 0

January 18, 2024 45 / 66

Statically indeterminate structures
Example 4

First integral:

Zπ/2 " #π/2

R3 2 R 3 Ax θ sin 2θ
Ax (1 − cos θ) dθ = θ − 2 sin θ + +
EI EI 2 4
0 0
3 3
 
R 3π R
= Ax −2 = (0.3562Ax )
EI 4 EI

January 18, 2024 46 / 66

Statically indeterminate structures
Example 4

Second integral:
3π/4 !
R3
Z  
Ax (1 + sin θ) − FB (1 − cos θ) 1 + sin θ dθ
EI
0
R3
  
θ sin 2θ
= Ax θ − 2 cos θ + −
EI 2 4
3π/4
sin2 θ
 
− FB θ − cos θ − sin θ −
2 0
3
  
R 9
= Ax π + 1.6642 + 2
EI 8
R3
 
3π
− FB − 0.25 + 1 = (7.1985Ax − 3.1062FB )
4 EI

January 18, 2024 47 / 66

Statically indeterminate structures
Example 4

Putting together

R3 R3
(0.3562Ax + 7.1985Ax − 3.1062FB ) = (7.5547Ax − 3.1062FB )
EI EI
Finally, the horizontal support reaction at A is
3.1062
Ax = FB = 0.411FB ←
7.5547

January 18, 2024 48 / 66

Statically indeterminate structures
Example 5

Determine the support reactions and plot the bending moment diagram in
the bent statically indeterminate beam shown in the figure (next page). It
is recommended to use the unit force method (Modified Castigliano
Theorem). The flexural stiffness EI is constant for the entire structure.
M0 A

30◦

January 18, 2024 49 / 66

Statically indeterminate structures
Example 5

Solution: It is convenient to use the support reactions at A as redundant

forces.

The system is two times statically indeterminate. Remove the support at

A:
M0 A X2

R(1 − cos θ) X1

R sin θ
R θ

January 18, 2024 50 / 66

Statically indeterminate structures
Example 5

The governing system of equations:

δ11 X1 + δ12 X2 + ∆1p = 0

δ21 X1 + δ22 X2 + ∆2p = 0

Bending moments in the primary structure when the supports are removed
and replaced by unknown forces, due X1 , X2 and M0 :

M 1 = R sin θ, M 2 = R(1 − cos θ), Mp = −M0

Displacements:

Zπ/2 Zπ/2
M 1M 1 M 1M 2
δ11 = ds, δ12 = δ21 = ds,
EI EI
0 0

January 18, 2024 51 / 66

Statically indeterminate structures
Example 5

Zπ/2 Zπ/2
Mp M 1 Mp M 2
∆1p = ds, ∆2p = ds
EI EI
π/6 π/6

Where ds = Rdθ. Note that the moment due to M0 is zero on the interval
from θ = 0 to θ = π/6. Calculation of displacements:

Zπ/2 Zπ/2
R3 R 3 θ sin 2θ π/2
 
1 2 2 2
δ11 = R sin θ R dθ = sin θ dθ = −
EI EI EI 2 4 0
0 0
πR 3 0.785398R 3
= =
4EI EI

January 18, 2024 52 / 66

Statically indeterminate structures
Example 5

Zπ/2 π/2
3 Z
1 R
δ22 = R 2 (1 − cos θ)2 R dθ = (1 − 2 cos θ + cos2 θ)dθ
EI EI
0 0

R3 θ sin 2θ π/2 R 3 3π 0.356194R 3

   
= θ − 2 sin θ + + = −2 =
EI 2 4 0 EI 4 EI

Zπ/2 Zπ/2
1 R3
δ12 = (R sin θ)(R(1 − cos θ))R dθ = (sin θ)(1 − cos θ)dθ
EI EI
0 0
Zπ/2 π/2
R3 R3 sin2 θ 0.5R 3

= (sin θ − sin θ cos θ)dθ = − cos θ − =
EI EI 2 0 EI
0

January 18, 2024 53 / 66

Statically indeterminate structures
Example 5

Zπ/2 Zπ/2
1 M0 R 2
∆1p = − M0 R sin θ R dθ = − sin θ dθ =
EI EI
π/6 π/6
√
M0R2 π/2 M0 R2 3 0.866025M0 R 2
− [− cos θ]π/6 = − =−
EI 2EI EI
Zπ/2 Zπ/2
1 M0 R 2
∆2p = − M0 R(1 − cos θ)R, dθ = − (1 − cos θ)dθ =
EI EI
π/6 π/6

R2 R2 0.547197M0 R 2
 
M0 π/2 M0 π 1
− [θ − sin θ]π/6 = − − =−
EI EI 3 2 EI

January 18, 2024 54 / 66

Statically indeterminate structures
Example 5

R3
Divide through by we have
EI
0.866025M0
0.785398X1 + 0.5X2 − =0
R
0.547197M0
0.5X1 + 0.356194X2 − =0
R
Solving the system gives us
1.172M0 0.109M0
X1 = ; X2 = −
R R

January 18, 2024 55 / 66

Statically indeterminate structures
Example 5

The bending moment diagram:

−0.4286M0 M0
A 0.109
R

M0
1.172
R
0.5714M0
30◦

B 0.0630M0

January 18, 2024 56 / 66

Statically indeterminate structures
Example 6

A thin ring is loaded by forces which are uniformly distributed along the
horizontal projection of the ring. Determine the decrease in the vertical
diameter.

January 18, 2024 57 / 66

Statically indeterminate structures
Example 6

Solution: The system is symmetrical, therefore, only one quadrant should

be considered. This significantly simplifies the solution. We take an
arbitrary quadrant:

C ds = Rdθ

R
θ M0

wR

January 18, 2024 58 / 66

Statically indeterminate structures
Example 6

The bending moment at any arbitrary point C is given by

wR 2
Mc = M0 − wR 2 (1 − cos θ) + (1 − cos θ)2
2
But M0 is unknown and is redundancy (the slope is zero at this point).
We use the second Castigliano theorem:

Zπ/2
∂U Mc ∂Mc
=0 or =0
∂Mo EI ∂M0
0

Obviously,
∂Mc
=1
∂M0

January 18, 2024 59 / 66

Statically indeterminate structures
Example 6

Thus,
Zπ/2
1 wR 2 
M0 − wR 2 (1 − cos θ) + (1 − cos θ)2 Rdθ = 0
EI 2
0

Simplify the expression under the integral:

Zπ/2
wR 3 wR 3 cos2 θ

1 3
M0 R − wR + + dθ
EI 2 2
0

Integrate, the factor 1/EI can be omitted:

π/2
wR 3 θ wR 3 θ sin 2θ
 
3
M0 Rθ − wR θ + + + =0
2 2 2 4 0

January 18, 2024 60 / 66

Statically indeterminate structures
Example 6

And after substituting the bounds (limits) of integration we arrive at

π π wR 3 π wR 3  π 
M0 R − wR 3 + + +0 =0
2 2 2 2 2 4
wR 3  π π π  2 wR 2
M0 = − − =
2 2 2 4 π 4
Finally, the bending moment at the current point C is given by

wR 2 wR 2
Mc = − wR 2 (1 − cos θ) + (1 − cos θ)2
4 2
After simplification:

wR 2 wR 2 wR 2 cos2 θ
Mc = − +
4 2 2
January 18, 2024 61 / 66
Statically indeterminate structures
Example 6

To calculate the change in the diameter we need to apply unit forces at

the top and bottom of the ring and, again, calculate the current moment
M c . We repeat the above derivations, but only for the unit loads:
R
Mc = M0 − (1 − cos θ)
2
Zπ/2 
1 R
M 0 − (1 − cos θ) Rdθ = 0
EI 2
0
from which
R R
M0 = −
2 π
and  
cos θ 1
Mc = R −
2 π
January 18, 2024 62 / 66
Statically indeterminate structures
Example 6

The desired displacement can be now calculated using the Modified

Castigliano theorem. We consider only one quadrant and the result
multiply by four:

Zπ/2 Zπ/2
wR 2 wR 2 wR 2 cos2 θ

4 4
∆d = Mc M c Rdθ = − +
EI EI 4 2 2
0 0
  
cos θ 1
× R − Rdθ
2 π

January 18, 2024 63 / 66

Statically indeterminate structures
Example 6

After multiplication the integral will take the following form

Zπ/2
4 wR 4 cos θ wR 4 wR 4 cos θ wR 4
∆d = − − +
EI 8 4π 4 2π
0
wR 4 cos3 θ wR 4 cos2 θ

+ − dθ
4 2π

After the integral evaluation we get

4 wR 4 wR 4 wR 4 wR 4

∆d = sin θ − θ− sin θ + θ
EI 8 4π 4 2π
wR 4 wR 4 sin3 θ wR 4 θ wR 4 sin 2θ

+ sin θ − − −
4 4 3 2π 2 2π 4

January 18, 2024 64 / 66

Statically indeterminate structures
Example 6

Substitute π/2 for θ:

4 wR 4 wR 4 wR 4 wR 4

∆d = − − +
EI 8 8 4 4
4 4 4 WR 4

wR wR wR
+ − − −0 = ←
4 12 8 6EI

January 18, 2024 65 / 66

Most common integrals used in calculations

Z
sin x dx = − cos x
Z
cos x dx = sin x
Z
x sin 2x
sin2 x dx = −
2 4
Z
x sin 2x
cos2 x dx = +
2 4
cos3 x
Z
sin3 x dx = − cos x +
3
Z 3
sin x
cos3 x dx = sin x −
3
sin2 x
Z
sin x cos x dx =
2
January 18, 2024 66 / 66
 Introduction to
Finite Element Method

Appendix

May 3, 2023

May 3, 2023 1/4

Appendix
Equivalent joint forces f0 for different types of load

Positive nodal force conventions: m1 m2

1 2
L
f1y f2y

P/2 P/2
P

L/2 L/2 P L/8 P L/8

P b2 (L + 2a) P a2 (L + 2b)
P L3 L3
a<b

P a2 b P ab2
a b
L2 L2

May 3, 2023 2/4

Appendix
Equivalent joint forces f0 for different types of load
P P P P

αL αL α(1 − α)P L α(1 − α)P L

q qL qL
2 2

L qL2 qL2
12 12
q
3qL 7qL
20 20

L qL2 qL2
30 20

May 3, 2023 3/4

Appendix
Equivalent joint forces f0 for different types of load
q
qL qL
4 4

L 5qL2 5qL2
96 96
parabolic load q
qL qL
3 3

L qL2 qL2
15 15
6Mab 6Mab
M L3 L3

a b Mb2 Ma
(2a − b) (2b − a)
L2 L2

May 3, 2023 4/4

 Introduction to
Finite Element Method

Part One

May 2, 2023

May 2, 2023 1 / 49
Introduction

The finite element method (FEM) is a numerical method for solving

problems of engineering and mathematical physics. Typical problem areas
of interest in engineering and mathematical physics that are solvable by
use of the finite element method include structural analysis, heat transfer,
fluid flow, mass transport, and electromagnetic potential.

Typically, for the structural stress-analysis problem, the engineer seeks to

determine displacements and stresses throughout the structure, which is in
equilibrium and is subjected to applied loads. For many structures, it is
difficult to determine the distribution of deformation using conventional
methods, and thus the finite element method is necessarily used.

May 2, 2023 2 / 49
Introduction

The finite element method involves modelling the structure using small
interconnected elements called finite elements.

A displacement function is associated with each finite element. Every

interconnected element is linked, directly or indirectly, to every other
element through common (or shared) interfaces, including nodes and/or
boundary lines and/or surfaces.

By using known stress/strain properties for the material making up the

structure, one can determine the behaviour of a given node in terms of the
properties of every other element in the structure. The total set of
equations describing the behaviour of each node results in a series of
algebraic equations best expressed in matrix notation.

May 2, 2023 3 / 49
Basic Steps of the Finite Element Method

1. Discretize and select the element types: This involves dividing the
body into an equivalent system of finite elements with associated nodes
and choosing the most appropriate element type to model most closely the
actual physical behaviour.

The total number of elements used and their variation in size and type
within a given body are primarily matters of engineering judgement. The
elements must be made small enough to give usable results and yet large
enough to reduce computational effort.

May 2, 2023 4 / 49
Basic Steps of the Finite Element Method

2. Select a displacement function: This step involves choosing a

displacement function within each element. The function is defined within
the element using the nodal values of the element. Linear, quadratic, and
cubic polynomials are frequently used functions because they are simple to
work with in finite element formulation.

However, trigonometric series can also be used. For a two-dimensional

element, the displacement function is a function of the coordinates in its
plane (say, the x − y plane). The functions are expressed in terms of the
nodal unknowns (in the two-dimensional problem, in terms of an x and a y
component). The same general displacement function can be used
repeatedly for each element.

May 2, 2023 5 / 49
Basic Steps of the Finite Element Method

Hence the finite element method is one in which a continuous quantity,

such as the displacement throughout the body, is approximated by a
discrete model composed of a set of piecewise-continuous functions
defined within each finite domain or finite element.

For the one-dimensional spring and bar elements, the displacement

function is a function of a single coordinate (say x, along the axis of the
spring or bar). For the spring and bar elements one can then skip this step
and start to derive the element stiffness matrix and equations, which is the
next step.

May 2, 2023 6 / 49
Basic Steps of the Finite Element Method

3. Define the strain-displacement and stress-strain relationships:

These relationships are necessary for deriving the equations for each finite
element. In the case of one-dimensional deformation, say, in the x
direction, we have strain εx related to displacement u by
du
εx =
dx
for small strains. In addition, the stresses must be related to the strains
through the stress-strain law – generally called the constitutive law. The
ability to define the material behaviour accurately is most important in
obtaining acceptable results.

May 2, 2023 7 / 49
Basic Steps of the Finite Element Method

The simplest of stress-strain laws,Hooke’s law, which is often used in

stress analysis, is given by
σx = E εx
where σx is stress in the x direction and E is modulus of elasticity.

May 2, 2023 8 / 49
An example of discretized (finite element mesh)
two-dimensional thin plate.
P2 P3
y

P1

i j

1
m
h 2 Pn
k

m−1
m−2
x
O
y
f2 u2
i j f1 u1
i i

h k

O x

May 2, 2023 9 / 49
Basic Steps of the Finite Element Method
4. Derive the element stiffness matrix and equations: The
development of element stiffness matrices and element equations can be
based on the concept of stiffness influence coefficients, which presupposes
a background in structural analysis. It will produce the equations to
describe the behaviour of an element, and these equations can be written
conveniently in a matrix form as
    

 f1  k11 k12 k13 . . . k1n   u1  
 f2 


   k21 k22 k23 . . . k2n  
  u2  
   

f3 =
 k 31 k32 k 33 . . . k3n

 u 3
 ..   .. .. .. .. ..   .. 
. . . . . .  . 

 
   


 
  
fn k31 k32 k33 . . . k3n un
   

where {f } is the vector of element nodal forces, [k] is the element stiffness
matrix (normally square and symmetric), and {u} is the vector of unknown
element nodal degrees of freedom or generalized displacements n.
May 2, 2023 10 / 49
Basic Steps of the Finite Element Method

5. Assemble the element equations to obtain the global or total

equations and introduce boundary conditions: In this step the
individual element nodal equilibrium equations generated in step four are
assembled into the global nodal equilibrium equations. The final
assembled or global equation written (in compact form) in matrix form is

{F} = [K]{u}

where {F} is the vector of global nodal forces, [K] is the structure global
or total stiffness matrix, (for most problems, the global stiffness matrix is
square and symmetric) and {u} is now the vector of known and unknown
structure nodal degrees of freedom or generalized displacements.

May 2, 2023 11 / 49
Basic Steps of the Finite Element Method

6. Solve for the unknown degrees of freedom (or generalized

displacements): The previous equation modified to account for the
boundary conditions, is a set of simultaneous algebraic equations that can
be written in expanded matrix form as
    

 F 1 
 K 11 K 12 K 13 . . . K 1n 
 u 1 

F K K K . . . K u
    
2 21 22 23 2n 2

 
  

    
F3  K31 K32 K33 . . . K3n 
=  u 3
..  . .. .. .. ..   .. 
 ..
  
.  . . . .  . 

 

 
 
 

Fn K31 K32 K33 . . . K3n un
   

where n is the structure total number of unknown nodal degrees of

freedom.

May 2, 2023 12 / 49
Basic Steps of the Finite Element Method

7. Solve for the element strains and stresses: For the structural
stress-analysis problem, important secondary quantities of strain and stress
(or moment and shear force) can be obtained because they can be directly
expressed in terms of the displacements determined in the previous step.

8. Interpret the results: The final goal is to interpret and analyze the
results for use in the design/analysis process. Determination of locations
in the structure where large deformations and large stresses occur is
generally important in making design/analysis decisions. Post-processor
computer programs help the user to interpret the results by displaying
them in graphical form.

May 2, 2023 13 / 49
Applications of the Finite Element Method

The finite element method can be used to analyse both structural and
non-structural problems.

Typical structural areas include:

1 Stress analysis, including truss and frame analysis (such as pedestrian
walk bridges, high rise building frames, and windmill towers), and
stress concentration problems, typically associated with holes, fillets,
or other changes in geometry in a body (such as automotive parts,
pressures vessels, medical devices, aircraft, and sports equipment).

May 2, 2023 14 / 49
Applications of the Finite Element Method

The finite element method can be used to analyse both structural and
non-structural problems.

Typical structural areas include:

1 Stress analysis, including truss and frame analysis (such as pedestrian
walk bridges, high rise building frames, and windmill towers), and
stress concentration problems, typically associated with holes, fillets,
or other changes in geometry in a body (such as automotive parts,
pressures vessels, medical devices, aircraft, and sports equipment).
2 Buckling, such as in columns, frames, and vessels.

May 2, 2023 14 / 49
Applications of the Finite Element Method

The finite element method can be used to analyse both structural and
non-structural problems.

Typical structural areas include:

1 Stress analysis, including truss and frame analysis (such as pedestrian
walk bridges, high rise building frames, and windmill towers), and
stress concentration problems, typically associated with holes, fillets,
or other changes in geometry in a body (such as automotive parts,
pressures vessels, medical devices, aircraft, and sports equipment).
2 Buckling, such as in columns, frames, and vessels.
3 Vibration analysis, such as in vibratory equipment.

May 2, 2023 14 / 49
Applications of the Finite Element Method

The finite element method can be used to analyse both structural and
non-structural problems.

Typical structural areas include:

1 Stress analysis, including truss and frame analysis (such as pedestrian
walk bridges, high rise building frames, and windmill towers), and
stress concentration problems, typically associated with holes, fillets,
or other changes in geometry in a body (such as automotive parts,
pressures vessels, medical devices, aircraft, and sports equipment).
2 Buckling, such as in columns, frames, and vessels.
3 Vibration analysis, such as in vibratory equipment.
4 Impact problems, including crash analysis of vehicles, projectile
impact, and bodies falling and impacting objects.

May 2, 2023 14 / 49
Applications of the Finite Element Method

Non-structural problems include:

1 Heat transfer, such as in electronic devices emitting heat as in a
personal computer microprocessor chip, engines, and cooling fins in
radiators.

May 2, 2023 15 / 49
Applications of the Finite Element Method

Non-structural problems include:

1 Heat transfer, such as in electronic devices emitting heat as in a
personal computer microprocessor chip, engines, and cooling fins in
radiators.
2 Fluid flow, including seepage through porous media (such as water
seeping through earthen dams), cooling ponds, and in air ventilation
systems as used in sports arenas, etc., air flow around racing cars,
yachting boats, and surfboards, etc.

May 2, 2023 15 / 49
Applications of the Finite Element Method

Non-structural problems include:

1 Heat transfer, such as in electronic devices emitting heat as in a
personal computer microprocessor chip, engines, and cooling fins in
radiators.
2 Fluid flow, including seepage through porous media (such as water
seeping through earthen dams), cooling ponds, and in air ventilation
systems as used in sports arenas, etc., air flow around racing cars,
yachting boats, and surfboards, etc.
3 Distribution of electric or magnetic potential, such as in antennas and
transistors.

May 2, 2023 15 / 49
Applications of the Finite Element Method

Non-structural problems include:

1 Heat transfer, such as in electronic devices emitting heat as in a
personal computer microprocessor chip, engines, and cooling fins in
radiators.
2 Fluid flow, including seepage through porous media (such as water
seeping through earthen dams), cooling ponds, and in air ventilation
systems as used in sports arenas, etc., air flow around racing cars,
yachting boats, and surfboards, etc.
3 Distribution of electric or magnetic potential, such as in antennas and
transistors.
4 Some biomechanical engineering problems (which may include stress
analysis) typically include analyses of human spine, skull, hip joints,
jaw/gum tooth implants, heart, and eye.

May 2, 2023 15 / 49
Basic concepts of the Finite Element Method
Definition of the stiffness matrix

The stiffness matrix is an integral part of the finite element method.

Without understanding the stiffness matrix is impossible to understand the
finite element method.

We define the stiffness matrix as follows: For an element, a stiffness

matrix [k] is a matrix such that
{f } = [k]{u} (1)
where [k] relates nodal displacements {u} to nodal forces {f } of a single
element.

For a whole structure comprising a series of elements:

{F} = [K]{u} (2)
where [K] represents the stiffness matrix of the whole assemblage.
May 2, 2023 16 / 49
Basic concepts of the Finite Element Method
A single spring element

When a pin-jointed tie or strut is part of a structure, its ends will be able
to move due to displacement of the structure and deformation of the
member. This may be modelled by a spring element as shown in Fig. 3.

The points of attachment of the element to other parts of the structure

are called nodes and are indicated by the points 1 and 2 in Fig. 3. Here, f
and u are the force and displacement values and their subscripts indicate
the node to which the apply as well as the coordinate direction.

k
f1x 1 2 f2x
u1 u2
L

Linear spring element with positive nodal displacements and forces.

May 2, 2023 17 / 49
Basic concepts of the Finite Element Method
A single spring element

In situations where the spring element is used to model a tie or strut of

length L and cross-sectional area A, the stiffness of the spring is calculated
as
AE
k=
L
where E is Young’s (elasticity) modulus for the material.

The stiffness matrix relates the nodal force matrix to the nodal
displacement matrix as follows:
    
f1x k11 k12 u1
= (3)
f2x k21 k22 u2

where the element stiffness coefficients kij = kji of the [k] matrix. An
important property of the stiffness matrix for an element is that it is
symmetrical.
May 2, 2023 18 / 49
Basic concepts of the Finite Element Method
A single spring element

Using the sign convention that forces and displacements are positive in the
x-direction, then the forces may be related to the displacements as follows

f1x = k(u1 − u2 ) = ku1 − ku2 (4)

f2x = k(u2 − u1 ) = −ku1 + ku2 (5)

The above equations can be written in a matrix form as

    
f1x k −k u1
= (6)
f2x −k k u2

where  
k −k
[k] = (7)
−k k
is the stiffness matrix for a linear spring element.
May 2, 2023 19 / 49
Basic concepts of the Finite Element Method
An assembly of spring elements

Consider now a system consisting of two spring elements as shown in

Fig. 4. Using the above force-displacement equations for each element
may be written as
         
f1x k1 −k1 u1 f2x k2 −k2 u2
= ; =
f2x −k1 k1 u2 f3x −k2 k2 u3
(8)
f2x , u2
f1x 1 2 3 f3x
u1 u3
k1 k2

(1) (1) (2) (2)

f2x f2x 2 f2x f2x

F3x
Nodal forces consistent with element force sign convention.
May 2, 2023 20 / 49
Basic concepts of the Finite Element Method
An assembly of spring elements

Based on the free-body diagrams of each node and the fact that external
forces must equal internal forces at each node, we can write nodal
equilibrium equations at nodes 1, 2, and 3 as
(1) (1) (2) (2)
F1x = f1x ; F2x = f2x + f2x ; F3x = f3x (9)

Here Newton’s third law, of equal but opposite forces, is applied in moving
from a node to an element associated with the node. Thus, we obtain

F1x = k1 u1 − k1 u2 (10)
F2x = (−k1 u1 + k1 u2 ) + (k2 u2 − k2 u3 ) (11)
F3x = −k2 u2 + k2 u3 (12)

May 2, 2023 21 / 49
Basic concepts of the Finite Element Method
An assembly of spring elements

For each spring element, the nodal forces in a matrix form can be written
down as     
 F1x  k1 −k1 0  u1 
F2x =  −k1 k1 0  u2 (13)
F3x 0 0 0 u3
   
    
 F1x  0 0 0  u1 
F2x =  0 k2 −k2  u2 (14)
F3x 0 −k2 k2 u3
   

As the spring elements are in series, these matrices may be added together
    
 F1x  k1 −k1 0  u1 
F2x =  −k1 k1 + k2 −k2  u2 (15)
F3x 0 −k2 k2 u3
   

May 2, 2023 22 / 49
Basic concepts of the Finite Element Method
An assembly of spring elements

The above equation can be written as the single matrix equation

{F } = [K ]{u} (16)
 
 F1x 
where {F } = F2x is called the global nodal force matrix,
F
 
  3x
u
 1 
{u} = u2 is called
u3
 

the global nodal displacement matrix, and

 
k1 −k1 0
[K ] =  −k1 k1 + k2 −k2  (17)
0 −k2 k2
is called the total or global or system stiffness matrix.
May 2, 2023 23 / 49
Examples: Springs

Example 1: For the spring assemblage with arbitrarily numbered nodes

shown below, obtain
(a) the global stiffness matrix;
(b) the displacements of nodes 3 and 4;
(c) the reaction forces at nodes 1 and 2;
(d) the forces in each spring.
A force of 5000 N is applied at node 4 in the x direction. The spring
constants are given in the figure. Nodes 1 and 2 are fixed.

1 3 4 P = 5000 N 2 x

k1 = 1000 N/mm k2 = 2000 N/mm k3 = 3000 N/mm

May 2, 2023 24 / 49
Examples: Springs
Ex1: Solution

By using the earlier derived stiffness matrix

 
k −k
[k] =
−k k

we express each element stiffness matrix as follows:

   
(1) 1000 −1000 1 (2) 2000 −2000 3
[k ] = ; [k ] =
−1000 1000 3 −2000 2000 4
 
(3) 3000 −3000 4
[k ]=
−3000 3000 2

May 2, 2023 25 / 49
Examples: Springs
Ex1: Solution

Or  (1) (1)   (2) (2) 

k11 k13 k33 k34
[k (1) ] =  ; [k (2) ] =  
(1) (1) (2) (2)
k31 k33 k43 k44
 (3) (3) 
k44 k42
[k (3) ] =  
(3) (23)
k24 k22
For clarity sake, we rewrite these matrix in expanded form, to see the
position of each matrix component in 4 × 4 matrix:
 
1000 0 −1000 0 1
0 0 0 0 
2
[k (1) ] = 

 −1000 0 1000 0  3
0 0 0 0 4
May 2, 2023 26 / 49
Examples: Springs
Ex1: Solution

 
0 0 −0 0 1
0 0 0 0 2
[k (2) ] = 


 0 0 2000 −2000  3
0 0 −2000 2000 4
 
0 0 −0 0 1
(3)
 0 3000 0 −3000 
2
[k ] =   0 0 0 0 3
0 −3000 0 3000 4
Using the concept of superposition, we obtain the global stiffness matrix as

[K ] = [k (1) ] + [k (2) ] + [k (3) ]

May 2, 2023 27 / 49
Examples: Springs
Ex1: Solution

Each element of the global stiffness matrix is calculated as the sum of

individual stiffnesses:
(1) (2) (3)
[Kij ] = [kij ] + [kij ] + [kij ]

For example,
(1) (2) (3)
[K33 ] = [k33 ] + [k33 ] + [k33 ] = 1000 + 2000 + 0 = 3000

Finally, the global stiffness matrix is

 
1000 0 −1000 0 u1
 0 3000 0 −3000  u2

[K ] = 
 −1000 0 3000 −2000  u3
0 −3000 −2000 5000 u4

May 2, 2023 28 / 49
Examples: Springs
Ex1: Solution

(b) The global stiffness matrix relates global forces to global displacements
as follows:
    

 F1x 
 1000 0 −1000 0 
 u1 
F2x 0 3000 0 −3000   u2 
  
=  −1000

F3x  0 3000 −2000   u 
 3 

  
F4x 0 −3000 −2000 5000 u4
  

Applying the boundary conditions u1 = 0 and u2 = 0 to the above

equations, substituting applied nodal forces, and partitioning the first two
equations (or deleting the first two rows of {F } and {u} and the first two
rows and columns of [K ] corresponding to the zero-displacement boundary
conditions), we obtain
    
0 3000 −2000 u3
=
5000 −2000 5000 u4
May 2, 2023 29 / 49
Examples: Springs
Ex1: Solution

Solving for u3 and u4 gives us

10 15
u3 = mm; u4 = mm
11 11
(c) To obtain the global nodal forces (which include the reactions at nodes
1 and 2), we back-substitute the found displacements and the boundary
conditions u1 = 0 and u2 = 0 into the equations for the global forces:
    

 F1x 
 1000 0 −1000 0 
 0  
F2x 0 3000 0 −3000  0 
  
= −1000


 F3x 
 0 3000 −2000    10/11 
F4x 0 −3000 −2000 5000 15/11
   

Multiplying the matrices and simplifying, we obtain the reaction forces at

each node
F1x = −909.09 N; F2x = −4090.91 N; F3x = 0; F4x = 5000 N
May 2, 2023 30 / 49
Examples: Springs
Ex1: Solution

From these results, we observe that F1x + F2x = F4x . This result verifies
equilibrium of the whole spring assemblage.

(d) Next we use stiffness matrices for each element to obtain the forces in
each element.

Element 1:
 
 f (1)    
1x 1000 −1000 0
=
 f (1)  −1000 1000 10/11
3x

Solving it, we obtain

(1) (1)
f1x = −909.09 N; f3x = 909.09 N

May 2, 2023 31 / 49
Examples: Springs
Ex1: Solution

1 3 (1)
F1x 1 f1x
909.09 N 909.09 N
k1

Element 2:
 
 f (2)    
3x 2000 −2000 10/11
=
 f (2)  −2000 2000 15/11
4x
Solving it, we obtain
(2) (2)
f3x = −909.09 N; f4x = 909.09 N

3 4
909.09 N 909.09 N
k2
May 2, 2023 32 / 49
Examples: Springs
Ex1: Solution

Element 3:
 
 f (3)    
4x 3000 −3000 15/11
=
 f (3)  −3000 3000 0
2x

Solving it, we obtain

(3) (3)
f4x = 4090.91 N; f2x = −4090.91 N

4 2 (3)
f2x 2 F2x
4090.91 N 4090.91 N
k3

May 2, 2023 33 / 49
Examples: Springs

Example 2: For the spring assemblage shown in below in the figure, obtain
(a) the global stiffness matrix;
(b) the displacements of nodes 2–4;
(c) the global nodal forces;
(d) the local element forces.
Node 1 is fixed while node 5 is given a fixed, known displacement
δ = 20.0 mm. The spring constants are all equal to k = 200 kN/m.

F5x
1 2 3 4 5 5

1 2 3 4 20 mm

May 2, 2023 34 / 49
Examples: Springs
Ex2: Solution

(a) We express each element stiffness matrix as

 
(1) (2) (3) (4) 200 −200
[k ] = [k ] = [k ] = [k ] =
−200 200

Using superposition, we obtain the global stiffness matrix as the sum of

individual matrices
 
200 −200 0 0 0
 −200 400 −200 0 0 
 
 0
[K ] =  −200 400 −200 0  kN/m

 0 0 −200 400 −200 
0 0 0 −200 200

May 2, 2023 35 / 49
Examples: Springs
Ex2: Solution

(b) The global stiffness matrix relates the global forces to the global
displacements as follows:
    

 F1x 
 200 −200 0 0 0 
 u1 

F −200 400 −200 0 0 u2
   
 2x  
  

 

F3x =  0 −200 400 −200 0  u3
  
F 0 0 −200 400 −200 u4
 
 4x 

 
   

  
F5x 0 0 0 −200 200 u5
  

Applying the boundary conditions u1 = 0 and u5 = 20 mm, substituting

known global forces F2x = 0, F3x = 0, and F4x = 0, and partitioning the
first and fifth equations corresponding to these boundary conditions, we
obtain

May 2, 2023 36 / 49
Examples: Springs
Ex2: Solution

 
    0
 

 0  −200 400 −200 0 0  u2

 


0 = 0 −200 400 −200 0  u3
0 0 0 −200 400 −200  u
   
 4
 

 
0.02


By multiplying the matrices we get

400u2 − 200u3 + 0 = 0
−200u2 + 400u3 − 200u4 = 0
0 − 200u3 + 400u4 = 4

May 2, 2023 37 / 49
Examples: Springs
Ex2: Solution

Or, in matrix form

    
 0  400 −200 0  u2 
0 =  −200 400 −200  u3
4 0 −200 400 u4
   

Solving, we obtain

u2 = 0.005 m u3 = 0.01 m u4 = 0.015 m

May 2, 2023 38 / 49
Examples: Springs
Ex2: Solution

(c) The global nodal forces are obtained by back-substituting the boundary
condition displacements.
    

 F1x  200 −200 0 0 0 
 0 
 F2x    −200 400 −200 0 0  0.005

   
 

  
F3x = 0
 −200 400 −200 0   0.01
F  0 0 −200 400 −200   0.015 
    
 4x 

   
  
F5x 0 0 0 −200 200 0.02
  

This substitution yields

F1x = −200 × 0.005 = −1.0 kN
F2x = 400 × 0.005 − 200 × 0.01 = 0
F3x = −200 × 0.005 + 400 × 0.01 − 200 × 0.015 = 0
F4x = −200 × 0.01 + 400 × 0.015 − 200 × 0.02 = 0
F5x = −200 × 0.015 + 200 × 0.02 = 1 kN
May 2, 2023 39 / 49
Examples: Springs
Ex2: Solution

The nodal force F5x = 1 kN is required to displace node 5 by δ = 20.0 mm.

This result verifies equilibrium of the whole spring assemblage.

Remember if the displacement is known at a node in a given direction (in

this example,u5 = 20 mm), then the force F5x at that same node and in
that same direction is not initially known. The force is determined after
solving for the unknown nodal displacements.

(d) Next, we make use of local element stiffness matrix to obtain the
forces in each element.

May 2, 2023 40 / 49
Examples: Springs
Ex2: Solution

Element 1:  
 f (1)    
1x 200 −200 0
=
 f (1)  −200 200 0.005
2x

which gives us
(1) (1)
f1x = −1.0 kN f2x = 1.0 kN

Element 2:  
 f (2)    
2x 200 −200 0.005
=
 f (2)  −200 200 0.01
3x

which gives us
(2) (2)
f2x = −1.0 kN f3x = 1.0 kN

May 2, 2023 41 / 49
Examples: Springs
Ex2: Solution

Element 3:  
 f (3)    
3x 200 −200 0.01
=
 f (3)  −200 200 0.015
4x

which gives us
(3) (3)
f3x = −1.0 kN f4x = 1.0 kN

Element 4:  
 f (4)    
4x 200 −200 0.015
=
 f (4)  −200 200 0.02
5x

which gives us
(4) (4)
f4x = −1.0 kN f5x = 1.0 kN

May 2, 2023 42 / 49
Examples: Springs

Example 3: Formulate the global stiffness matrix for the spring

assemblage shown below in the figure.

k2
2 3

1 k1 2 P x
k3
2 4
Rigid bar

May 2, 2023 43 / 49
Examples: Springs
Ex3: Solution

The boundary conditions are

u1 = u3 = u4 = 0

The compatibility condition at node 2 is

(1) (2) (3)
u2 = u2 = u2 = u2

The nodal equilibrium conditions are

(1)
F1x = f1x
(2)
F3x = f3x
(3)
F4x = f4x
(1) (2) (3)
P = f2x + f2x + f2x

May 2, 2023 44 / 49
Examples: Springs
Ex3: Solution

Free-body diagrams of elements and nodes of spring assemblage are shown

next

(2) 2 (2)
2 f2x 2 3 f3x 3 F3x

(1) 1 (1)
F1x f1x f2x 2 P
1 1 2
(3) 3 (3)
2 f2x 2 4 f4x 4 F4x

May 2, 2023 45 / 49
Examples: Springs
Ex3: Solution

Using the local stiffness matrix applied to each element and compatibility
condition, we obtain the total or global equilibrium equations as

F1x = k1 u1 − k1 u2
F3x = −k2 u2 + k2 u3
F4x = −k3 u2 + k3 u4
P = (−k1 u1 + k1 u2 ) + (k2 u2 − k2 u3 ) + (k3 u2 − k3 u4 )

In matrix form,
    

 F1x 
 k1 −k1 0 0 
 u1 

P −k1 k1 + k2 + k3 −k2 −k3  u2
     
=
F 3 0 −k 2 k2 0 u
 3
 
 x 

  
 

F4x 0 −k3 0 k3 u4
 

May 2, 2023 46 / 49
Examples: Springs
Ex3: Solution

Making use of the boundary conditions, and then considering the equation
for the force P, we solve for it u2 as
P
u2 =
k1 + k2 + k3
Next, we solve for the global forces as

F1x = −ku2 F3x = −k2 u2 F4x = −k3 u2

The forces obtained can be interpreted as the global reactions in this

example. The negative signs in front of these forces indicate that they are
directed to the left (opposite the x axis).

May 2, 2023 47 / 49
Examples: Springs
Ex3: Solution

Using the direct stiffness method, we formulate the global stiffness matrix.
First, we express each element stiffness matrix as
   
(1) k1 −k1 u1 (2) k2 −k2 u2
[k ] = [k ] =
−k1 k1 u2 −k2 k2 u3
 
(3) k3 −k3 u2
[k ]=
−k3 k3 u4
where the particular degrees of freedom associated with each element are
listed in the columns on the right hand side of each matrix.

May 2, 2023 48 / 49
Examples: Springs
Ex3: Solution

Using the direct stiffness method add terms from each element stiffness
matrix into the appropriate corresponding row and column in the global
stiffness matrix to obtain
 
k1 −k1 0 0 u1
 −k1 k1 + k2 + k3 −k2 −k3  u2
[K ] =  
 0 −k2 k2 0  u3
0 −k3 0 k3 u4

We observe that each element stiffness matrix [k] has been added into the
location in the global [K ] corresponding to the identical degree of freedom
associated with the element [k].

May 2, 2023 49 / 49
 Introduction to
Finite Element Method

Part Two

May 15, 2023

May 15, 2023 1 / 62

Analysis of trusses (frameworks)

We will now derive the stiffness matrix for a linear-elastic bar (or truss)
element using the general steps outlined in previous sections. Although the
spring elements considered so far have been collinear this need to be the
case. The spring element could, for example, be used to analyse the forces
and deformation of a truss.

However, as the spring elements in the model are at different angles to one
another it is necessary to express the forces and deformations for each
element relative to the global co-ordinate system x and y .

May 15, 2023 2 / 62

Analysis of trusses (frameworks)
Derivation of the stiffness matrix for a bar element in local coordinates

The bar element is assumed to have constant cross-sectional area A,

modulus of elasticity E , and initial length L. The nodal degrees of freedom
are local axial displacements (longitudinal displacements directed along the
length of the bar) represented by u1 and u2 at the ends of the element as
shown in Fig. 5. From Hooke’s law and the strain-displacement
relationship we write

x, u
T 1 L 2 T

u1 , f1x u2 , f2x
Bar subjected to tensile forces T .

May 15, 2023 3 / 62

Analysis of trusses (frameworks)
Derivation of the stiffness matrix for a bar element in local coordinates

du
σ x = E εx εx =
dx
From force equilibrium, we have

Aσx = T = contant

for a bar with loads applied only at the ends. Using the above relations
and differentiating with respect to x, we obtain the differential equation
governing the linear-elastic bar behaviour as
 
d du
EA =0
dx dx

where u is the axial displacement function along the element in the x

direction.
May 15, 2023 4 / 62
Analysis of trusses (frameworks)
Derivation of the stiffness matrix for a bar element in local coordinates

The following assumptions are used in deriving the bar element stiffness
matrix:

1 The bar cannot sustain shear force or bending moment, that is,
f1y = 0, f2y = 0, M1 and M2 = 0.
2 Any effect of transverse displacement is ignored.
3 Hooke’s law applies; that is, axial stress σx is related to axial strain εx
by σx = E εx .
4 No intermediate applied loads.

Next, the major steps required for the derivation of the stiffness matrix are
outlined.

May 15, 2023 5 / 62

Analysis of trusses (frameworks)
Derivation of the stiffness matrix for a bar element in local coordinates

1. Select the element type. Represent the bar by labelling nodes at

each end and in general by labelling the element number (Fig. 5).

2. Define the strain-displacement and stress-strain relationships.

The strain-displacement relationship is
u2 − u1
εx =
L
and the stress/strain relationship by Hooke’s law is
σx = E εx
3. Derive the element stiffness matrix and equations. The element
stiffness matrix is derived as follows. From elementary mechanics, we have
 
u2 − u1
T = Aσx = EA
L
May 15, 2023 6 / 62
Analysis of trusses (frameworks)
Derivation of the stiffness matrix for a bar element in local coordinates

Also, by the nodal force sign convention,

EA
f1x = −T = − (u2 − u1 )
L
Similarly,
EA
(u2 − u1 )
f2x = T =
L
Expressing the above two equations together in matrix form, we have
    
f1x EA 1 −1 u1
=
f2x L −1 1 u2
Now, because {f } = [k]u, we get the stiffness matrix written as
 
EA 1 −1
[k] =
L −1 1
It represents the stiffness matrix for a bar element in local coordinates.
May 15, 2023 7 / 62
Analysis of trusses (frameworks)
Derivation of the stiffness matrix for a bar element in local coordinates

4. Assemble element equations to obtain global or total equations.

Assemble the global stiffness and force matrices and global equations using
the direct stiffness method described earlier on. This step applies for
structures composed of more than one element such that
N
X N
X
[K ] = [k (e) ] and {F } = {f (e) }
e=1 e=1

where now all local element stiffness matrices [k (e) ] must be transformed
to global element stiffness matrices [k] (unless the local axes coincide with
the global axes) before the direct stiffness method is applied.

May 15, 2023 8 / 62

Analysis of trusses (frameworks)
Derivation of the stiffness matrix for a bar element in local coordinates

5. Solve for the nodal displacements Determine the displacements by

imposing boundary conditions and simultaneously solving a system of
equations, {F} = [K] {u}.

6. Solve for the element forces. Finally, determine the strains and
stresses in each element by back-substitution of the displacements into
equations for stresses and deformations.

May 15, 2023 9 / 62

Analysis of trusses: Example

For the three-bar assemblage shown in Fig. 6, determine

(a) the global stiffness matrix;
(b) the displacements of nodes 2 and 3;
(c) the reactions at nodes 1 and 4. A force of 13.3 kN is applied in the x
direction at node 2. The length of each element is 0.8 m. Let E = 250 GPa
and A = 6.45 × 10−4 m2 for elements 1 and 2, and let E = 125 GPa and
A = 12.9 × 10−4 m2 for element 3. Nodes 1 and 4 are fixed.

13.3 kN
1 1 2 2 3 3 4 x
0.8m 0.8m 0.8m

Three-bar assemblage.

May 15, 2023 10 / 62

Solution
a) The stiffness matrix of each element is given as
 
EA 1 −1
[k] =
L −1 1
Thus,
(250 × 106 )(6.45 × 10−4 )
 
(1) (2) 1 −1
[k ] = [k ] =
0.8 −1 1
 
1 −1
= 201.5625 × 103 kN/m
−1 1

(125 × 106 )(12.9 × 10−4 )

 
1 −1
[k (3) ] =
0.8 −1 1
 
1 −1
= 201.5625 × 103 kN/m
−1 1

May 15, 2023 11 / 62

Solution
Assembling the element stiffness matrices by the direct stiffness method,
we obtain the global stiffness matrix as (kPa)
 
1 −1 0 0 u1
 −1 2 −1 0 
[K ] = 201.5625 × 103   u2
 0 −1 2 −1  u3
0 0 −1 1 u4

b) The governing equation which relates global nodal forces to global

nodal displacements is written as follows:
    

 F1x  1 −1 0 0   u1 
F2x 3  −1 2 −1 0  u2
    
= 201.5625 × 10  
F  0 −1 2 −1   u 
 3x   3 

 
F4x 0 0 −1 1 u4
 

May 15, 2023 12 / 62

Solution

The boundary conditions are

u1 = 0 u4 = 0

Therefore,
    
13.3 3 2 −1 u2
= 201.5625 × 10
0 −1 2 u3

Solving simultaneously for displacements we obtain

u2 = 4.3990 × 10−5 m = 0.044 mm u3 = 2.1995 × 10−5 m = 0.022 mm

May 15, 2023 13 / 62

Solution
c) Back-substituting the displacements into the governing equations, we
obtain the global nodal forces, which include the reactions at nodes 1 and
4, as follows:
   

 F1x  −1 0  

F2x
  2 −1  4.3990
3 −5
= 201.5625 × 10   2.1995 × 10


 F 3x 
 −1 2
F4x 0 −1
 
 

 −8.87  
13.30
 
= kN

 0 
−4.43
 

The obtained results show that the sum of the reactions F1x and F4x is
equal in magnitude but opposite in direction to the applied nodal force of
13.3 kN at node 2. Equilibrium of the bar assemblage is thus verified.
May 15, 2023 14 / 62
Trusses
Global stiffness matrix for bar arbitrarily oriented in the plane

We now consider a bar inclined at an angle θ from the global x axis

identified by the local axis x ′ directed from node 1 to node 2 along the
direction of the bar, as shown below in the figure. Here positive angle θ is
taken counter-clockwise from x to x ′ .

y, v x′ , u′
′
′ f 2x
u 2,
2
y′ L

θ
1 x, u
′
u′1 , f1x

Bar element arbitrarily oriented in the global x–y plane.

May 15, 2023 15 / 62

Trusses
Global stiffness matrix for bar arbitrarily oriented in the plane

Next, we write down the governing equations in the coordinate system, –

the prime notations denotes the local system.
( ′ ) ( ′ )
f1x u1

EA 1 −1
= (1)
′
f2x L −1 1 u2′
or
{f ′ } = [k ′ ]{u ′ } (2)
We now want to relate the global element nodal forces {f } to the global
nodal displacements {u}
   

 f1x 
 
 u1 

f1y v1
   
= [k] (3)

 f2x 
 
 u2 

f2y v2
   

May 15, 2023 16 / 62

Trusses
Global stiffness matrix for bar arbitrarily oriented in the plane

or, in simplified matrix form

{f } = [k]{u} (4)

The relationship between the local and global displacement and forces
components is given as (see the figure)

u1′ = u1 cos θ + v1 sin θ

u2′ = u2 cos θ + v2 sin θ

In matrix form, it can be written as

 
 u1 
u1′
    
c s 0 0 v1
 
= (5)
u2′ 0 0 c s  u2 

v2
 

May 15, 2023 17 / 62

Trusses
Global stiffness matrix for bar arbitrarily oriented in the plane

or as
{u ′ } = [T ∗ ]{u} (6)
where [T ∗ ] is the transformation matrix
 
∗ c s 0 0
[T ] = (7)
0 0 c s

Similarly, these equations can be written for the forces

 
( ′ )   f1x 
f1x
 
c s 0 0 f1y
 
= (8)
′
f2x 0 0 c s  f 
 2x 

f2y


{f ′ } = [T ∗ ]{f } (9)
May 15, 2023 18 / 62
Trusses
Global stiffness matrix for bar arbitrarily oriented in the plane

Now, substituting (6) into (2), we obtain

{f ′ } = [k ′ ][T ∗ ]{u} (10)

and using (9) in (10) yields

[T ∗ ]{f } = [k ′ ][T ∗ ]{u} (11)

The equations for each nodal displacement transformation can be written

as  ′  
 u1  c s 0

0  u1

  
 v′ 
  
−s c 0 0  v1
   
1
′
=  (12)
u2  0 0 c s  u
 2
   
   
0 0 −s c v2

 ′   
v 2

{u ′ } = [T ]{u} (13)
May 15, 2023 19 / 62
Trusses
Global stiffness matrix for bar arbitrarily oriented in the plane

Here [T ] is the transformation matrix

 
c s 0 0
 −s c 0 0 
[T ] = 
 0 0
 (14)
c s 
0 0 −s c
{f ′ } = [T ]{f } (15)
Because forces are like displacements – both are vectors. Also, must [k ′ ]
be expanded to a 4 × 4 matrix. Therefore, (1) in expanded form becomes
 ′ 
u1′ 
 
 f1x   

 ′ 
  1 0 −1 0   

′
  
 f 
  EA  
v

0 0 0 0
 
1y  1
=   (16)
 f ′ 
2x
L  −1 0 1 0  u ′ 
   2 
 
0 0 0 0 

 
 
 ′ 
  ′ 

f2y v2
 

May 15, 2023 20 / 62

Trusses
Global stiffness matrix for bar arbitrarily oriented in the plane

In equation (16), because f1y′ and f ′ are zero, rows of zeros corresponding
2y
to these forces appear in [k ′ ]. Now, using (13) and (15) in (2), we obtain

[T ]{f } = [k ′ ][T ]{u} (17)

Premultiplying both sides of (17) by [T ]−1 , we have

{f } = [T ]−1 [k ′ ][T ]{u} (18)

where [T ]−1 is the inverse of [T ]. Also,

[T ]−1 = [T ]T (19)

where [T ]T is the transpose of [T ]. Substituting (19) into (18), we arrive

at
{f } = [T ]T [k ′ ][T ]{u} (20)
May 15, 2023 21 / 62
Trusses
Global stiffness matrix for bar arbitrarily oriented in the plane

Equating equations (4) and (20), we obtain the global stiffness matrix for
an element as
[k] = [T ]T [k ′ ][T ] (21)
Substituting (14) for [T ] and the expanded form of [k ′ ] given in (16) into
(21), we obtain [k] given in explicit form by
 2
cs −c 2

c −cs
EA  cs
 s 2 −cs −s 2 
[k] =  (22)
L  −c −cs c 2
2 cs 
−cs −s 2 cs s2

It is important to note that the above matrix must be symmetrical.

May 15, 2023 22 / 62

Trusses
Global stiffness matrix for bar arbitrarily oriented in the plane

The global stiffness and force matrices can be obtained using the direct
stiffness method
N
X
[K ] = [k (e) ] (23)
e=1
N
X
{F } = {f (e) (24)
e=1

and
{F} = [K]{u} (25)

May 15, 2023 23 / 62

Example

For the bar element shown below, evaluate the global stiffness matrix with
respect to the x–y coordinate system if θ = 30◦ . Let the bar’s
cross-sectional area equal 25 cm2 , length equal 150 cm, and modulus of
elasticity equal 2.0 × 105 MPa.

y, v x′ , u′
′
′ f 2x
u 2,
2
y′ L

θ
1 x, u
′
u′1 , f1x

May 15, 2023 24 / 62

Solution

√
◦ 3
For θ = 30◦ we have c = cos 30 = and s = sin 30◦ = 0.5.
2
 √ √ 
3 3 3 3
− −
 4 4 4 4 
 √ 
1 3 1 
 
− −

(2.0 × 105 )(0.0025)  ... 
[k] =
 4 4 4  MN/m
1.5
 √ 
 3 3 
 ... ... 
4 4 
 

 1 
... ... ...
4

May 15, 2023 25 / 62

Solution

The above matrix is symmetrical, only one half is shown therefore.

Simplifying, we have
 
0.250 0.144 −0.250 −0.144
 . . . 0.083 −0.144 −0.083 
[k] = 103   MN/m
 
 ... ... 0.250 0.144 
... ... ... 0.083

May 15, 2023 26 / 62

Trusses
Computation of stress for a bar in the x–y plane

The usual definition of axial tensile stress is axial force divided by

cross-sectional area.
f′
σ = 2x (26)
A
′ is used because it is the axial force that pulls on the bar as
where f2x
shown in the above example. By (1),
 ′ 
′ EA u1
f2x = [−1 1] (27)
L u2′
Combining (26) and (27) yields
E
{σ} = [−1 1]{u ′ } (28)
L
Using (6) we obtain
E
{σ} = [−1 1][T ∗ ]{u} (29)
L
May 15, 2023 27 / 62
Trusses
Computation of stress for a bar in the x–y plane

Equation (29) can be expressed in simpler form as

{σ} = [C ′ ]{u} (30)

where (when we use [T ∗ ])

 
′E c s 0 0
C = [−1 1] (31)
L 0 0 c s

After multiplying the matrices in (31), we have

E
[C ′ ] = [−c −s c s] (32)
L

May 15, 2023 28 / 62

Example

For the bar shown in the previous example, determine the axial stress. Let
A = 4 × 10−4 m2 , E = 210 GPa, and L = 2 m, and let the angle between x
and x ′ be 60◦ . Assume the global displacements have been previously
determined to be u1 = 0.25 mm, v1 = 0.0, u2 = 0.50 mm, and
v2 = 0.75 mm.

May 15, 2023 29 / 62

Solution

We can use equation (30) to evaluate the axial stress. Therefore, we first
calculate [C ′ ] from (32) as

6
" √ √ #
210 × 10 1 3 1 3
[C ′ ] = − −
2 2 2 2 2
√
◦ 1 ◦ 3
where c = cos 60 = and s = sin 60 = in the above equation. Now
2 2
{u} is given by

0.25 × 10−3 
   

 u1 
 
 
v1 0.0
   
{u} = = m
u   0.50 × 10−3 
 2 

   
v2 0.75 × 10−3
 

May 15, 2023 30 / 62

Solution

Using equation (30), we obtain the bar axial stress as

0.25 × 10−3
 
" √ √ # 
210 × 106
 
1 3 1 3 
0.0

σx = − −
2 2 2 2 2   0.50 × 10−3 

0.75 × 10−3
 

= 81.32 × 103 kPa = 81.32 MPa

May 15, 2023 31 / 62

Trusses
Inclined supports

If a support is inclined, or skewed, at an angle from the global x axis, as

shown below, the resulting boundary conditions on the displacements are
not in the global x – y directions.

There are two methods used to handle inclined supports. In the first
method, to account for inclined boundary conditions, we must perform a
transformation of the global displacements at into the local nodal
coordinate system, while keeping all other displacements in the x y global
system. This procedure is a bit cumbersome, but an accurate one.

May 15, 2023 32 / 62

Trusses
Inclined supports

The second method is much simple,more popular, an is used in some

computer programmes. The support is replaced by an additional member
which is much shorter than other members, and has a very large stiffness,
compared to the rest of members as shown below.

? vi

j
?
i i ui
uj = vj = 0

May 15, 2023 33 / 62

Examples: Solution of a plane truss
Example 1. For the plane truss composed of the three elements shown
below determine the x and y displacements at node 1, and the stresses in
each element. Let E = 210 GPa and A = 1.3 × 10−3 m2 for all elements.

2 3

2
3m

45◦

45◦ 3 4 x
1
3m
45 kN
May 15, 2023 34 / 62
Solution of a plane truss
Ex1: Solution

Data for the truss:

Element θ c s c2 s2 cs

1 90◦ 0 1 0 1 0
√ √
2 45◦ 2/2 2/2 0.5 0.5 0.5

3 0 1 0 1 0 0

May 15, 2023 35 / 62

Solution of a plane truss
Ex1: Solution

For element 1, using equation (22), we obtain

 
0 0 0 0 u1
× 6 )(1.3 × 10−3 ) 
(210 10 0 1 0 −1 
 v1
[k (1) ] = 
3  0 0 0 0  u2
0 −1 0 1 v2

Similarly, for element 2:

 
0.5 0.5 −0.5 −0.5 u1
(210 × 106 )(1.3 × 10−3 ) 
 0.5 0.5 −0.5 −0.5 
 v1
[k (2) ] = √
3× 2  −0.5 −0.5 0.5 0.5  u3
−0.5 −0.5 0.5 0.5 v3

May 15, 2023 36 / 62

Solution of a plane truss
Ex1: Solution

and for element 3:

 
1 0 −1 0 u1
(210 × 106 )(1.3 × 10−3 ) 
[k (3) ] =  0 0 0 0 
 v1
3  −1 0 1 0  u4
0 0 0 0 v4

The common factor of (210 × 106 )(1.3 × 10−3 )/3 can be taken from each
of three matrices, √where each term in the square bracket of the second
matrix divided by 2. After adding terms from the individual element
stiffness matrices into their corresponding locations in [K ], we obtain the
total stiffness matrix.
0.5
(210 × 106 )(1.3 × 10−3 )/3 = 91000 and √ = 0.354
2

May 15, 2023 37 / 62

Solution of a plane truss
Ex1: Solution

To properly understand how the total stiffness matrix is assembled, let us

show the position of the elements of local stiffness matrices in the global
matrix.
 (1) (1) (1) (1)

k11 k12 k13 k14 0 0 0 0 u1
 
 (1) (1) (1) (1)
 k21 k22 k23 k24 0 0 0 0  v1

 
 (1) (1) (1) (1)  u2

 k31 k32 k33 k34 0 0 0 0
 
 v2
 (1) (1) (1) (1)

(1)

[k ] =  41
 k k42 k43 k44 0 0 0 0 
u3

 0 0 0 0 0 0 0 0 
 
v3
 
 0 0 0 0 0 0 0 0 
 
u4
 
 0 0 0 0 0 0 0 0 
 
0 0 0 0 0 0 0 0 v4
May 15, 2023 38 / 62
Solution of a plane truss
Ex1: Solution

(2) (2) (2) (2)

 
k11 k12 0 0 k13 k14 0 0 u
  1
 (2) (2) (2) (2)
 k21 k22 0 0 k23 k24 0 0  v1

 
 0 0 0 0 0 0 0 0  u2
 
 
 0 0 0 0 0 0 0 0  v2
 
(2)
[k ] = 
 (2)

(2) (2) (2)
 u3

 k31 k32 0 0 k33 k34 0 0
 
 v3
 (2) (2) (2) (2)

 k
 41 k42 0 0 k43 k44 0 0 
 u4
 
 0 0 0 0 0 0 0 0
 
0 0 0 0 0 0 0 0 v4

May 15, 2023 39 / 62

Solution of a plane truss
Ex1: Solution

(3) (3) (3) (3)

 
k11 k12 0 0 0 0 k13 k14 u
  1
 (3) (3) (3) (3)
 k21 k22 0 0 0 k23 k24  v1

 
 0 0 0 0 0 0 0 0  u2
 
 
 0 0 0 0 0 0 0 0  v2
 
(3)
[k ] =  
 0 0 0 0 0 0 0 0  u3
 
 
 0 0 0 0 0 0 0 0  v3
 
 
 (3) (3) (3) (3)
 u4

 k31 k32 0 0 0 0 k33 k34
 
(3) (3)
k41 k42 0 0 0 0 k43
(3) (3)
k44 v4

May 15, 2023 40 / 62

Solution of a plane truss
Ex1: Solution

By summing up these three matrices we obtain the global stiffness matrix.

 
1.354 0.354 0 0 −0.354 −0.354 −1 0 u1
 
 0.354
 1.354 0 −1 −0.354 −0.354 0 0  v1

 

 0 0 0 0 0 0 0 0  u2

 
 0 −1 0 1 0 0 0 0   v2
[K ] = 91000×


 −0.354 −0.354 0 0 0.354 0.354 0 0 
  u3
 
 −0.354 −0.354 0 0 0.354 0.354 0 0 
  v3
 
 −1 0 0 0 0 0 1 0 
  u4
0 0 0 0 0 0 0 0 v4

May 15, 2023 41 / 62

Solution of a plane truss
Ex1: Solution

Next, calculate the displacements at joint 1:

−0.354 −0.354 −1
   
0  1.354 0.354 0 0 0 u1 


  
 

 
   




 −45 



 0.354
 1.354 0 −1 −0.354 −0.354 0 0 


 v1 




 
   

F2x  0 0 0 0 0 0 0 0 0 

    

 
   


 
   

   
 F2y 
  
 0 −1 0 1 0 0 0 0 
 0 
= 91000 ×  
F3x  −0.354 −0.354 0 0 0.354 0.354 0 0 0 
    

 
  

 
   
    



 F3y 

  −0.354 −0.354 0 0 0.354 0.354 0 0  0 


 
   

    

F4x −1 0 0 0 0 0 1 0 0 

 
   

 
   


 
 
 

   
F4y 0 0 0 0 0 0 0 0 0

May 15, 2023 42 / 62

Solution of a plane truss
Ex1: Solution

Thus, we have
( ) " #( )
0 1.354 0.354 u1
= 91000 ×
−45 0.354 1.354 v1

u1 = 1.025×10−4 m = 0.103 mm v1 = −3.9205×10−4 m = −0.392 mm

Using (30), we determine stresses in each element as follows

u1 = 1.025 × 10−4
 

 

 
 v1 = −3.9205 × 10−4
 
6
 
(1) 210 × 10 
σ = [0 − 1 0 1]
3 
 u2 = 0 


 

 
v2 = 0
 
= 27444 kPa = 27.44 MPa
May 15, 2023 43 / 62
Solution of a plane truss
Ex1: Solution

" √ √ √ √ #
210 × 106 2 2 2 2
σ (2) = √ − −
3 2 2 2 2 2
u = 1.025 × 10−4
 
 1

 


 v1 = −3.9205 × 10−4 

 

× = 10134 kPa = 10.13 MPa


 u 3 = 0 



 

v3 = 0
 

May 15, 2023 44 / 62

Solution of a plane truss
Ex1: Solution

u1 = 1.025 × 10−4
 

 

 
 v1 = −3.9205 × 10−4
 
210 × 106
 

σ (3) = [−1 0 1 0]
3 
 u4 = 0 


 

 
v4 = 0
 
= −7175 kPa = −7.18 MPa

We now verify our results by examining force equilibrium at node 1:

X
Fx = 0 : (10134)(1.3 × 10−3 ) cos 45◦ − (7175)(1.3 × 10−3 ) = 0.012 ≈ 0
X
Fy = 0 : (27444)(1.3 × 10−3 ) + (10134)(1.3 × 10−3 ) cos 45◦ − 45 = −0.007 ≈ 0

May 15, 2023 45 / 62

Solution of a plane truss
Example 2. For the two-bar truss shown below in the figure determine the
displacement in the y direction of node 1 and the axial force in each
element. A force of P = 1000 kN is applied while node 1 settles an
amount δ = 50 mm in the negative x direction. Let E = 210 GPa and
A = 6.0 × 10−4 m2 for each element.

2
x

1
y
3m

θ(1)

3 2 1 1000 kN
δ = 50 mm

4m

May 15, 2023 46 / 62

Solution of a plane truss
Ex2: Solution

We begin by using equations (22) to determine each element stiffness

matrix.
3 4
Element 1: cos θ(1) = = 0.6 and sin θ(1) = = 0.8.
5 5
EA (210 × 106 )(6 × 10−4 )
= = 25200
L 5
 
0.36 0.48 −0.36 −0.48 u1
. . . 0.64 −0.48 −0.64  v1
[k (1) ] = 25200 × 


 ... ... 0.36 0.48  u2
... ... ... 0.64 v2

May 15, 2023 47 / 62

Solution of a plane truss
Ex2: Solution

Element 2: cos θ(2) = 0.0 and sin θ(2) = 1.0.

EA (210 × 106 )(6 × 10−4 )
= = 31500 = 25200 × 1.25
L 4
  
0 0 0 0 0 0 0
. . . 1 0 −1 . . . 1.25 0 −
[k (2) ] = 25200×1.25×
  
 = 25200×
 ... ... 0 0   ... ... 0
... ... ... 1 ... ... ...
Governing equations
   
0.36 0.48 −0.36 −0.48


 F1x   0 0   u1 



 F 1y


  . . . 1.89 −0.48 −0.64 0 −1.25  v1 

    

F2x ... ... 0.36 0.48 0 0  u2
     
= 25200×

 F2y    ... ... ... 0.64 0 0 
 v2 

F . . . . . . . . . . . . 0 0 u3
   
3x

 
  
 


   
F3y

... ... ... ... ... 1.25

v3

May 15, 2023 48 / 62
Solution of a plane truss
Ex2: Solution

Boundary conditions:

u1 = δ u2 = u3 = 0 v2 = v3 = 0

Thus, we are left only with one equation (second line),

F1y = 1000 = 25200(0.48u1 + 1.89v1 )

Taking into account that P = 1000 kN and δ = −0.05 m we compute the

displacement v1 as

v1 = (2.01 × 10−5 )P − 0.254δ =⇒ v1 = 0.0337 m

where the positive value indicates horizontal displacement to the left.

The local element forces are obtained by using (10).

May 15, 2023 49 / 62
Solution of a plane truss
Ex2: Solution

Element 1:
 
u1 = −0.05
′ 
( ) 
f1x
   
1 −1 0.6 0.8 0 0 v1 = 0.0337
 
= 25200×
′
f2x −1 1 0 0 0.6 0.8  u =0
 2
 

v2 = 0


Which gives us
′ ′
f1x = −76.6 kN f2x = 76.6 kN
Element 2:
 
u1 = −0.05
′ 
( ) 
f1x
   
1 −1 0 1 0 0 v1 = 0.0337
 
= 31500 ×
′
f3x −1 1 0 0 0 1  u =0
 3
 

v3 = 0


Which gives us
′ ′
f1x = 1061 kN f3x = −1061 kN
May 15, 2023 50 / 62
Solution of a plane truss

Example 3. To illustrate how we can combine spring and bar elements in

one structure, we now solve the two-bar truss supported by a spring shown
below. Both bars have E = 210 GPa and A = 5.0 × 10−4 m2 . Bar one has
a length of 5 m and bar two a length of 10 m. The spring stiffness is
k = 2000 kN/m.

2
y
1 5m
25 kN
45◦
3 10 m x
1
2

3 k

May 15, 2023 51 / 62

Solution of a plane truss
Ex3: Solution

First, determine each element stiffness matrix.

√ √
◦ (1) 2 2
(1)
Element 1: θ = 135 , cos θ = − , sin θ(1) = .
2 2
 
0.5 −0.5 −0.5 0.5 u1
× 6 )(5.0 × 10−4 ) 
(210 10 −0.5 0.5 0.5 −0.5 
 v1
[k (1) ] = 
5  −0.5 0.5 0.5 −0.5  u2
0.5 −0.5 −0.5 0.5 v2
Simplifying,
 
1 −1 −1 1 u1
 −1 1 1 −1 
[k (1) ] = 10500   v1
 −1 1 1 −1  u2
1 −1 −1 1 v2

May 15, 2023 52 / 62

Solution of a plane truss
Ex3: Solution

Element 2: θ(2) = 180◦ , cos θ(2) = −1, sin θ(2) = 0.

EA (210 × 106 )(5.0 × 10−4 )

= = 10500
L(2) 10
 
1 0 −1 0 u1
 0 0 0 0 
[k (2) ] = 10500   v1
 −1 0 1 0  u3
0 0 0 0 v3

May 15, 2023 53 / 62

Solution of a plane truss
Ex3: Solution

Element 3: θ(2) = 270◦ , cos θ(3) = 0, sin θ(2) = −1.

Using (22) but replacing AE /L with the spring constant k, we obtain the
stiffness matrix of the spring as
 
0 0 0 0 u1
 0 1 0 −1   v1
[k (3) ] = 2000 
 0 0 0 0  u4
0 −1 0 1 v4

Applying the boundary conditions, we have

u2 = u3 = u4 = 0 v2 = v3 = v4 = 0

May 15, 2023 54 / 62

Solution of a plane truss
Ex3: Solution

The reduced governing equations are

    
F1x = 0 21000 −10500 u1
=
F1y = −25 −10500 12500 v1
Solving for the global displacements, we obtain
u1 = −1.724 × 10−3 m v1 = −3.448 × 10−3 m
The stresses in the bar elements are obtained using (30) as
210 × 103 
σ (1) =

0.707 − 0.707 − 0.707 0.707
 5
−1.724 × 10−3

 
−3.448 × 10−3

 

× = 51.2 MPa

 0 

0
 

May 15, 2023 55 / 62

Solution of a plane truss
Ex3: Solution

−1.724 × 10−3
 
 
210 × 103  −3.448 × 10−3

 

σ (2) =

1 0 −1 0 = −36.2 MPa
10 
 0 

0
 

May 15, 2023 56 / 62

Solution of a plane truss
Example 4. Find the displacements in the truss shown below. The truss is
composed of eight elements and five nodes as shown. A vertical load of
2P is applied at node 4. Nodes 1 and √ 5 are pin supports. Bar elements 1,
2, 7, and 8 have axial stiffnesses of 2EA, and bars 3, 4, 5 and 6 have
axial stiffness of EA.

L L
2 2

x′4
4 y 4
1 7 1 P
L

2P
x′1
1 3 6 x′3 3 4 x
5 1
4
x′2
5 5
8
L

2 2
x′5

a) b)
3 3

a) symmetrical plane truss; b) the truss reduced by symmetry.

May 15, 2023 57 / 62
Solution of a plane truss
Ex4: Solution

The vertical plane perpendicular to the plane truss passing through nodes
2, 4, and 3 is the plane of reflective symmetry because identical geometry,
material, loading, and boundary conditions occur at the corresponding
locations on opposite sides of this plane. For loads such as 2P, occurring
in the plane of symmetry, half of the total load must be applied to the
reduced structure.

For elements occurring in the plane of symmetry, half of the

cross-sectional area must be used in the reduced structure.
Furthermore, for nodes in the plane of symmetry, the displacement
components normal to the plane of symmetry must be set to zero
in the reduced structure; that is, we set u2 = 0, u3 = 0, and u4 = 0.

May 15, 2023 58 / 62

Solution of a plane truss
Ex4: Solution

 
xe − xb ye − yb
Data for the truss cos θ = , sin θ = :
L L

El. θ◦ c s c2 s2 cs Length
√ √ √
1 45◦ 2/2 2/2 1/2 1/2 1/2 2L
√ √ √
2 315◦ 2/2 − 2/2 1/2 1/2 −1/2 2L
3 0 1 0 1 0 0 L
4 270◦ 0 −1 0 1 0 L
5 90◦ 0 1 0 1 0 L

May 15, 2023 59 / 62

Solution of a plane truss
Ex4: Solution

There are a total of eight nodal components of displacement for the truss
before boundary constraints are imposed. Therefore, [K ] must be of order
8 × 8. Using (22) along with the table for the direction cosines, we write
down the local stiffness matrices.
 
0.5 0.5 −0.5 −0.5 u1
EA 
 0.5 0.5 −0.5 −0.5   v1
[k (1) ] = (33)
L  −0.5 −0.5 0.5 0.5  u2
−0.5 −0.5 0.5 0.5 v2
 
0.5 −0.5 −0.5 0.5 u1
[k (2) ] =  −0.5
EA  0.5 0.5 −0.5 
 v1 (34)
L  −0.5 0.5 0.5 −0.5  u3
0.5 −0.5 −0.5 0.5 v3

May 15, 2023 60 / 62

Solution of a plane truss
Ex4: Solution

 
1 0 −1 0 u1
EA  0 0 0 0 
 v1
[k (3) ] =

(35)
L  −1 0 1 0  u4
0 0 0 0 v4
 
0 0 0 0 u4
EA 0 0.5 0 −0.5 
 v4
[k (4) ] =

 (36)
L  0 0 0 0  u2
0 −0.5 0 0.5 v2
 
0 0 0 0 u3
EA  0 0.5 0 −0.5 
 v3
[k (5) ] =

(37)
L  0 0 0 0  u4
0 −0.5 0 0.5 v4

May 15, 2023 61 / 62

Solution of a plane truss
Ex4: Solution

Considering the boundary constraints that result in zero-displacement

components, we can immediately obtain the reduced set of equations by
eliminating rows and columns in each element stiffness matrix
corresponding to a zero-displacement component. That is, because u1 = 0,
v1 = 0 (support), and u2 = 0 , u3 = 0, and u4 = 0 (due to symmetry), we
can cancel rows and columns corresponding to these displacement
components in each element stiffness matrix before assembling the total
stiffness matrix. The resulting set of stiffness equations is
    
1 0 −0.5  v2   0 
AE 
0 1 −0.5  v3 = 0 (38)
L
−0.5 −0.5 1 v4 −P
   

Solving for the displacements, we get

PL PL 2PL
v2 = − v3 = − v4 = − (39)
EA EA EA
May 15, 2023 62 / 62
 Introduction to
Finite Element Method

Part Three

May 15, 2023

May 15, 2023 1 / 98

Analysis of beams
Beam stiffness

A beam is a long, slender structural member generally subjected to

transverse loading that produces significant bending effects as opposed to
twisting or axial effects.

This bending deformation is measured as a transverse displacement and a

rotation. Hence, the degrees of freedom considered per node are a
transverse displacement and a rotation.

y, v
m2 θ2
x
m1 , θ1 1 2
L
f1y , v1 f2y , v2

Beam element with positive nodal displacements, rotations, forces, and

moments.
May 15, 2023 2 / 98
Analysis of beams
Beam stiffness

Consider the beam element shown here. The beam is of length L with
axial local coordinate x and transverse local coordinate y . The local
transverse nodal displacements are given by vi and the rotations by θi .
The local nodal forces are given by fiy and the bending moments by mi as
shown. We initially neglect all axial effects. At all nodes, the following
sign conventions are used:

1 Moments are positive in the counter-clockwise direction.

2 Rotations are positive in the counter-clockwise direction.
3 Forces are positive in the positive y direction.
4 Displacements are positive in the positive y direction.

May 15, 2023 3 / 98

Analysis of beams
Derivation of the stiffness matrix. Beam equations.

From the course of the Strength of Materials we know that the differential
equation of a beam for constant EI and only nodal forces and moments
can be written as
d 4v
EI 4 = 0 (1)
dx
From elementary beam theory, we also know that the shear force and the
bending moment can be obtained by the integrating equation (1) as

d 3v d 2v
V (x) = EI and M(x) = EI (2)
dx 3 dx 2

May 15, 2023 4 / 98

Analysis of beams
Derivation of the stiffness matrix. Displacement function.

Assume the transverse displacement variation through the element length

to be
v (x) = a1 x 3 + a2 x 2 + a3 x + a4 (3)
The complete cubic displacement function (3) is appropriate because there
are four total degrees of freedom (vi and θi at each node). The cubic
function also satisfies the basic beam differential equation and the
conditions of displacement and slope continuity at nodes shared by two
elements.

May 15, 2023 5 / 98

Analysis of beams
Derivation of the stiffness matrix. Displacement function.

Now, we express v as a function of the nodal degrees of freedom v1 , v2 ,

θ1 , and θ2 as follows

v (0) = v1 = a4

dv (0)
= θ 1 = a3
dx
v (L) = v2 = a1 L3 + a2 L2 + a3 L + a4 (4)

dv (L)
= θ2 = 3a1 L2 + 2a2 L + a3
dx

May 15, 2023 6 / 98

Analysis of beams
Derivation of the stiffness matrix. Displacement function.

Solving (4) for a1 through a4 in terms of the nodal degrees of freedom and
substituting into (3), we have
 
2 1
v = 3 (v1 − v2 ) + 2 (θ1 + θ2 ) x 3
L L
 
3 1
+ − 2 (v1 − v2 ) − (2θ1 + θ2 ) x 2 + θ1 x + v1 (5)
L L

May 15, 2023 7 / 98

Analysis of beams
Derivation of the stiffness matrix. Elements of stiffness matrix.

We now relate the nodal and beam theory sign conventions for shear forces
and bending moments (Fig. 8), along with equations (5) and (2), to obtain

d 3 v (0) EI
f1y = V = EI 3
= 3 (12v1 + 6Lθ1 − 12v2 + 6Lθ2 )
dx L
d 2 v (0) EI
m1 = −m = −EI = 3 (6Lv1 + 4L2 θ1 − 6Lv2 + 2L2 θ2 )
dx 2 L

d 3 v (L) EI
f2y = −V = −EI 3
= 3 (−12v1 − 6Lθ1 + 12v2 − 6Lθ2 ) (6)
dx L
d 2 v (L) EI
m2 = m = EI = 3 (6Lv1 + 2L2 θ1 − 6Lv2 + 4L2 θ2 )
dx 2 L

May 15, 2023 8 / 98

Analysis of beams
Derivation of the stiffness matrix. Elements of stiffness matrix.

The minus signs in the second and third of (6) are the result of opposite
nodal and beam theory positive bending moment conventions at node 1
and opposite nodal and beam theory positive shear force conventions at
node 2. In matrix form, equations (6) become
    
 f1y  12 6L −12 6L  v1 
2 −6L 2  θ 
 
 EI   
m1 6L 4L 2L

1
=   (7)
f  L3  −12 −6L 12 −6L   v2 
 2y 

  
m2 6L 2L2 −6L 4L2 θ2
  

where the stiffness matrix is

 
12 6L −12 6L
EI  6L 4L2 −6L 2L2 
[k] = 3   (8)
L  −12 −6L 12 −6L 
6L 2L2 −6L 4L2
May 15, 2023 9 / 98
Analysis of beams
Derivation of the stiffness matrix. Elements of stiffness matrix.

In the beam element stiffness matrix (8), it is assumed that the beam is
long and slender; that is, the length, L, to depth, h, dimension ratio of the
beam is large. In this case, the deflection due to bending that is predicted
by using the stiffness matrix from (8) is quite adequate.

However, for short, deep beams the transverse shear deformation can be
significant and can have the same order of magnitude contribution to the
total deformation of the beam.

May 15, 2023 10 / 98

Analysis of beams
Example: Assemblage of beam stiffness matrices

Consider the beam in shown below as an example to illustrate the

procedure for assemblage of beam element stiffness matrices. Assume EI
to be constant throughout the beam. A force of 100 kN and a moment of
100 kNm are applied to the beam at mid span. The left end is a fixed
support and the right end is a pin support.

100 kNm
1 1 2 2 3 x

L L
100 kN

May 15, 2023 11 / 98

Analysis of beams
Example: Assemblage of beam stiffness matrices: Solution.

We discretize the beam into two elements with nodes 1 – 3 as shown. We

include a node at mid span because applied force and moment exist at mid
span and, at this time, loads are assumed to be applied only at nodes.

Using Eq. (8), we find that the global stiffness matrices for the two
elements are now given by
 
12 6L −12 6L v1
EI  6L 4L2 −6L 2L2   θ1
[k (1) ] = 3 
L  −12 −6L 12 −6L  v2
6L 2L2 −6L 4L2 θ2
 
12 6L −12 6L v2
EI  6L 4L2 −6L 2L2   θ2
[k (2) ] = 3 
L  −12 −6L 12 −6L  v3
6L 2L2 −6L 4L2 θ3
May 15, 2023 12 / 98
Analysis of beams
Example: Assemblage of beam stiffness matrices: Solution.

The total stiffness matrix can now be assembled for the beam by using the
direct stiffness method. For clarity, we show the position of each local
stiffness matrix in the global one, then add them up.

6L −12
 
12 6L 0 0 v1
 6L 4L2 −6L 2L2 0 0  θ1
 
(1) EI  −12 −6L 12 −6L 0 0  v2
[k ] = 3  2 2

L  6L 2L −6L 4L 0 0 

 θ2
 0 0 0 0 0 0  v3
0 0 0 0 0 0 θ3

May 15, 2023 13 / 98

Analysis of beams
Example: Assemblage of beam stiffness matrices: Solution.

 
0 0 0 0 0 0 v1

 0 0 0 0 0 0  θ1

(2) EI  0 0 12 6L −12 6L  v2
[k ] = 3  
L  0 0 6L 4L2 −6L 2L2  θ2

 0 0 −12 −6L 12 −6L  v3
0 0 6L 2L2 −6L 4L2 θ3
Thus, the governing equations for the beam are obtained as
   
−12

 F1y  12 6L 6L 0 0  v1 
2 −6L 2L2
   



 M 1 


  6L 4L 0 0 

 θ1 


 EI   −12 −6L
 
F2y 24 0 −12 6L v2
  
= 3 2


 M2   L  6L 2L 0 8L2 −6L 2L2 
 θ2 

F  0 0 −12 −6L 12 −6L v3
   
 3y 

   
 

  
M3 6L 2L2 −6L 4L2

0 0

θ3


May 15, 2023 14 / 98

Analysis of beams
Example: Assemblage of beam stiffness matrices: Solution.

Boundary conditions (constraints):

v1 = 0 θ1 = 0 v3 = 0

By deleting the first, the second and the fifth rows and columns, we obtain
    
 −100  EI 24 0 6L  v2 
100 =  0 8L2 2L2  θ2
 L3
0 6L 2L2 4L2 θ3
  

where F2y = −100 kN, M2 = 100 kNm and M3 = 0. The above equations
are solved simultaneously for the nodal displacements v2 , θ2 and θ3 . The
final solution is left for you to obtain independently.

May 15, 2023 15 / 98

Analysis of beams
Examples: Beam analysis.

Example 1. Using the direct stiffness method, solve the problem of the
propped cantilever beam subjected to end load P shown in the figure
below. The beam is assumed to have constant flexural stiffness EI and
length 2L. It is supported by a roller at mid-length and is built in at the
right end.

P
2 3
1
L L

May 15, 2023 16 / 98

Analysis of beams
Ex1: Solution.

We will determine the nodal displacements and rotations, the reactions,

and the complete shear force and bending moment diagrams.

Quite clear, that the stiffness matrix will be the same as in the previous
example. The governing equations for the beam are then given by
   
6L −12

 F1y  12 6L 0 0  v1 
 6L 4L2 −6L 2L2

 M    


 1 

  0 0  



 θ1 


−12 −6L −12
 EI 
F2y 24 0 6L v
  
2

= 3  

 M2   L  6L 2L2 0 8L2 −6L 2L2   θ2 
F 0 0 −12 −6L 12 −6L v3 
   
3y 

 
   
 
  
6L 2L2 −6L 4L2

M3

0 0

θ3


May 15, 2023 17 / 98

Analysis of beams
Ex1: Solution.

On applying the boundary conditions

v2 = 0 v3 = 0 θ3 = 0

and partitioning the equations associated with unknown displacements

(the first, second, and fourth equations) from those equations associated
with known displacements in the usual manner, we obtain the final set of
equations as
    
 −P  EI 12 6L 6L  v1 
0 =  6L 4L2 2L2  θ1
 L3 2 2
0 6L 2L 8L θ2
  

where F1y = −P, M1 = 0, and M2 = 0. We will now solve it for the nodal
displacement and nodal slopes.
May 15, 2023 18 / 98
Analysis of beams
Ex1: Solution.

Node 1:
7PL3 3PL2
v1 = − θ1 =
12EI 4EI
Node 2:
PL2
θ2 =
4EI
Where the minus sign indicates that the displacement is downwards. The
positive sign of the slopes θ1 and θ2 indicate counter-clock rotation.

May 15, 2023 19 / 98

Analysis of beams
Ex1: Solution.

Next,determine the global nodal forces.

7PL3 
 

 −
 12EI 
 

   

−12

 F1y  12 6L 6L 0 0 

 3PL 2 

 6L 4L2 −6L 2L2
   



 M1 


  0 0  


 4EI




F2y EI  −12 −6L 24 0 −12 6L 
    
= 3 0
M2 L  6L 2L2 0 8L2 −6L 2L2 
PL2 

 
   
 F3y 0 0 −12 −6L 12 −6L 

 
   
 

 
4EI
   

M3

0 0 2
6L 2L −6L 4L 2 





 



 0 



 
0

May 15, 2023 20 / 98

Analysis of beams
Ex1: Solution.

Multiplying the matrices on the right-hand side, we obtain the global

nodal forces and moments as
F1y = −P M1 = 0 F2y = 2.5P
M2 = 0 F3y = −1.5P M3 = 0.5PL

It is generally necessary to determine the local nodal forces associated with

each element of a large structure to perform a stress analysis of the entire
structure. treated similarly).

May 15, 2023 21 / 98

Analysis of beams
Ex1: Solution.

Using the {f } = [k]{u} equation for each element, we have

7PL3
 
−
 (1)  
 


 f1y    

 12EI



  12 6L −12 6L  
 3PL2
   

 m (1)   
6L 4L2 −6L 2L2 
 
 EI   

1
= 3   4EI
 f2y
  L
(1)  −12 −6L 12 −6L   
0
 
6L 2L2 −6L 4L2 

 
  

   
 (1) 
  2

m2 PL
 
 


 

4EI
Which gives us

f1y = −P m1 = 0 f2y = P m2 = −PL

May 15, 2023 22 / 98

Analysis of beams
Ex1: Solution.

 (2)   

 f2y 
   0 

 
 12 6L −12 6L 

PL2 

 m(2)
   
2 −6L 2
   
2
 EI  6L 4L 2L  
= 3  4EI
 (2)
f3y  L  −12 −6L 12 −6L  
 
2 −6L
 0 
6L 2L 4L2

 
 
 

   
 (2)
  
0

m3


Which gives us

f2y = 1.5P m2 = PL f3y = −1.5P m3 = 0.5PL

May 15, 2023 23 / 98

Analysis of beams
Ex1: Solution.

PL PL 0.5PL
1 1 2 2 2 3

L L
P P 1.5P 1.5P

1.5P 1.5P
SFD

−P −P
PL

BM

0.5PL

May 15, 2023 24 / 98

Analysis of beams
Examples: Beam analysis.

Example 2. Determine the displacement and rotation under the force and
moment located at the center of the beam shown in the figure. The beam
has been discretized into the two elements. It is fixed at each end. A
downward force of 10 kN and an applied moment of 20 kNm act at the
center of the beam. Let E = 210 GPa and I = 4 × 10−4 m4 throughout
the beam length.

10 kN

1 1 2 2 3

3m 20 kNm 3m

May 15, 2023 25 / 98

Analysis of beams
Ex2: Solution.

The governing equation will look identical to those in the Example 1.

   
6L −12

 F1y  12 6L 0 0  v1 
4L2 −6L 2L2
   

 M
 1 
 



 6L 0 0  



 θ 1




F2y EI  −12 −6L 24 0 −12 6L  v2
   
= 3

 M 2 
 L  6L
 2L2 0 8L2 −6L 2L2   θ2 
F  0 0 −12 −6L 12 −6L  v3 
   
 3y 

     
  
M3 6L 2L2 −6L 4L2

0 0

θ3


The boundary conditions are v1 = v3 = 0 and θ1 = θ3 = 0.

The global forces are F2y = −10 kN and M2 = 20 kNm.

May 15, 2023 26 / 98

Analysis of beams
Ex2: Solution.

Applying the boundary conditions and the global forces we arrive at

(210 × 106 )(4 × 10−4 ) 28000

=
27 9
    
−10 28000 24 0 v2
=
20 9 0 72 θ2
Solving for the displacements, we obtain

v2 = −1.339 × 10−4 m θ2 = 8.928 × 10−5 rad

May 15, 2023 27 / 98

Analysis of beams
Ex2: Solution.

Using the local equations for each element, we obtain the local nodal
forces and moments for element one as follows:
 (1) 
f
 1y 

    

 (1)


 12 18 −12 18  0 
 m
   
−18
 28000 
18 36 18  0

1
=
−1.339 × 10−4
 

 (1)
f2y   9  −12 −18 12 −18  
 

18 18 −18 36 8.928 × 10−5

 
  
 
 (1) 

m2


Calculating, we have
(1) (1) (1) (1)
f1y = 10 kN m1 = 12.5 kNm f2y = −10 kN m2 = 17.5 kNm

May 15, 2023 28 / 98

Analysis of beams
Ex2: Solution.

Similarly, for element two the local nodal forces and moments are
 (2) 
 f2y 
−1.339 × 10−4
    

 
 12 18 −12 18 
 m(2)
  
8.928 × 10−5
   
36 −18
 28000 
 18 18 
 
2
= 

 (2)
f3y 
 9  −12 −18 12 −18    0 

18 18 −18 36 0

 
  
 
 (2)
 
m3


(2) (2) (2) (2)

f2y = 0 m2 = 2.5 kNm f3y = 0 m3 = −2.5 kNm

May 15, 2023 29 / 98

Analysis of beams
Ex2: Solution.

10

SFD

−12.5
−2.5
BM

17.5

May 15, 2023 30 / 98

Analysis of beams
Examples: Beam analysis.

Example 3. Determine the nodal displacements and rotations and the

global and element forces for the beam shown in below. The beam is
discretized as shown by the node numbering. It is fixed at node 1, has a
roller support at node 2, and has an elastic spring support at node 3. A
downward vertical force of P = 50 kN is applied at node 3. Let
E = 210 GPa and I = 2 × 10−4 m4 throughout the beam, and let
k = 200 kN/m.
P = 50 kN
1 1 2 2 3

3 k = 200 kN/m
3m 3m
4

May 15, 2023 31 / 98

Analysis of beams
Ex3: Solution.

Using (8) for each beam element and the stiffness matrix for the spring
element as well as the direct stiffness method, we can obtain the global
stiffness matrix. Since, different types of local stiffness matrices are used,
we show them separately.

6L −12
 
12 6L 0 0 0 v1
 6L 4L2 −6L 2L2 0 0 0  θ1
 
 −12 −6L 12 −6L 0 0 0 
(1) EI 
 2 2
 v2
[k ] = 3  6L 2L −6L 4L 0 0 0   θ2
L 
 0 0 0 0 0 0 0  v3

 0 0 0 0 0 0 0  θ3
0 0 0 0 0 0 0 v4

May 15, 2023 32 / 98

Analysis of beams
Ex3: Solution.
 
0 0 0 0 0 0 0 v1

 0 0 0 0 0 0 0  θ1

 0 0 12 6L −12 6L 0  v2
(2) EI  
[k ] = 3  0 0 6L 4L2 −6L 2L2 0  θ2
L  

 0 0 −12 −6L 12 −6L 0  v3

 0 0 6L 2L2 −6L 4L2 0  θ3
0 0 0 0 0 0 0 v4
 
0 0 0 0 0 0 0 v1

 0 0 0 0 0 0 0 
 θ1
 0 0 0 0 0 0 0 
 v2
(3)

[k ] = 
 0 0 0 0 0 0 0 
 θ2

 0 0 0 0 k 0 −k  v3
 0 0 0 0 0 0 0  θ3
0 0 0 0 −k 0 k v4
May 15, 2023 33 / 98
Analysis of beams
Ex3: Solution.

EI
For the global matrix we use a common factor . Therefore, me must
L3
L3
multiply each component of matrix [k (3) ] by . To simplify the notation
EI
kL3
we will use k ′ =
.
EI
6L −12
    
 F1y  12 6L 0 0 0  v1 
2 −6L 2
   



 M 
1 



 6L 4L 2L 0 0 0 

 θ
 1
 



 F2y  −12 −6L 24 0 −12 6L 0   v2

   
 EI 

 

M2 6L 2L 2 0 8L2 −6L 2L 2 0 
= 3
L    θ2

 F3y 

  0 ′ ′
0 −12 −6L 12 + k −6L −k   v3 


    


 M

 0 0 6L 2L2 −6L 4L2 0

 θ


3
 3
     
   

F4y

0 0 0 0 −k ′ 0 k ′ v4


May 15, 2023 34 / 98

Analysis of beams
Ex3: Solution.

Let us apply the boundary conditions

v1 = v2 = v4 = 0 θ1 = 0

We delete the first three equations and the seventh equation The
remaining three equations are

8L2 −6L 2L2  θ2 

    
 0  EI
−P =  −6L 12 + k ′ −6L  v3
 L3
0 2L2 −6L 4L2 θ3
  

Solving simultaneously, we obtain

7PL3 3PL2
   
1 1
v3 = − θ2 = −
EI 12 + 7k ′ EI 12 + 7k ′

May 15, 2023 35 / 98

Analysis of beams
Ex3: Solution.

9PL2
 
1
θ3 = −
EI 12 + 7k ′
By substituting the given numerical values, we obtain

200 × 33
k′ = = 0.1286
(210 × 106 )(2 × 10−4 )

1 1 10
= =
12 + 7k ′ 12 + 7 × 0.1289 129

7 × 50 × 33 10 3
v3 = − 6 −4)
× =− = 0.01744 m
(210 × 10 )(2 × 10 129 172

May 15, 2023 36 / 98

Analysis of beams
Ex3: Solution.

PL2 50 × 32
 
1 10
= ×
EI 12 + 7k ′ 6 −4
(210 × 10 )(2 × 10 ) 129
1
= = 8.3056 × 10−4
1204
Finally,
1 3
θ2 = − ×3=− = −0.00249 rad
1204 1204
9
θ3 = 3θ2 = − = −0.00747 rad
1204

May 15, 2023 37 / 98

Analysis of beams
Ex3: Solution.

We now back-substitute the obtained displacements into the governing

equations, and obtain the global nodal forces.
    
 F1y  12 6L −12 6L 0 0 0  0 
4L2 2L2
   



 M1





 6L −6L 0 0 0 


 0




 F2y −12 −6L 24 0 −12 6L 0 0

 
   

EI 
  
M2 = 3 6L 2L2 0 8L2 −6L 2L2 0  −3/1204
L 
12 + k ′ −k ′

F 0 0 −12 −6L −6L  −3/172 
  
3y

 
   
    
 M3 0 0 6L 2L2 −6L 4L2 0 −9/1204 

 
   
 
 
 

F4y 0 0 0 0 −k ′ 0 k′ 0
   

May 15, 2023 38 / 98

Analysis of beams
Ex3: Solution.

Which gives us

F1y = −69.8 kN M1 = −69.8 kNm

F2y = 116.3 kN M2 = 0
F3y = −50.0 kN M3 = 0
F4y = 3.5 kN

Check for the balance of forces:

X
Fy = 116.3 + 3.5 − 69.8 − 50 = 0
X
M1 = 50 × 6 − 116.3 × 3 − 3.5 × 6 = −69.9 ≈ −69.8 kNm

May 15, 2023 39 / 98

Analysis of beams
Examples: Beam analysis.

Example 4. Determine the nodal displacements and rotations, global

nodal forces, and element forces for the beam shown in the figure. The
beam is discretized as indicated by the node numbering. The beam is fixed
at nodes 1 and 5 and has a roller support at node 3. Vertical loads of
45 kN each are applied at nodes 2 and 4. Let 210 GPa and
I = 2.08 × 10−4 m4 . throughout the beam.

45 kN 45 kN

1 2 3 4 5

3m 3m 3m 3m

May 15, 2023 40 / 98

Analysis of beams
Ex4: Solution.

Here the lengths and stiffness of each element are the same. Thus, we can
EI
factor an 3 out of the superimposed (global) stiffness matrix.
L
12 6L −12 6L 0 0 0 0 0 0
    
F1y v1
4L2 2L2
   


 M1


  6L −6L 0 0 0 0 0 0 

 θ1



 −12 −6L −12
    


 F2y


 24 0 6L 0 0 0 0  v2


 
2L2 8L2 2L2
    
M2 6L 0 −6L 0 0 0 0 θ2

 
   

   
EI 
    

F3y
 0 0 −12 −6L 24 0 −12 6L 0 0  v3

=
2L2 8L2 2L2
 

 M3 
 3
L 
 0 0 6L 0 −6L 0 0 
 θ3 




 F4y 




 0 0 0 0 −12 −6L 24 0 −12 6L 


 v4 



M4 2L2 8L2 −6L 2L2 θ4
   







 0 0 0 0 6L 0 








 F5y 



 0 0 0 0 0 0 −12 −6L 12 −6L 

 v5 



M5 0 0 0 0 0 0 6L 2L2 −6L 4L2 θ5

Taking into account the boundary conditions,

v1 = v3 = v5 = 0 θ1 = θ5 = 0

the resulting system of equations is given next

May 15, 2023 41 / 98
Analysis of beams
Ex4: Solution.

   
−45 

 24 0 6L 0 0  v2 
2 2
   
 0  0 8L 2L 0 0 θ2

   
 EI  
 

0 = 3  6L 2L
 2 8L −6L 2L2
2  θ3
L  
−45  0 0 −6L 24 0 v4

 
  


  
 

0 0 0 2L2 0 8L2 θ4
   

The rotations (slopes) at nodes 2, 3 and 4 are equal to zero because of

symmetry in loading, geometry, and material properties. Therefore,
θ2 = θ3 = θ4 = 0, and we can further reduce the system of equations to
    
−45 EI 24 0 v2
= 3
−45 L 0 24 v4

May 15, 2023 42 / 98

Analysis of beams
Ex4: Solution.

Moreover, since v2 = v4 , it is enough to solve only one equation

EI (210 × 106 )(2.08 × 10−4 )

× 24v2 = −45 or × 24v2 = −45
L3 27
EI (210 × 106 )(2.08 × 10−4 )
Where = = 1617.7778.
L3 27
Solving for the displacement,

v2 = v4 = −1.159 × 10−3 m = −1.16 mm ↓

May 15, 2023 43 / 98

Analysis of beams
Ex4: Solution.

We can now back-substitute the obtained displacements, along with the

numerical values for E , I , and L, into the global system of equations and
determine the global nodal forces.

12 6L −12 6L 0 0 0 0 0 0
    
F1y 0
4L2 2L2
   


 M1


  6L −6L 0 0 0 0 0 0 

 0



 −12 −6L −12
    


 F2y


 24 0 6L 0 0 0 0  −0.001159 
 

2L2 8L2 2L2
    
M2 6L 0 −6L 0 0 0 0 0

 
   

   
EI 
    

F3y
 0 0 −12 −6L 24 0 −12 6L 0 0  0

=
2L2 8L2 2L2
 

 M3 
 L3 
 0 0 6L 0 −6L 0 0 
 0 




 F4y 




 0 0 0 0 −12 −6L 24 0 −12 6L 


 −0.001159 



M4 2L2 8L2 −6L 2L2 0
   







 0 0 0 0 6L 0 







 F5y 



 0 0 0 0 0 0 −12 −6L 12 −6L 

 0 



M5 0 0 0 0 0 0 6L 2L2 −6L 4L2 0

May 15, 2023 44 / 98

Analysis of beams
Ex4: Solution.

F1y = 1617.7778 × (−12) × (−0.001159) = 22.5 kN

M1 = 1617.7778 × (−6) × 3 × (−0.001159) = 33.75 kNm
F2y = 1617.7778 × 24 × (−0.001159) = −45 kN
M2 = 0
 
F3y = 1617.7778 × (−12) × (−0.001159) + (−12) × (−0.001159) = 45 kN
M3 = 0
F4y = 1617.7778 × 24 × (−0.001159) = −45 kN
M4 = 0
F5y = 1617.7778 × (−12) × (−0.001159) = 22.5 kN
M5 = 1617.7778 × 6 × 3 × (−0.001159) = −33.75 kNm

May 15, 2023 45 / 98

Analysis of beams
Ex4: Solution.

The solution for the external forces can be considerably simplified if the
columns corresponding to zero displacements are deleted in the stiffness
matrix, i.e.:

−12
 
F1y  0
 

 −6L

 M  




1 
  0 

 F2y 



  24 0 
   



 M 2




 0 0 
 EI   
F3y −12 −12  −0.001159
 
= 3  

 M3  L  6L −6L   −0.001159
 F4y  0 24 

 
 
   

M



 4




 0 0 

0 −12 
 


 F5y 



 
M5 0 6L

May 15, 2023 46 / 98

Analysis of beams
Ex4: Solution.

Once again, the global nodal forces (and moments) at the support nodes
(nodes 1, 3, and 5) can be interpreted as the reaction forces, and the
global nodal forces at nodes 2 and 4 are the applied external forces.

For large structures we must obtain the local element shear force and
bending moment at each node end of the element because these values are
used in the design/analysis process. They can be obtained using the local
equations for each element. For example, for the element 1 we have
 (1) 
f
 1y 

    

 (1)


 12 6L −12 6L  0 
 m
 
 EI  2 −6L 2 
 
6L 4L 2L 0

1
= 3  

 (1)
f2y   L −12 −6L 12 −6L    −0.001159  





 6L 2L2 −6L 4L 2 
0

 (1) 
 
m2

May 15, 2023 47 / 98

Analysis of beams
Ex4: Solution.

Which gives us
(1) (1)
f1y = 22.5 kN m1 = 33.75 kNm
(1) (1)
f2y = −22.5 kN m2 = 33.75 kNm
Similarly, the rest of internal forces can be calculated.

May 15, 2023 48 / 98

Analysis of beams
Distributed loading

Beam members can support distributed loading as well as concentrated

nodal loading. Therefore, we must be able to account for distributed
loading. Consider the fixed-fixed beam subjected to a uniformly distributed
loading w shown here. The reactions, determined from structural analysis
theory, are shown in the figure.
w (N/m)
A B

L
wL2 wL2
MA = MB =
12 12
A B
wL wL
RA = RB =
2 2
Fixed-fixed beam subjected to a uniformly distributed load.

May 15, 2023 49 / 98

Analysis of beams
Distributed loading

Since the finite element method can only handle the nodal loads, we
replace the load by concentrated nodal forces and moments tending to
have the same effect on the beam as the actual distributed load. Next
figure illustrates this idea for a beam. We have replaced the uniformly
distributed load by a statically equivalent force system consisting of a
concentrated nodal force and moment at each end of the member carrying
the distributed load.

wL wL
2 2
wL2 wL2
12 12
Example of the equivalent load. The whole beam is considered as a finite
element.

May 15, 2023 50 / 98

Analysis of beams
Distributed loading

wL wL wL
+ wL
2 2 2
2
wL2 wL2
12 12
wL2 wL2
12 12
Example of the equivalent load. The whole beam is divided into two finite
elements.

These statically equivalent forces are always of opposite sign from the
fixed-end reactions. If we want to analyse the behaviour of loaded member
in better detail, we can place a node at midspan and use the same
procedure.
May 15, 2023 51 / 98
Analysis of beams
Distributed loading

In general, we can account for distributed loads or concentrated loads

acting on beam elements by starting with the following formulation
application for a general structure:

{F} = [K]{u} − {F0 } (9)

where {F} are the concentrated nodal forces and {F0 } are called the
equivalent nodal forces, now expressed in terms of global-coordinate
components, which are of such magnitude that they yield the same
displacements at the nodes as would the distributed load. Because we now
assume that concentrated nodal forces are not present ({F} = 0), as we
are solving beam problems with distributed loading only here, we can write
(9) as
{F0 } = [K]{u} (10)

May 15, 2023 52 / 98

Analysis of beams
Distributed loading

For a uniformly distributed load w acting over a one-element beam, we

have  wL 
 − 
2 

 

 

2
 

 wL 

 −

 

12

{F0 } = (11)
wL 
 −

 
 
2 
 

 

 2 
 wL 

 

12
This concept can be applied on a local basis to obtain the local nodal
forces {f } in individual elements of structures by applying(9) locally as
{f } = [k]{u} − {f0 } (12)
where {f0 } are the equivalent local nodal forces.
May 15, 2023 53 / 98
Analysis of beams
Examples: Distributed load

Example 1. For the cantilever beam subjected to the uniform load w ,

solve for the right-end vertical displacement and rotation and then for the
nodal forces. Assume the beam to have constant bending stiffness EI
throughout its length.

wL wL
y − −
2 2
w
x 1
L
L wL2 wL2
−
12 12

May 15, 2023 54 / 98

Analysis of beams
Ex1: Distributed load: Solution.

Using equations (10) and the beam element stiffness matrix, we obtain
 wL 
 − 
2 

 

 

2 
    
12 6L −12 6L  v1 
 wL 
 −
  
2 2
  
 
EI  6L 4L −6L 2L 
  
θ1 12

= (13)
L3  −12 −6L 12 −6L   v wL 
 2 −
 
 
2 2   
6L 2L −6L 4L θ2
 
2

 


 


 wL 
 2 

 
12
By applying the boundary conditions

v1 = 0 and θ1 = 0

May 15, 2023 55 / 98

Analysis of beams
Ex1: Distributed load: Solution.

we arrive at
 
wL
    − 

EI 12 −6L v2 2
 
= (14)
L3 −6L 4L2 θ2 2
 wL

 


12
Solving 14) for the displacements, we obtain
 
wL
 −
 
2L2
    
v2 L 3L 2

= 2
(15)
θ2 6EI 3L 6 
 wL 

 
12

May 15, 2023 56 / 98

Analysis of beams
Ex1: Distributed load: Solution.

Which gives us the displacement and slope as

wL4
 
   −
  

v2 8EI

= (16)
θ2 3
 − wL

 


6EI
The negative signs indicate that v2 is downward and θ2 is clockwise. In
this case, the method of replacing the distributed load by discrete
concentrated loads gives exact solutions for the displacement and rotation
as could be obtained by classical methods.

May 15, 2023 57 / 98

Analysis of beams
Ex1: Distributed load: Solution.

Next, we obtain the effective global nodal forces.

   wL 
 (e) 0  
2 
 
 
 
 
 F1y  
 
 
 

2
      

 
 12 6L −12 6L 
 0

 
 5wL 

 M (e)
     
2 −6L 2
     
1
 EI  6L 4L 2L    
12

= 3  wL4 =
(e)
F2y  L  −12 −6L 12 −6L  − wL 
−
  

8EI 
    
 
6L 2L2 −6L 4L2
  
2

 
 
 
 
 

 (e) 3
     
M2
 
 wL 
 
 2


 −
 
   wL 

6EI
 
12
(17)

May 15, 2023 58 / 98

Analysis of beams
Ex1: Distributed load: Solution.

We then use (11) and (17) in (9) to obtain the correct global nodal forces
as  wL   wL 
   − 
2 2
     
wL
     
F
   
1y 
     
2  2 
       

  
 5wL  
 wL  
 2


− wL
       
M1
        
12 12
      
= − = 2 (18)
F2y  wL  wL 
− − 0

  
  
  
 

       
2  2 

 
      
    
M2
      
0
     


 wL 2 




 wL 2 


   
12 12
Concluding remarks to the solved problem: The nodal force and
moment reactions obtained by equations (18) illustrate the importance of
using (9) to obtain the correct global nodal forces and moments.

May 15, 2023 59 / 98

Analysis of beams
Ex1: Distributed load: Solution.

The general method can be described as follows:

1 Replace the distributed load by its work-equivalent to identify the

nodal force and moment used in the solution.
2 Assemble the global force and stiffness matrices and global equations
illustrated by (13).
3 Apply the boundary conditions to reduce the set of equations as done
in previous problems and illustrated by (14) where the original four
equations have been reduced to two equations to be solved for the
unknown displacement and rotation.
4 Solve for the unknown displacement and rotation given by (15) and
(16).
5 Use (9) as illustrated by (18) to obtain the final correct global nodal
forces and moments. Those forces and moments at supports will be
the reactions in this problem.
May 15, 2023 60 / 98
Analysis of beams
Examples: Distributed load

Example 2. For the cantilever beam subjected to the concentrated load

P, solve for the right-end vertical displacement and slope, and the nodal
forces, including reactions, by replacing the concentrated load with
equivalent nodal forces acting at each end of the beam. Assume EI
constant throughout the beam.

P P
y − −
2 2

L/2 L/2 x 1
L PL
PL
−
8 8

May 15, 2023 61 / 98

Analysis of beams
Ex2: Distributed load: Solution.

We begin by discretizing the beam. Here only one element is used with
nodes at each end of the beam. We then replace the concentrated load as
shown in the above figure.
 P 
     − 
EI 12 −6L v2 2

= (19)
L3 −6L 4L2 θ2  PL 
 
8
where the boundary conditions v1 = 0 and θ1 = 0 have been taken into
account. Solving (19) for displacements, we obtain
 P 
 
L
 2  − 

v2 2L 3L 2

= (20)
θ2 6EI 3L 6   PL  
8
May 15, 2023 62 / 98
Analysis of beams
Ex2: Distributed load: Solution.

Simplifying,
3
 
5PL
 −
    

v2 48EI

= (21)
θ2  PL 2 
 −
 

8EI
To obtain the unknown nodal forces, we begin by evaluating the effective
nodal forces {F (e)} = [K ]{u} as   P 
 
 (e)   0  
 

 F 1y 
  











 2 





 (e)


 12 6L −12 6L 

 0





 3PL 


 M
  2 2
   
−6L
 EI 
6L 4L 2L 8
   
1 
= 3   5L3 =
(e)
F2y  L −12 −6L 12 −6L   − P 
− 
 
 
48EI 
   
 
6L 2L 2 −6L 4L 2   
2

 
 
 
 
 

 (e)  2
     
M2


 −
 PL 


 


 PL




8EI
 
8
May 15, 2023
(22)
63 / 98
Analysis of beams
Ex2: Distributed load: Solution.

The correct nodal forces can now be obtained by applying (9) as

P   P 
   

   −   
   2   2 
P
 F1y 

 
 
 
  
    3PL     
   
 
 PL 
 
 


 M1  
 
 
   −

 
 
 
 PL 

8 8
  
= − = 2 (23)
F2y  P   P  0 
−  −

  
   
 
       
2 2

 
 
 
 
 
 
 

M2
    
 
 
 
 0  
 PL   PL 

 
 
 

   
8 8

May 15, 2023 64 / 98

Analysis of beams
Examples: Distributed load

Example 3. For the cantilever beam subjected to the concentrated

free-end load P and the uniformly distributed load w acting over the whole
beam as shown in the figure, determine the free-end displacements and the
nodal forces.

wL wL
y P − − −P
2 2
w
x 1
2 L
L wL wL2
−
12 12

May 15, 2023 65 / 98

Analysis of beams
Ex3: Distributed load: Solution.

Using the beam element stiffness (8), we obtain

 
wL
     − −P 

EI 12 −6L v2 2
 
= (24)
L3 −6L 4L2 θ2 
 wL2 

 
12
where the boundary conditions v1 = 0 and θ1 = 0 have been taken into
account. Solving (24) for displacements, we obtain
4 3
 
wL PL
 − −
    

v2 8EI 3EI

= (25)
θ2 3 2 
 − wL − PL 

 
6EI 2EI

May 15, 2023 66 / 98

Analysis of beams
Ex3: Distributed load: Solution.

Next, we obtain the effective nodal forces using {F (e) = [K ]{u} as

wL 
 
0 P+
  
 
 (e)
F1y
 







 2 

     
      2 

 
 12 6L −12 6L

 0 
 
 5wL 

 M1(e)
     PL + 
4L2 2L2
     
EI  6L −6L 12
    
= 3   wL
 4 3
PL  =
 (e)
F2y  L  −12 −6L 12 −6L − − wL 

 

6L 2L2 −6L 4L2

 8EI
 
3EI   −P − 

2

 
 
 
 
 

 (e)
    
   3 2 
   
M2  − wL −
 PL  2

wL
   
 


6EI 2EI
 
12
(26)

May 15, 2023 67 / 98

Analysis of beams
Ex3: Distributed load: Solution.

Finally, subtracting the equivalent nodal force matrix from effective force
matrix, we obtain the correct nodal forces as
 wL   wL 
 P +   − 
2 2 
     
P + wL
    
F1y 

 
 
 
  
2  2 
      

 
 
 5wL  
 wL  
 2


 M1   PL +   − wL

 
 
 
 
 
 
 

12 12 PL +
  
= − = 2 (27)
F2y  wL  wL 
 −P − − −P

  
  
  
 

       
2  2 

 
     
    
M2
      
0
     


 wL 2 




 wL 2 


   
12 12

May 15, 2023 68 / 98

Analysis of beams
Examples: Distributed load

Example 4. For the fixed–fixed beam subjected to the linear varying

distributed loading acting over the whole beam shown in the figure below
determine the displacement and rotation at the center and the reactions.

w
y w/2

L L

May 15, 2023 69 / 98

Analysis of beams
Ex4: Distributed load: Solution.

The beam is now modelled using two elements with nodes 1, 2, and 3 and
the distributed load is replaced as shown below. The first element (1 – 2)
has a triangular load. The second element (2 – 3) is loaded with a
uniformly distributed load and a triangular load.

w
w/2
w/2
L L
1 2 3

The equivalent joint forces f0 are shown below for both the types of
loading.
May 15, 2023 70 / 98
Analysis of beams
Ex4: Distributed load: Solution.

w
w

L 7wL wL L wL
3wL − − −
− 20 2 2
20

wL2 wL2 wL2 wL2

− −
30 20 12 12

Taking into account that the load magnitude in our case is w /2, we obtain
the following equivalent nodal force replacement system.

Element 1:
(1) 3 (1) 1 (1) 7 (1) 1
f1y = − wL m1 = − wL2 f2y = − wL m2 = wL2
40 60 40 40
(28)
May 15, 2023 71 / 98
Analysis of beams
Ex4: Distributed load: Solution.

Element 2:
(2) 1 3 13 (2) 1 1 7
f2y = − wL− wL = − wL m2 = − wL2 − wL2 = − wL2
4 40 40 24 60 120
(29)
(2) 1 7 17 (2) 1 1 1
f3y = − wL− wL = − wL m3 = − wL2 − wL2 = − wL2
4 40 40 24 40 15
(30)
Using the beam element stiffness for each element, we obtain
 
12 6L −12 6L
EI  6L 4L2 −6L 2L2 
[k (1) ] = [k (2) ] = 3   (31)
L  −12 −6L 12 −6L 
6L 2L2 −6L 4L2

The boundary conditions are v1 = 0, θ1 = 0, v3 = 0, and θ3 = 0.

May 15, 2023 72 / 98
Analysis of beams
Ex4: Distributed load: Solution.

Using the direct stiffness method to assemble the global stiffness matrix,
and applying the boundary conditions, we obtain
 
wL
 −
  ( )
F2y
    
EI 24 0 v2 2

= = (32)
L3 0 8L2 θ2 2 
 wL  M2

 
30
Solving (32) for displacements, we obtain

wL4 wL3
v2 = − θ2 = − (33)
48EI 240EI

May 15, 2023 73 / 98

Analysis of beams
Ex4: Distributed load: Solution.

Next, we obtain the effective nodal forces.

 

(e)
  0 


 F1y 











 (e)

 
 0 

6L −12
 


 M1


 12 6L 0 0 




 6L 4L2 −6L 2L2
   

 
 0 0 
wL 4 

 F (e) −

 
   

EI  −12 −6L 24 0 −12 6L
  
2y 
48EI
= 3 2
 =
 M2
(e)  L  6L 2L 0 8L2 6L 2L2 
wL 3 
   − 
0 0 −12 −6L 12 −6L

 
   

240EI
   
(e)
   
F3y 6L 2L −6L 4L2
2
 
0 0
  
 

   








 0 


 (e)
   
M3
 
 

 
0

(34)
May 15, 2023 74 / 98
Analysis of beams
Ex4: Distributed load: Solution.

Finally, using (9) we subtract the equivalent nodal force matrix based on
the equivalent load replacement given in (28), (29) and (30) from the
effective force matrix given by the results in (34), to obtain the correct
nodal forces and moments as
 9wL   3wL 
   −   12wL 
40 40

 
 
 
  
40

 
 
 
 
 

 
    
F1y 7wL2  wL 2
  
 
 
 
 

− 8wL2
  
 
 
 
 
 

       

 
 
 60   
 60 
 
 


 M1 
 
 
 
  
 
 60 

wL  wL

 
 
 
 
 
 

− −
       

 F2y
 
 
  
  
 
 0 

  2   2   
= 2 
− = (35)
 M2   wL   wL2   0 
− −

 
 
  
 
 
 

30  30

 
 
 
 
 
 
 

        

 F3y 
 
  
 
 
 28wL


11wL  17wL

 
 
 
 
 
 

−
       

M3
 
  
  
 
 40 


 40   
 40 
 
 

3wL2

 
 
 
 
 

2wL2  2
     

  
 wL 
  − 
 −   
15
15 15

May 15, 2023 75 / 98

Analysis of beams
Ex4: Distributed load: Solution.

We used symbol L to represent one-half the length of the beam. If we

replace L with L/2, we obtain the reactions shown in the previous figure,
thus verifying the correctness of our result.

May 15, 2023 76 / 98

Analysis of beams
Beam element with nodal hinge

In some beams an internal hinge may be present. In general, this internal

hinge causes a discontinuity in the slope or rotation of the deflection curve
at the hinge.

1 1 2 2 3
L Hinge L
m1 , θ1 y
m2 = 0, θ2
1 x
1
2
f1y , v1 f2y , v2
Beam with 2 elements and nodal hinge. The hinge considered to be at
right end of element 1.
May 15, 2023 77 / 98
Analysis of beams
Beam element with nodal hinge

To model the hinge, we consider the hinge to be placed on either the

right end of element 1 or on the left end of element 2 but not on
both elements at node 2. Also, the bending moment is zero at the
hinge. We could construct other types of connections that release other
generalized end forces; that is, connections can be designed to make the
shear force or axial force zero at the connection.

These special conditions can be treated by starting with the generalized

unreleased beam stiffness matrix (8) and eliminating the known zero force
or moment. This yields a modified stiffness matrix with the desired force
or moment equal to zero and the corresponding displacement or slope
eliminated.

May 15, 2023 78 / 98

Analysis of beams
Beam element with nodal hinge

For the beam element with a hinge at its right end, the moment m2 is zero
and we partition the [k] matrix (8) to eliminate the degree of freedom θ2
(which is not zero, in general) associated with m2 = 0 as follows:
 . 
12 6L −12 .. 6L

2
.. 2

 6L 4L −6L . 2L 
 
EI  . 
[k] = 3 
L  −12 −6L 12 .. −6L   (36)

 ... .. 
 ... ... . ... 

2
.. 2
6L 2L −6L . 4L

May 15, 2023 79 / 98

Analysis of beams
Beam element with nodal hinge

Partitioning allows us to condense out the degree of freedom θ2 associated

with m2 = 0. That is, equation. (36) is partitioned as shown below:
 .. 
[K11 ] . [K12 ]
 .. 
 3×3 3×1
.
 

[k] =  . . . . . . . . . (37)
 

 [K ] ... [K ]
 

 21 22 
..
1×3 . 1×1

May 15, 2023 80 / 98

Analysis of beams
Beam element with nodal hinge

The condensed stiffness matrix is then found by using the equation

{f } = [k]{u} partitioned as follows:
 .. 

{f1 } 
 [K11 ] . [K12 ]

{u1 }


   .. 



3 × 1 3×3 3 × 1 
. 3×1

 
   

    
... = ... ... ...  ... (38)
 

{f2 } 
  . 
{u2 }

[K21 ] .. [K22 ] 

   
 

   
1×1 1×1
    
..
1×3 . 1×1

where  
 v1 
{u1 } = θ1 {u2 } = {θ2 } (39)
v2
 

May 15, 2023 81 / 98

Analysis of beams
Beam element with nodal hinge

Equations (38) in expanded form are

{f1 } = [K11 ] {u1 } + [K12 ] {u2 }

(40)
{f2 } = [K21 ] {u1 } + [K22 ] {u2 }

Solving for {u2 } in the second of equations (40), we obtain

 
{u2 } = [K22 ]−1 {f2 } − [K21 ] {u1 } (41)

Substituting (41) into the first of equations (40), we obtain

 
{f1 } = [K11 ] − [K12 ] [K22 ]−1 [K21 ] {u1 } + [K12 ] [K22 ]−1 {f2 } (42)

May 15, 2023 82 / 98

Analysis of beams
Beam element with nodal hinge

Combining the second term on the right side of equation (42) with {f1 },
we obtain
{fc } = [Kc ] {u1 } (43)
where the condensed stiffness matrix is

[Kc ] = [K11 ] − [K12 ] [K22 ]−1 [K21 ] (44)

and the condensed force matrix is

{fc } = {f1 } − [K12 ] [K22 ]−1 {f2 } (45)

May 15, 2023 83 / 98

Analysis of beams
Beam element with nodal hinge

Substituting the partitioned parts of [k] from (36) into (44), we obtain the
condensed stiffness matrix as
   
12 6L −12 6L 
EI  2 −6L  − EI

2 1  2

[Kc ] = 6L 4L 2L 6L 2L − 6L
L3 L3   4L2
−12 −6L 12 −6L
 
1 L −1 v1
3EI 
= L L2 −L  θ1 (46)
L3
−1 −L 1 v2
and the element equations (force/displacement equations) with the hinge
at node 2 are
    
 f1y  3EI 1 L −1  v1 
m1 = 3  L L2 −L  θ1 (47)
L
f2y −1 −L 1 v2
   

May 15, 2023 84 / 98

Analysis of beams
Beam element with nodal hinge

The generalized rotation θ2 has been eliminated from the equation and is
not calculated using this scheme. However, θ2 is not zero in general. We
can expand (47) to include θ2 by adding zeros in the fourth row and
column of the [k] matrix to maintain m2 = 0, as follows:
  
  1 L −1 0  v1 
f
 1y  3EI 
L L2 −L 0 
 
θ1
 
m1 = 3    (48)
L −1 −L 1 0  v 
f2y  2 
  
0 0 0 0 θ2


May 15, 2023 85 / 98

Analysis of beams
Beam element with nodal hinge

For the beam element with left node 1 and right node 2 with a hinge at its
left end, the moment m1 is zero, and we partition the [k] matrix (8) to
eliminate the zero moment m1 and its corresponding rotation θ1 to obtain
    
 f1y  3EI 1 −1 L  v1 
f2y = 3  −1 1 −L  v2 (49)
L 2
m2 L −L L θ2
   

The expanded form of (49) including θ1 is

    

 f1y 
 1 0 −1 L   v1 

 3EI 
m1 0 0 0 0  θ
  
1
= 3   (50)
f L  −1 0 1 −L   v
 2y 2

 
  

m2 L 0 −L L2 θ2
  

May 15, 2023 86 / 98

Analysis of beams
Examples: Beams with nodal hinges

Example 1. Determine the displacement and rotation at node 2 and the

element forces for the uniform beam with an internal hinge at node 2. Let
EI be a constant.

P
1 1 2 2 3
a b

May 15, 2023 87 / 98

Analysis of beams
Ex1: Beams with nodal hinges: Solution.

We can consider the hinge as part of element 1. Therefore, with the hinge
located at the right end of element 1, the equation (48) contains the
correct stiffness matrix for element 1. The stiffness matrix of element 1
with L = a is then
 
1 a −1 0 v1
3EI  a a2 −a 0   θ1
[k (1) ] = 3  (51)
a  −1 −a 1 0  v2
0 0 0 0 θ2

As the hinge is considered to be part of element 1, we do not consider it

again as part of element 2.

May 15, 2023 88 / 98

Analysis of beams
Ex1: Beams with nodal hinges: Solution.

So we use the standard beam element stiffness matrix obtained from (8)
for element 2 as
 
12 6b −12 6b v2
EI  6b 4b 2 −6b 2b 2   θ2
[k (2) ] = 3  (52)
b  −12 −6b 12 −6b  v3
6b 2b 2 −6b 4b 2 θ3
Superimposing (51) and (52) and applying the boundary conditions
v1 = 0 θ1 = 0 v3 = 0 θ3 = 0
we obtain the total stiffness matrix and total set of equations as
 3 12 6 
+ 3    
 a 3 b b2  v2 −P
EI  = (53)
6 4 θ2 0

b 2 b
May 15, 2023 89 / 98
Analysis of beams
Ex1: Beams with nodal hinges: Solution.

Solving it, we get

a3 b 3 P a3 b 2 P
v2 = − θ2 = (54)
3(a3 + b 3 )EI 2(a3 + b 3 )EI
(2)
The value θ2 is actually θ2 that associated with element 2. The value of
(1) (2)
θ2 at the right end of element 1 (θ2 ) is, in general, not equal to θ2 .
(1)
Unfortunately, to obtain θ2 we need to consider the hinge on the left end
of the second element. Using (47) for element 1, we obtain the element
forces as
 
    0 

 f1y  3EI 1 a −1 
 

0
 
m1 = 3  a a −a  2 (55)
a a 3 b3 P
f2y −1 −a 1 
  
 

 − 3
 

3(a + b 3 )EI


May 15, 2023 90 / 98

Analysis of beams
Ex1: Beams with nodal hinges: Solution.

which gives us

b3 P ab 3 P b3 P
f1y = m1 = f2y = − (56)
a3 + b 3 a3 + b 3 a3 + b 3
Using equation (52) and the results from (54), we obtain the element 2
forces as
a3 b 3 P
 

 − 

3 + b 3 )EI 
3(a
   
 f2y  12 6b −12 6b 





  2 2
 3
a b P2 
−6b
 EI 
m2 6b 4b 2b
   
= 3  
3 3 (57)
 f3y 
  b −12 −6b 12 −6b    2(a + b )EI  
m3 6b 2b 2 −6b 4b 2  0
   


 


 

0

May 15, 2023 91 / 98

Analysis of beams
Ex1: Beams with nodal hinges: Solution.

Simplifying,

a3 P a3 P a3 bP
f2y = − m2 = 0 f2y = m3 = − (58)
a3 + b 3 a3 + b 3 a3 + b 3

May 15, 2023 92 / 98

Analysis of beams
Examples. Beams with nodal hinges.

Example 2. Determine the slope at node 2 and the deflection and slope
at node 3 for the beam with internal hinge located at node 3, as shown in
the figure below. Nodes 1 and 4 are fixed, and there is a roller support at
node 2. Let E = 210 GPa and I = 2 × 10−4 m4 .

q = 10 kN/m

1 2 3 4
1 2 3
2m 1m 1m

May 15, 2023 93 / 98

Analysis of beams
Ex2: Beams with nodal hinges: Solution.

Use equation (8) to determine element one stiffness matrix as

   
12 12 −12 12 1.5 1.5 −1.5 1.5 v1
EI 
 12 16 −12 8   1.5 2 −1.5 1 
 θ1
[k (1) ] =  = EI 
8  −12 −12 12 −12   −1.5 −1.5 1.5 −1.5  v2
12 8 −12 16 1.5 1 −1.5 2 θ2
(59)
Assume the hinge as part of element 2 and use (48) to obtain the element
2 stiffness matrix as
 
1 1 −1 0 v2
3EI  1 1 −1 0   θ2
[k 2) ] = 3  (60)
1  −1 −1 1 0  v3
0 0 0 0 θ3

May 15, 2023 94 / 98

Analysis of beams
Ex2: Beams with nodal hinges: Solution.

As the hinge is considered to be at the right end of element 2, we do not

consider it to be part of element 3. So we use (8) to obtain the stiffness
matrix as  
12 6 −12 6 v3
3EI  6 4 −6 2 
 θ3
[k 3) ] = 3  (61)
1  −12 −6 12 −6  v4
6 2 −6 4 θ4

May 15, 2023 95 / 98

Analysis of beams
Ex2: Beams with nodal hinges: Solution.

Using the direct stiffness method and the element stiffness matrices in
equations (59) through (61), we assemble the global stiffness matrix as

1.5 −1.5 1.5

 
1.5 0 0 0 0 v1

 1.5 2 −1.5 1 0 0 0 0  θ1


 −1.5 −1.5 4.5 1.5 −3 0 0 0  v2

 1.5 1 1.5 5 −3 0 0 0  θ2
[K ] = EI   (62)

 0 0 −3 −3 15 6 −12 6  v3


 0 0 0 0 6 4 −6 2  θ3

 0 0 0 0 −12 −6 12 −6  v4
0 0 0 0 6 2 −6 4 θ4

May 15, 2023 96 / 98

Analysis of beams
Ex2: Beams with nodal hinges: Solution.

Boundary conditions:

v1 = θ1 = v2 = v4 = θ4 = 0

The reduced stiffness matrix and the equations for solution take the
following form
    
5 −3 0  θ2   0 
EI  −3 15 6  v3 = −5 (63)
0 6 4 θ3 −0.833
   

where
qL 10 × 1 qL2 10 × 12
F3y = − =− = 5 kN M3 = − =− = −0.833 kNm
2 2 12 12

May 15, 2023 97 / 98

Analysis of beams
Ex2: Beams with nodal hinges: Solution.

Substituting numerical values for E and I into (63), and solving

simultaneously, we obtain
(3)
v3 = −2.126×10−5 m θ2 = −1.276×10−5 rad θ3 = θ3 = 2.693×10−5 rad
(64)
(2)
Note θ3 was condensed out of the stiffness matrix for element two.

May 15, 2023 98 / 98