Plastic Analysis

Structural Analysis III
Dr. C. Caprani 1

Plastic Analysis
3rd Year
Structural Engineering

2007/8

Dr. Colin Caprani
Dr. C. Caprani 2
Contents
1. Introduction ......................................................................................................... 3
1.1 Background......................................................................................................3
2. Development......................................................................................................... 4
2.1 Material Behaviour..........................................................................................4
2.2 Cross Section Behaviour .................................................................................6
2.3 Formation of Hinges for Collapse.................................................................14
2.4 Plastic Hinge Development...........................................................................18
2.5 Important Definitions....................................................................................22
2.6 Virtual Work in Plastic Analysis...................................................................24
2.7 Theorems of Plastic Analysis........................................................................29
2.8 Plastic Design................................................................................................39
2.9 Summary of Important Points.......................................................................42
3. Beams.................................................................................................................. 43
3.1 Example 1 Fixed-Fixed Beam with Point Load.........................................43
3.2 Example 2 Propped Cantilever with Two Point Loads..............................46
3.3 Example 3 Propped Cantilever under UDL ...............................................51
3.4 Problems........................................................................................................56
4. Frames ................................................................................................................ 57
4.1 Collapse Mechanisms....................................................................................57
4.2 Example 4 Frame........................................................................................59
4.3 Example 5 Frame, Summer 1997...............................................................65
4.4 Example 6 Frame, Sumer 2000..................................................................69
4.5 Problems........................................................................................................70
5. References .......................................................................................................... 75

Dr. C. Caprani 3
1. Introduction
1.1 Background
Up to now we have concentrated on the elastic analysis of structures. In these
analyses we used superposition often, knowing that for a linearly elastic structure it
was valid. However, an elastic analysis does not give information about the loads that
will actually collapse a structure. An indeterminate structure may sustain loads
greater than the load that first causes a yield to occur at any point in the structure. In
fact, a structure will stand as long as it is able to find redundancies to yield. It is only
when a structure has exhausted all of its redundancies will extra load causes it to fail.
Plastic analysis is the method through which the actual failure load of a structure is
calculated, and as will be seen, this failure load can be significantly greater than the
elastic load capacity.

To summarize this, Prof. Sean de Courcy (UCD) used to say:

a structure only collapses when it has exhausted all means of standing.

Before analysing complete structures, we review material and cross section behaviour
beyond the elastic limit.
Dr. C. Caprani 4
2. Development
2.1 Material Behaviour
A uniaxial tensile stress on a ductile material such as mild steel typically provides the
following graph of stress versus strain:

As can be seen, the material can sustain strains far in excess of the strain at which
yield occurs before failure. This property of the material is called its ductility.

Though complex models do exist to accurately reflect the above real behaviour of the
material, the most common, and simplest, model is the idealised stress-strain curve.
This is the curve for an ideal elastic-plastic material (which doesnt exist), and the
graph is:

Dr. C. Caprani 5

As can be seen, once the yield has been reached it is taken that an indefinite amount
of strain can occur. Since so much post-yield strain is modelled, the actual material
(or cross section) must also be capable of allowing such strains. That is, it must be
sufficiently ductile for the idealised stress-strain curve to be valid.

Next we consider the behaviour of a cross section of an ideal elastic-plastic material
subject to bending. In doing so, we seek the relationship between applied moment
and the rotation of a cross section.

Dr. C. Caprani 6
2.2 Cross Section Behaviour
Moment-Rotation Characteristics of General Cross Section
We consider an arbitrary cross-section with a vertical plane of symmetry, which is
also the plane of loading. We consider the cross section subject to an increasing
bending moment, and assess the stresses at each stage.

Cross-Section and Stresses

Moment-Rotation Curve

Dr. C. Caprani 7
Stage 1 Elastic Behaviour
The applied moment causes stresses over the cross-section that are all less than the
yield stress of the material.

Stage 2 Yield Moment
The applied moment is just sufficient that the yield stress of the material is reached at
the outermost fibres of the cross-section. All other stresses in the cross section are
less than the yield stress. This is limit of applicability of an elastic analysis and of
elastic design. Since all fibres are elastic, the ratio of the depth of the elastic to plastic
regions, 1.0 = .

Stage 3 Elasto-Plastic Bending
The moment applied to the cross section has been increased beyond the yield
moment. Since by the idealised stress-strain curve the material cannot sustain a stress
greater than yield stress, the fibres at the yield stress have progressed inwards
towards the centre of the beam. Thus over the cross section there is an elastic core
and a plastic region. The ratio of the depth of the elastic core to the plastic region is
1.0 0 < < . Since extra moment is being applied and no stress is bigger than the yield
stress, extra rotation of the section occurs: the moment rotation curve losses its
linearity and curves, giving more rotation per unit moment (i.e. looses stiffness).

Stage 4 Plastic Bending
The applied moment to the cross section is such that all fibres in the cross section are
at yield stress. This is termed the Plastic Moment Capacity of the section. Since there
are no fibres at an elastic stress, 0 = . An attempt at increasing the moment at this
point simply results in more rotation, once the cross-section has sufficient ductility.
That is, in steel the cross section classification must be plastic and in concrete the
section must be under-reinforced.

Dr. C. Caprani 8
Stage 5 Strain Hardening
Due to strain hardening of the material, a small amount of extra moment can be
sustained.

The above moment-rotation curve represents the behaviour of a cross section of a
regular elastic-plastic material. However, it is usually further simplified as follows:

With this idealised moment-rotation curve, the cross section linearly sustains moment
up to the plastic moment capacity of the section and then yields in rotation an
indeterminate amount. Again, to use this idealisation, the actual section must be
capable of sustaining large rotations that is it must be ductile.

Plastic Hinge
Note that once the plastic moment capacity is reached, the section can rotate freely
that is, it behaves like a hinge, except with moment of
P
M at the hinge. This is
termed a plastic hinge, and is the basis for plastic analysis. AT the plastic hinge
stresses remain constant, but strains and hence rotations can increase.
Dr. C. Caprani 9
Analysis of Rectangular Cross Section
Since we now know that a cross section can sustain more load than just the yield
moment, we are interested in how much more. In other words we want to find the
yield moment and plastic moment, and we do so for a rectangular section. Taking the
stress diagrams from those of the moment-rotation curve examined previously, we
have:

Elastic Moment
From the diagram:

2
3
Y
M C d =

But, the force (or the volume of the stress block) is:

1
2 2
Y
d
C T b = =

Hence:

Dr. C. Caprani 10

2
1 2
2 2 3
6
Y Y
Y
Y
d
M b d
bd
Z

=

=
=

The term
2
6 bd is thus a property of the cross section called the elastic section
modulus and it is termed Z.

Elasto-Plastic Moment
The moment in the section is made up of plastic and elastic components:

' '
EP E P
M M M = +

The elastic component is the same as previous, but for the reduced depth, d instead
of the overall depth, d:

'
2
2
1 2
2 2 3
6
E Y
Y
d d
M
bd

=

=

The plastic component is:

'
P P
M C s =

The lever arm, s, is:

p
s d h = +
Dr. C. Caprani 11
But

( ) 1
2 2
p
d d d
h

= =

Thus,

( )
2 2
1
2
d d
s d
d
= +
= +

The force is:

( )
1
2
P Y p
Y
C h b
d
b

=
=

Hence,

( ) ( )
( )
'
2
2
1 1
2 2
1
4
P Y
Y
d d
M b
bd

= +

=

And so the total elasto-plastic moment is:

( )
( )
2 2
2 2
2
2
1
6 4
3
6 2
EP Y Y
Y
bd bd
M
bd

= +
=

Dr. C. Caprani 12
Plastic Moment
From the stress diagram:

2
P
d
M C =

And the force is:

2
Y
d
C T b = =

Hence:

2
2 2
4
P Y
Y
Y
bd d
M
bd
S

=

=
=

The term
2
4 bd is a property of the cross section called the plastic section modulus,
termed S.

Dr. C. Caprani 13
Shape Factor
Thus the ratio of elastic to plastic moment capacity is:

P Y
Y Y
M S S
M Z Z

= =

This ration is termed the shape factor, f, and is a property of a cross section alone.
For a rectangular cross-section, we have:

2
2
4
1.5
6
S bd
f
Z bd
= = =

And so a rectangular section can sustain 50% more moment than the yield moment,
before a plastic hinge is formed. Therefore the shape factor is a good measure of the
efficiency of a cross section in bending. Shape factors for some other cross sections
are:

Rectangle: 1.5 f = , as above;

Circle: 1.698 f = ;

Diamond: 2.0 f = ;

Steel I-beam: f is between 1.12 and 1.15.

Dr. C. Caprani 14
2.3 Formation of Hinges for Collapse
We investigate the collapse of a simply supported beam under central point load with
the information we now have.

The bending moment at the centre of the beam is given by:

4
C
PL
M =

Therefore the load at which yield first occurs is:
Dr. C. Caprani 15

4
4
Y
C Y
Y
Y
P L
M M
M
P
L
= =
=

Collapse of this beam occurs when the plastic hinge forms at the centre of the beam,
since the extra hinge turns the statically determinate beam into a mechanism. The
collapse load occurs when the moment at the centre reaches the plastic moment
capacity:

4
4
P
C P
P
P
P L
M M
M
P
L
= =
=

The ratio collapse to yield load is:

4
4
P P P
Y Y Y
P M L M
P M L M
= =

But since,
P
Y
M S
f
M Z
= =

The ratio is just the shape factor of the section.

We are also interested in the plastic hinge, and the zone of elasto-plastic bending. As
can be seen from the diagram, the plastic material zones extend from the centre out to
the point where the moment equals the yield moment.

Dr. C. Caprani 16
Using similar triangles from the bending moment diagram at collapse, we see that:

2
P P Y P EP
p
M M M M M
L l z

= =

In which
EP
M is the elasto-plastic moment at a distance z from the plastic hinge, and
where
2
p
l
z , where
p
l is the total length of the plastic region.

Equating the first two equations gives:

( )
1
1 1
Y
p P Y
P P
L M
l M M L L
M M f

= = =

And so for a beam with a rectangular cross section ( 1.5 f = ) the plastic hinge extends
for a length:

1
1
1.5 3
p
L
l L

= =

Lastly, the shape of the hinge follows from the first and third equation:

( )
2
1
2
1
1
2
P P EP
P EP
P
EP
P
M M M
L z
z
M M
L M
z M
L M
=
=

=

Dr. C. Caprani 17
From our expressions for the elasto-plastic and plastic moments, we have:

( )( )( )
( )
( )
2 2
2
2
2
6 12 3
1
1
2 4
1 2 1
1 3
2 3 2
6
Y
Y
bd
z
L bd
z
L

=

=

=

This shows that the plastic region has a parabolic profile, and confirms that the total
length of the hinge, 2
p
l z = , is 3 L at the location where 1.0 = .

Using a similar form of analysis, we can show that under a UDL the plastic hinge has
a linear profile given by 2 3 z L = and that its length is 3 L .

Dr. C. Caprani 18
2.4 Plastic Hinge Development
Illustrative Example Propped Cantilever
We now assess the behaviour of a simple statically indeterminate structure under
increasing load. We consider a propped cantilever with mid-span point load:

From previous analyses we know that:

3 5
16 32
A C
PL PL
M M = =

We will take the span to be 1 m L = and the cross section to have the following
capacities:

7.5 kNm 9.0 kNm
Y P
M M = =

Further, we want this beam to be safe at a working load of 32 kN, so we start there.

Load of 32 kN
At this value of load the BMD is as shown, with:

( )( ) ( )( )
3 32 1 5 32 1
6kNm 5 kNm
16 32
A C
M M = = = =
Dr. C. Caprani 19
Since the peak moments are less than the yield moments, we know that yield stress
has not been reached at any point in the beam. Also, the maximum moment occurs at
A and so this point will first reach the yield moment.

Load of 40 kN
At this load the BDM becomes that as shown. The moment at A has now reached the
yield moment and so the outer fibres at A are at yield stress.

Load of 48 kN
The BMD is as shown. The moment at A is now 9 kNm the plastic moment
capacity of the section and so the cross section at A has fully yielded. Thus a plastic
hinge has formed at A and so no extra moment can be taken at A, but A can rotate
freely with constant moment of 9 kNm. Also, the moment at C has reached the yield
Dr. C. Caprani 20
moment. Note that the structure does not collapse since there are not sufficient hinges
for it to be a mechanism yet.

Load of 54 kN
Since the moment at A has already reached the plastic moment of the section, no
extra moment can be taken there and
A
M must remain 9 kNm whilst allowing
rotation to freely occur. Therefore, all of the extra moment caused by the increase in
load of 54 48 6 kN = must be taken by the structure as if it were a simply-supported
beam. That is, a beam free to rotate at both ends. The extra moment at C is thus
6 14 1.5 kNm = bring the total moment at C to 9 kNm the plastic moment capacity
of the section. Therefore a plastic hinge forms at C and the structure is not capable of
sustaining anymore load becomes a mechanism and so collapse ensues.

Dr. C. Caprani 21
Discussion
There are several things to note from this analysis:
1. The actual load carried by the beam is 54 kN, greater than the load at which
yield first occurs, 40 kN, the elastic limit. This difference of 35% represents
the extra capacity of the structure over the elastic capacity, so to ignore it
would be inefficient.
2. At the end of the analysis 9 kNm
A C
M M = = and so 1
A C
M M = . Since for an
elastic analysis
( ) ( )
3 16 5 32 1.2
A C
M M PL PL = = , it is evident that our
analysis is not an elastic analysis and so is a plastic analysis.
3. The height of the free bending moment diagram was 4 PL throughout, as
required by equilibrium only the height of the reactant bending moment
diagram varied.
4. At the point of collapse we had 4 reactions and 2 plastic hinges giving a statical
indeterminacy of 3 4 2 3 1 R C = = which is a mechanism and so
collapse occurs.
5. The load can only increase from 48 kN to 54 kN once the cross section at A has
sufficient ductility to allow it rotate thereby allowing the extra load to be taken
at C. If there was not sufficient ductility there may have a brittle-type
catastrophic failure at A resulting in the beam failing by rotating about B before
the full plastic capacity of the structure is realized. Therefore it is only by
having sufficient ductility that a plastic analysis can be used.

Some of these points are general for any plastic analysis and these generalities are
known as the Theorems of Plastic Analysis. However, before looking at these
theorems we need a simpler way of analysing for the collapse of structures: the
incremental loading approach works, but is very laborious.

Dr. C. Caprani 22
2.5 Important Definitions
Load Factor
The load factor for a possible collapse mode i, denoted
i
, is of prime importance in
plastic analysis:

Collapse Load for mode
Working Load
i
i
=

The working load is the load which the structure is expected to carry in the course of
its lifetime.

The collapse load factor,
C
, is the load factor at which the structure will actually fail.
It is therefore the minimum of the load factors for the
m
n different possible collapse
modes:

m
1
min
C i
i n

=

In our previous analysis the working load was 32 kN and the collapse load for the
single mode was found to be 54 kN. Hence:

54
1.6875
32
C
= =

Dr. C. Caprani 23
Factor of Safety
This is defined as:

First yield load
FoS
Working Load
=

The FoS is an elastic analysis measure of the safety of a design. For our example:

40
FoS 1.25
32
= =

Dr. C. Caprani 24
2.6 Virtual Work in Plastic Analysis
Introduction
The easiest way to carry out a plastic analysis is to use virtual work. To do this we
allow the presumed shape at collapse to be the compatible displacement set, and the
external loading and internal bending moments to be the equilibrium set. We can then
equate external and internal virtual work, and solve for the collapse load factor for
that supposed mechanism.

Remember:
Equilibrium set: the internal bending moments at collapse;
Compatible set: the virtual collapsed configuration (see below).

Note that in the actual collapse configuration the members will have elastic
deformation in between the plastic hinges. However, since a virtual displacement
does not have to be real, only compatible, we will choose to ignore the elastic
deformations between plastic hinges, and take the members to be straight between
them.

Dr. C. Caprani 25
Illustrative Example Contd
Actual Collapse Mode
So for our previous beam, we know that we require two hinges for collapse (one
more than its degree of redundancy), and we think that the hinges will occur under
the points of peak moment, A and C. Therefore impose a unit virtual displacement at
C and relate the corresponding virtual rotations of the hinges using S R = , giving:

Notice that the collapse load is the working load times the collapse load factor. So:

( )( ) ( )( ) ( )( )
At At
32 1 2 4
e I
P P
A C
W W
M M

=
= +
14243 14243

( )
32 6
6 9
1.69
32
P
M
=
= =

Since 9 kNm
P
M = and this is as found before.

Dr. C. Caprani 26
Other Collapse Modes
For the collapse mode looked at previously, it seemed obvious that the plastic hinge
in the span should be beneath the load. But why? Using virtual work we can examine
any possible collapse mode. So lets consider the following collapse modes and see
why the plastic hinge should have been beneath the load.

Plastic Hinge between A and C:
Imposing a unit virtual deflection at B, we get the following collapse mode:

And so the virtual work equation becomes:

( )( ) ( ) ( )
( )
At At
32 0.5 1
1 1
2 1
16
1
e I
P P
A D
P
W W
a a
M M
a a
a a
M
a

=

= + +

+
=

1442443 1442443

And since 9 kNm
P
M = :
Dr. C. Caprani 27

1 0.5
9 1
16 1
a
a
a
<
+

=

Eq. (1)

And so we see that the collapse load factor for this mode depends on the position of
the plastic hinge in the span.

Plastic Hinge between C and B:
Again imposing a unit virtual deflection at B we get:

And so the virtual work equation becomes:

( ) ( ) ( )
( )
( )
At At
0.5
32 1
1 1 1
2 1
16
1 1
16 1
e I
P P
A D
P
P
W W
a a a
M M
a a a
a a
a
M
a a
a M a

=

= + +

+

=

= +
1442443 1442443

Dr. C. Caprani 28
Using 9 kNm
P
M = :

0.5 0
9 1
16
a
a
a
<
+

=

Eq. (2)

And again we see that the load factor depends on the position of the hinge.

Summary
Plotting how the collapse load factor changes with the position of the hinge, we get:

1.6875
1
1.5
2
2.5
3
3.5
4
4.5
5
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Distance from Support A (m)
L
o
a
d

F
a
c
t
o
r

Eq 1
Eq 2

This tells us that when the load reaches 1.6875 times the working load (i.e. 54 kN) a
hinge will form underneath the load, at point C, 0.5 m from support A. It also tells us
that it would take more than 54 kN for a hinge to form at any other place, once it
hadnt already formed at C. Thus the actual collapse load factor is the smallest of all
the possible load factors. Hence we can see that in analysing proposed collapse
mechanisms, we are either correct ( 1.6875
C
= ) or we are unsafe (
C
> ). This is
why plastic analysis is an upperbound method.

Dr. C. Caprani 29
2.7 Theorems of Plastic Analysis
Criteria
In Plastic Analysis to identify the correct load factor, there are three criteria of
importance:

1. Equilibrium: the internal bending moments must be in equilibrium with the
external loading.

2. Mechanism: at collapse the structure, or a part of, can deform as a mechanism.

3. Yield: no point in the structure can have a moment greater than the plastic
moment capacity of the section it is applied to.

Based on these criteria, we have the following theorems.

Dr. C. Caprani 30
The Upperbound (Unsafe) Theorem
If a bending moment diagram is found which satisfies the conditions of equilibrium
and mechanism (but not necessarily yield), then the corresponding load factor is
either greater than or equal to the true load factor at collapse.

This is called the unsafe theorem because for an arbitrarily assumed mechanism the
load factor is either exactly right (when the yield criterion is met) or is wrong and is
too large, leading a designer to think that the frame can carry more load than is
actually possible.

Think of it like this: unless its exactly right, its dangerous.

Since a plastic analysis will generally meet the equilibrium and mechanism criteria,
by this theorem a plastic analysis is either right or dangerous. This is why plastic
analyses are not used as often in practice as one might suppose.

The above theorem can be easily seen to apply to the Illustrative Example. When we
varied the position of the hinge we found a collapse load factor that was either correct
( 1.6875
C
= = ) or was too big (
C
> ).

Dr. C. Caprani 31
The Lowerbound (Safe) Theorem
If a bending moment diagram is found which satisfies the conditions of equilibrium
and yield (but not necessarily that of mechanism), then the corresponding load factor
is either less than or equal to the true load factor at collapse.

This is a safe theorem because the load factor will be less than (or at best equal to)
the collapse load factor once equilibrium and yield criteria are met leading the
designer to think that the structure can carry less than or equal to its actual capacity.

Think of it like this: youre either wrong and safe, or youre right and safe.

Since an elastic analysis will always meet equilibrium and yield conditions, an elastic
analysis will always be safe. This is the main reason that it is elastic analysis that is
used, in spite of the extra capacity that plastic analysis offers.
Dr. C. Caprani 32
The Uniqueness Theorem
If a bending moment distribution can be found which satisfies the three conditions of
equilibrium, mechanism, and yield, then the corresponding load factor is the true
load factor at collapse.

So to have identified the correct load factor (and hence collapse mode) for a structure
we need to meet all three of the criteria:
1. Equilibrium;
2. Mechanism;
3. Yield.

The permutations of the three criteria and the three theorems are summarized in the
following table:

Criterion
Upperbound
(Unsafe) Theorem
Lowerbound
(Safe) Theorem
Unique Theorem
Mechanism
Equilibrium
C

Yield
C

Dr. C. Caprani 33
Corollaries of the Theorems
Some other results immediately apparent from the theorems are the following:
1. If the collapse loads are determined for all possible mechanisms, then the actual
collapse load will be the lowest of these (Upperbound Theorem);
2. The collapse load of a structure cannot be decreased by increasing the strength of
any part of it (Lowerbound Theorem);
3. The collapse load of a structure cannot be increased by decreasing the strength of
any part of it (Upperbound Theorem);
4. The collapse load is independent of initial stresses and the order in which the
plastic hinges form (Uniqueness Theorem);

The first point above is the basis for using virtual work in plastic analysis. However,
in doing so, it is essential that the designer considers the actual collapse more. To not
do so would lead to an unsafe design by the Upperbound Theorem.

Note that the Uniqueness Theorem does not claim that the BMD at collapse is unique
only that the collapse load factor is unique. Although rare, it is possible for more
than one BMD to satisfy the Uniqueness Theorem, but they will have the same load
factor.

Dr. C. Caprani 34
Illustrative Example Contd
Plastic Hinge Under the Load
We discovered previously that the collapse load factor was 1.6875 and this occurred
when the hinge was under the point load. Therefore, this collapse mode should meet
all three criteria of the Uniqueness Theorem:

1. Equilibrium: check on the moment at C say:

about 0 54 0.5 9 0 18 kN
B B
M A V V = = =

Thus, from a free-body diagram of CB , 18 0.5 9 kNm
C
M = = as expected.
Thus the equilibrium condition is met.

2. Mechanism: Given the number of hinges it is obvious the structure collapses:

3 4 2 3 1 R C = =

3. Yield: Check that there is no moment greater than 9 kNm
P
M = :

Dr. C. Caprani 35

And so the yield criterion is met.

Since all three conditions are met we are assured that the have the actual collapse
load factor by the Uniqueness Theorem.

Other Collapse Modes
Using the analyses carried out previously for different positions of the plastic hinge,
we can check these collapse modes against the Uniqueness Theorem. For the case of
the hinge between A and C:

Dr. C. Caprani 36
To determine this BMD, we calculate the reaction
B
V by considering the free body
diagram BCD:

( )
about 0 32 0.5 0
16
32
P B
P
B
M D M a V a
M
V
a a
= + =
= +

Thus the moment under the point load is:

8
0.5 16
2
P
C B
M
M V
a a
= = +

Substituting in the expression for from Eq. (1) previously:

8 1
16
2 16 1
P P
C
M M a
M
a a a
+

= +

Which after some algebra becomes:

1
C P
a
M M
a

=

And so because 0.5 1.0 a ,
C P
M M as shown in the BMD. Only when 0.5 a =
does
C P
M M = , which is of course the correct solution.

For the case of the hinge being between C and B, we have:

Dr. C. Caprani 37

Again, we find the reaction
B
V by considering the free body diagram DB:

about 0 0
P
P B B
M
M D M V a V
a
= = =

Thus the moment under the point load at C is:

1
2
C P
M M
a

=

And since 0 0.5 a then 12 1 a and so
C P
M M . Again only when 0.5 a =
does
C P
M M = .

Dr. C. Caprani 38
Summary
We have seen that for any position of the plastic hinge, other than at exactly C, the
yield condition is not met. Therefore, in such cases, the Uniqueness Theorem tells us
that the solution is not the correct one.

Notice that in these examples the mechanism and equilibrium conditions are always
met. Therefore the Upperbound Theorem tells us that our solutions in such cases are
either correct (as in when 0.5 a = ) or are unsafe (as in
C
> ).

In cases where one of the conditions of the Uniqueness Theorem is not met, we
assume a different collapse mode and try again.

Dr. C. Caprani 39
2.8 Plastic Design
When we come to design a structure using plastic methods, it is the load factor that is
known in advance and it is the plastic moment capacity that is the objective. The
general virtual work equations for a proposed collapse mode i is

e I
i j ji Pj j
W W
P M

=
=

In which j is an individual load and deflection or plastic moment and rotation pair. If
we take the
P
M of each member to be some factor, , of a nominal
P
M , then we
have:

i j ji Pj j j
P M =

Since work is a scalar quantity, and since the sum of work done on both sides is
positive, we can see that the load factor and plastic moment capacity have a linear
relationship of slope m for each collapse mode i:

j j
i P
j ji
i i P
M
P
m M

=
=

Thus for each collapse mode, 1
m
k n , we can plot the load factor against the plastic
moment capacity. We do so for two cases:

Dr. C. Caprani 40
1. Load Factor Required Design Plastic Moment Capacity Known:

We can see from this graph that for a particular value of the plastic moment capacity,
*
P
M , collapse mode k gives the lowest load factor and so by the Upperbound
Theorem is the true collapse mode.

2. Design Load Factor Known Plastic Moment Capacity Required:

Dr. C. Caprani 41

From this graph we can see that for a particular value of the load factor, * , collapse
mode k gives the highest design plastic moment capacity,
P
M . However, since by the
Upperbound Theorem we know collapse mode k to be the true collapse mode, it is
therefore the highest value of
P
M from each mode that is required.

Mathematically, using the Upperbound Theorem, the above is summarized as:

[ ]
min
min
min
C i
i P
P i
m M
M m
=
=
=

Hence when the desired
C
is specified:

min
max
max
C
P
i
C
i
C j ji
P
j j
M
m
m
P
M

=

=

=

In summary, if:
Design plastic moment capacity is known design for lowest load factor;
Design load factor is known design for highest plastic moment capacity.

Dr. C. Caprani 42
2.9 Summary of Important Points
Number of Hinges Required for Collapse:
In general we require sufficient hinges to turn the structure into a mechanism, thus:

No. of Plastic
Indet 1
Hinges Required
= +
o

However, this does not apply in cases of local partial collapses.

The Three Theorems of Plastic Analysis:
Criterion
Upperbound
(Unsafe) Theorem
Lowerbound
(Safe) Theorem
Unique Theorem
Mechanism
Equilibrium
C

Yield
C

Collapse Load Factor
By the Unsafe Theorem, which applies when the virtual work method is used:

m
1
min
C i
i n

=

Design Value of Plastic Moment Capacity
The design value of
P
M is the maximum of the design values for
P
M from each
collapse mode:

,
1
max
m
P P i
i n
M M

=

Dr. C. Caprani 43
3. Beams
3.1 Example 1 Fixed-Fixed Beam with Point Load
For the following beam, find the load at collapse, given that 60 kNm
P
M = :

To start the problem, we examine the usual elastic BMD to see where the plastic
hinges are likely to form:

We also need to know how many hinges are required. This structure is 3 statically
indeterminate and so we might expect the number of plastic hinges required to be 4.
However, since one of the indeterminacies is horizontal restraint, removing it would
not change the bending behaviour of the beam. Thus for a bending collapse only 2
indeterminacies apply and so it will only take 3 plastic hinges to cause collapse.

So looking at the elastic BMD, well assume a collapse mode with the 3 plastic
hinges at the peak moment locations: A, B, and C.
Dr. C. Caprani 44
Next, we impose a virtual rotation of to the plastic hinge at A and using the S R =
rule, relate all other displacements to it, and then apply the virtual work equation:

( ) ( ) ( ) ( )
At At At
6 3 3
6 8
8
6
e I
P P P
A C B
P
P
W W
P M M M
P M
P M

=
= + + +
=
=
123 14243 14243

Since 60 kNm
P
M = the load required for collapse is 80 kN P = and so the collapse
BMD for this mode is:

We need to check that this is the correct solution using the Uniqueness Theorem:
Dr. C. Caprani 45
1. Equilibrium:
Well check that the height of the free BMD is 120 kNm as per the collapse BMD:

about 0 80 6 8 0 60 kN
B B
M A V V = = =

Thus, using a free body diagram of CB:

about 0 2 0 120 kNm
C B C
M C M V M = = =

And so the applied load is in equilibrium with the free BMD of the collapse BMD.

2. Mechanism:
From the proposed collapse mode it is apparent that the beam is a mechanism.

3. Yield:
From the collapse BMD it can be seen that nowhere is
P
M exceeded.

Thus the solution meets the three conditions and so, by the Uniqueness Theorem, is
the correct solution.

Dr. C. Caprani 46
3.2 Example 2 Propped Cantilever with Two Point Loads
For the following beam, for a load factor of 2.0, find the required plastic moment
capacity:

Allowing for the load factor, we need to design the beam for the following loads:

Once again we try to picture possible failure modes. Since maximum moments occur
underneath point loads, there are two real possibilities:

Mode 1 Plastic Hinge at C Mode 2 Plastic Hinge at D
Dr. C. Caprani 47
Therefore, we analyse both and apply the Upperbound Theorem to find the design
plastic moment capacity.

Mode 1 Plastic Hinge at C:

( ) ( ) ( )
At
At
150 2 60
2
5
360
2
144 kNm
e I
P P
A
C
P
P
W W
M M
M
M

=

+ = + +

=
=
123
14243

Mode 2 Plastic Hinge at D:

Dr. C. Caprani 48

( ) ( ) ( ) ( )
At At
150 2 60 4 2
540 4
135 kNm
e I
P P
A D
P
P
W W
M M
M
M

=
+ = + +
=
=
123 14243

So by the application of the Upperbound theorem for the design plastic capacity, we
choose 144 kNm
P
M = as the design moment and recognize Mode 1 to be the correct
failure mode. We check this by the Uniqueness Theorem:

1. Equilibrium:
Using the BMD at collapse, well check that the height of the free BMD is that of the
equivalent simply-supported beam. Firstly the collapse BMD from Mode 1 is:

Hence, the total heights of the free BMD are:

96 144 240 kNm
48 132 180 kNm
C
D
M
M
= + =
= + =

Checking these using a simply-supported beam analysis:

Dr. C. Caprani 49

about 0 150 2 60 4 6 0 90 kN
0 150 60 90 0 120 kN
B B
y A A
M A V V
F V V
= + = =
= + = =

Thus, using appropriate free body diagrams of AC and DB:

120 2 240 kNm
90 2 180 kNm
C
D
M
M
= =
= =

And so the applied load is in equilibrium with the free BMD of the collapse BMD.

2. Mechanism:
From the proposed collapse mode it is apparent that the beam is a mechanism. Also,
since it is a propped cantilever and thus one degree indeterminate, we require two
plastic hinges for collapse, and these we have.

3. Yield:
From the collapse BMD it can be seen that nowhere is the design 144 kNm
P
M =
exceeded.
Dr. C. Caprani 50
Thus by the Uniqueness Theorem we have the correct solution.

Lastly, well examine why the Mode 2 collapse is not the correct solution. Since the
virtual work method provides an upperbound, then, by the Uniqueness Theorem, it
must not be the correct solution because it must violate the yield condition.

Using the collapse Mode 2 to determine reactions, we can draw the following BM<D
for collapse Mode 2:

From this it is apparent that Mode 2 is not the unique solution, and so the design
plastic moment capacity must be 144 kNm as implied previously from the
Upperbound Theorem.

Dr. C. Caprani 51
3.3 Example 3 Propped Cantilever under UDL
For the general case of a propped cantilever, find the locations of the plastic hinges at
collapse, and express the load at collapse in terms of the plastic moment capacity.

When considering UDLs, it is not readily apparent where the plastic hinge should be
located in the span. For this case of a propped cantilever we require 2 hinges, one of
which will occur at A, as should be obvious. However, we need to keep the location
of the span hinge variable at say, aL, from A:

Using S R = , we find the rotation at B:

( )
1
B
aL L a =
Dr. C. Caprani 52
And so:

( )
1
B
a
a
=

Thus, noting that the external work done by a UDL is the average distance it moves,
we have:

( ) ( )
At
At
2
2
2
2 1
2
2 1
2
2 1
2 2
1
e I
P P
A
C
P
P
P
W W
aL a
wL M M
a
waL a
M
a
waL a
M
a
M a
waL a

=

= + +

= +

123
144424443

If we introduce a non-dimensional quantity,
2
P
K M wL , wehave:

2 2
1
a
K
a a

Thus the collapse load factor is a function of the position of the hinge, a, as expected.
Also, we can plot the function K against a to visualize where the minimum might
occur:

Dr. C. Caprani 53
(0.586, 11.656)
0
2
4
6
8
10
12
14
16
18
20
0 0.2 0.4 0.6 0.8 1 1.2
Position Along Beam, aL
N
o
n
-
d
i
m
e
n
s
i
o
n
a
l

C
o
l
l
a
p
s
e

L
o
a
d

F
a
c
t
o
r
,

/
K

To determine the critical collapse load factor, suing the Upperbound Theorem, we
look for the minimum load factor using:

0
d
da
=

To do this, well expand the fraction:

2
2 2 4 2
1
a a
K K
a a a a

= =

Using the quotient rule for derivates:

( )( ) ( )( )
( )
2
2
2
2
2 4 2 1 2
0
du dv
v u
dy
dx dx
dx v
a a a a
d
da
a a
=

= =

Dr. C. Caprani 54
Thus multiplying across by
( )
2
2
a a and simplifying gives:

2
2 8 4 0 a a + =

Thus:

( )( )
( )
2
8 8 4 2 4
2 2
2 2
a

=
=

Since we know 0 1 a , then:

2 2 0.586 a = =

At this value for a, the collapse load factor is:

2
2
2 2 0586
0.586 1 0.586
11.656
P
C
P
M
wL
M
wL

=

These values are shown in the graph previously. The collapse BMD is:

Dr. C. Caprani 55
The propped cantilever is a good structure to illustrate the use of the Lowerbound
Theorem. Consider the standard elastic BMD for this structure which meets the
equilibrium condition:

2 2
9
8 128
A max
wL wL
M M = =

If we increase the load by a load factor so that
A P
M M = , and since
max A
M M < we
meet the yield condition, then we have:

2
2 2
8
8 11.656
P
P P
C
wL
M
M M
wL wL

=
= < =

By meeting the equilibrium and yield conditions, but not the mechanism condition,
we have a lowerbound on the critical load factor without doing the virtual work
analysis. This is one of the main reasons elastic analyses are mostly used in practice.
Dr. C. Caprani 56
3.4 Problems
1. For the following prismatic beam of 30 kNm
P
M = , find the load factor at
collapse. (Ans. 1.5)

P

P

Dr. C. Caprani 57
4. Frames
4.1 Collapse Mechanisms
In frames, the basic modes of collapse are:

Beam-type collapse:

Sway Collapse:

Combination Collapse:

Dr. C. Caprani 58
Combination of Mechanisms
One of the most powerful tools in plastic analysis is Combination of Mechanisms.
This allows us to work out the virtual work equations for the beam and sway
collapses separately and then combine them to find the collapse load factor for a
combination collapse mode.

Location of Plastic Hinge at Joints
In frames where members of different capacities meet at joints, it is the weaker
member that develops the plastic hinge. So, for example:

The plastic hinge occurs in the column and not in the beam section since the column
section is weaker.

This is important when calculating the external virtual work done.

Dr. C. Caprani 59
4.2 Example 4 Frame
Find the collapse load in terms of the plastic moment capacity:

Using the idea of Combination of Mechanisms, we will analyse the beam and sway
modes separately, and then combine them in various ways to achieve a solution.

Beam Collapse Mode:

Dr. C. Caprani 60
Notice that, as previously mentioned, we must take the plastic hinge at joint C to be
in the column which has the smaller
P
M . Applying the virtual work equation:

( ) ( ) ( )
{
( )
At
At At At
At At
4 1
3 2 2 2
3 3
6 5
5
6
e I
P P P
G
E F B
E C
P
P
W W
W W W M M M
W M
W M

=

+ + = + +

=
=
123 123 14243
14243 14243

Sway Collapse Mode:

Again notice how careful we are of the hinge location at joint C.

( ) ( )
At At
At
3
9 2
2
9 3.5
7
18
e I
P P
B B
C
P
P
W W
W M M
W M
W M

=

= +

=
=
123 14243
14243

Dr. C. Caprani 61
Combined Collapse Mechanism
To arrive at a solution, we want to try to minimize the collapse load value. Examining
the previous equations, this means that we should try to maximize the external work
done and minimize the internal work done. So:
To maximize the external work done we need to make every load move
through some displacement, unlike the sway mechanism;
To minimize the internal work done we try to remove a hinge, whilst
maintaining a mechanism.

Based on the above try the following:

Instead of using virtual work, we can combine the equations already found:
External virtual work: Since all forces move through displacements:

{ {
Beam Sway
6 9 15
e
W W W W = + =

Internal virtual work: we can add but we must remove the work done by the
hinge at B for both the beam and sway modes:

Dr. C. Caprani 62

{ { {
Beam Hinge - Beam Hinge - Sway
Sway
5 3.5 2 2 4.5
I P P P P P
B B
W M M M M M = + =
14243

Thus we have:

15 4.5
3
10
e I
P
P
W W
W M
W M

=
=
=

Since this is lower than either of the previous mechanisms, we think this is the
solution, and so check against the three conditions of the Uniqueness Theorem.

To prove that the combination of mechanisms works, we do the virtual work analysis:

( ) ( ) ( ) ( )
{
At
At At At
At At
4 3 1
9 3 2 2
3 2 3
15 4.5
3
10
e I
P P
G
B E F
E C
P
P
W W
W W W W M M
W M
W M

=

+ + + = + +

=
=
123 123 123
14243 1442443

Dr. C. Caprani 63
Check for the three conditions, recognizing that 3.33
0.3
P
W
M W = =

1. Equilibrium:
We start by determining the reactions:

about 0 6 0
3.33
0.55
6 6
D P
P
D
M C H M
M W
H W
= =
= = =

0 0.55 0.45
x A
F H W W W = = =

For the whole frame:

about 0 12 3 6 9 6 3 0 0.89
A A A
M D V H W W W W V W = + + = =

Thus the moment at E, from a free-body diagram of ABE, is:

about 0 3 9 0 6.71
A A E E
M E V H M M W = + = =

Since there is a plastic hinge at E of value
( )
2 2 3.33 6.67
P
M W W = = we have
equilibrium.

2. Mechanism:
The frame is obviously a mechanism since 3 4 2 3 1 R C = = .

3. Yield:
To verify yield we draw the collapse BMD from the reactions:

Dr. C. Caprani 64

From the diagram we see that there are no moments greater than 2 6.67
P
M W = in
members AB and BC, and no moments greater than 3.33
P
M W = in member CD.

Dr. C. Caprani 65
4.3 Example 5 Frame, Summer 1997
For the following frame, find the plastic moment capacity required for collapse under
the loads given.

The structure is 1 degree indeterminate so the number of plastic hinges required is 2.
We propose the following collapse mechanism:

Dr. C. Caprani 66
Also, looking closely at the relevant joints:

Thus we have:

( ) ( ) ( )
{
At
At At
At At
3 3
200 3 100 50 2
2 2
650 4.5
144.44 kNm
e I
P P
G
J F
J C
P
P
W W
M M
M
M

=

+ = +

=
=
14243 123
14243 14243

Notice that the 50 kN point load at G does negative external work since it moves
against its direction of action.

Note also that there are other mechanisms that could be tried, some of which are
unreasonable.

Next we check this solution to see if it is unique:

Dr. C. Caprani 67
1. Equilibrium:
For the whole frame, taking moments about D gives:

50 1 200 6 100 8 9 0 227.8 kN
A A
V V + + = =

Using a free body diagram of ABJ, and taking moments about the plastic hinge at J:

2 144.4 100 2 3 227.8 3 0 64.9 kN
A A
H H + = =

So for the whole frame:

0 0 64.9 kN
x A D D
F H H H = = =

Thus for the free body diagram of CD, taking moments about C:

50 1 3 0 144.7 kNm
C D C
M H M = =

Dr. C. Caprani 68
Since this is the value of
P
M we have a plastic hinge at C as expected. Thus the loads
are in equilibrium with the collapse mode.

2. Mechanism:
Since 3 4 2 3 1 R C = = we have a mechanism.

3. Yield:
Drawing the bending moment diagram at collapse shows that no section has a
moment greater than its moment capacity of either
P
M or 2
P
M :

Dr. C. Caprani 69
4.4 Example 6 Frame, Sumer 2000
For the following frame, find the collapse load factor when 120 kNm
P
M = :

To be done in class.

Dr. C. Caprani 70
4.5 Problems
1. (Summer 1999) The following rigid-jointed frame is loaded with working
loads as shown:
(a) Find the value of the collapse load factor when 120 kNm
P
M = ;
(b) Show that your solution is the unique solution;
(c) Sketch the bending moment diagram at collapse, showing all important
values.
(Ans. 2.0 = )

Dr. C. Caprani 71
loads as shown:
P
M = ;
values.
(Ans. 1.89 = )

Dr. C. Caprani 72
loads as shown:
P
M = ;
values.
(Ans. 2.13 = )

Dr. C. Caprani 73
loads as shown:
P
M = ;
values.
(Ans. 1.33 = )

Dr. C. Caprani 74
5. (Summer 2007) The following rigid-jointed frame is loaded so that the force
system shown is just sufficient to cause collapse in the main frame ABCD:
(a) Find the value of
P
M given that the relative plastic moment capacities
are as shown in the figure;
values.
(Ans. 175.8 kNm
P
M = )

3 m
3

m
A
B C
D
E
F
G
4 m 4 m
1

m
50 kN
200 kN
1 m
M
p
2M
p
M
p

Dr. C. Caprani 75
5. References
Baker, J .F., Horne, M.R. and Heyman, J ., The Steel Skeleton, Volume II, Plastic
Behaviour and Design, Cambridge University Press, 1956.
Heyman, J ., Beams and Framed Structures, 2nd Edn., Pergamon Press, 1974.
Heyman, J ., Elements of the Theory of Structures, Cambridge University Press,
1996.
Hodge, P.G., Plastic Analysis of Structures, McGraw-Hill, New York, 1959.
McKenzie, W.M.C., Examples in Structural Analysis, Taylor and Francis,
Abington, 2006.
Neal, B.G., Structural Theorems and their Applications, Pergamon Press, 1964.
Thompson, F., and Haywood, G.G., Structural Analysis Using Virtual Work,
Chapman and Hall, 1986.
Rees, D.W.A., Mechanics of Solids and Structures, Imperial College Press,
London, 2000.

Plastic Analysis

Uploaded by

Copyright:

Available Formats

Plastic Analysis

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Plastic Analysis

Uploaded by

Copyright:

Available Formats

Structural Analysis III

You might also like