1c. CVE 503 HO CHP 01
1c. CVE 503 HO CHP 01
1.2. DEVELOPMENT
A uniaxial tensile stress on a ductile material such as mild steel typically provides the following
graph of stress versus strain (Fig.1.1):
Fig. 1.1: Stress strain relationship Fig. 1.2: Idealized Stress-Strain Curve
As can be seen from Fig. 1.1, 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 doesn’t exist), and the graph is presented in Fig. 1.2:
As can be seen from Fig. 1.2, 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 idealized 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.
1
1.2.2 Cross Section Behaviour
From Fig. 1.3, 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.
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 Mp 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.
3
Fig. 1.5: Simplified Moment-Rotation Curve
2
The term 𝑏𝑑 ⁄6 is thus a property of the cross section called the elastic section modulus and it
is termed Z.
Shape Factor
Thus the ratio of elastic to plastic moment capacity is:
𝑀𝑃 𝜎𝑦 ∗ 𝑆 𝑆
= =
𝑀𝑌 𝜎𝑦 ∗ 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
𝑆 𝑏𝑑 ⁄4
𝑓 = = 2 = 1.5
𝑍 𝑏𝑑 ⁄
6
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:
Circle: f =1.698;
Diamond: f = 2.0;
5
Steel I-beam: f is between 1.12 and 1.15.
6
Fig. 1.7: Propped cantilever
From previous analyses we know that:
3𝑃𝐿 5𝑃𝐿
𝑀𝐴 = 𝑀𝐶 =
16 32
We will take the span to be L =1 m and the cross section to have the following capacities:
𝑀𝑌 = 7.5 𝑘𝑁𝑚 𝑀𝑃 = 9.0 𝑘𝑁𝑚
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
𝑀𝐴 = = 6.0 𝑘𝑁𝑚 𝑀𝐶 = = 5.0 𝑘𝑁𝑚
16 32
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 (Fig. 1.9). 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 (Fig. 1.10). 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
7
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 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 𝑀𝐴 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 ∗ 1⁄4 = 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.
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.
𝑀𝐴
2. At the end of the analysis MA = MC = 9 kNm and so ⁄𝑀 = 1. Since for an elastic
𝐶
𝑀𝐴 (3𝑃𝐿/16)
analysis ⁄𝑀 = ⁄(5𝑃𝐿/32) = 1.2, it is evident that our analysis is not an elastic
𝐶
8
4. At the point of collapse we had 4 reactions and 2 plastic hinges giving a statically
indeterminacy of R −C − 3 = 4 − 2 − 3 = −1 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.
1.2.5 Important Definitions
Load Factor
The load factor for a possible collapse mode i, denoted 𝜆𝑖 , is of prime importance in plastic
analysis:
𝐶𝑜𝑙𝑙𝑎𝑝𝑠𝑒 𝑙𝑜𝑎𝑑 𝑓𝑜𝑟 𝑚𝑜𝑑𝑒 𝑖
𝜆𝑖 =
𝑊𝑜𝑟𝑘𝑖𝑛𝑔 𝑙𝑜𝑎𝑑
The working load is the load which the structure is expected to carry in the course of its lifetime.
The collapse load factor, 𝜆𝐶 , is the load factor at which the structure will actually fail. It is
therefore the minimum of the load factors for the 𝑛𝑚 different possible collapse modes:
𝜆𝐶 = min 𝜆𝑖 1 < 𝑖 < 𝑛𝑚
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
𝜆𝐶 = = 1.6875
32
Factor of Safety
This is defined as:
𝐹𝑖𝑟𝑠𝑡 𝑦𝑖𝑒𝑙𝑑 𝑙𝑜𝑎𝑑
𝐹𝑆 =
𝑊𝑜𝑟𝑘𝑖𝑛𝑔 𝑙𝑜𝑎𝑑
The FS is an elastic analysis measure of the safety of a design. For our example:
40
𝐹𝑆 = = 1.25
32
9
1.2.6 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.
10
Notice that the collapse load is the working load times the collapse load factor. So:
𝛿𝑊𝑒 = 𝛿𝑊𝑖
(32𝜆)(1) = (𝑀𝑃 𝑎𝑡 𝐴 )(2) + (𝑀𝑃 𝑎𝑡 𝐶 )(4)
32𝜆 = 6𝑀𝑃
6(9)
𝜆= = 1.69
32
Since 𝑀𝑃 = 9 𝑘𝑁𝑚 and this is as found before.
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.
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 it’s exactly right, it’s dangerous.
11
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 (𝜆 = 𝜆𝐶 = 1.6875)
or was too big(𝜆 > 𝜆𝐶 ).
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: you’re either wrong and safe, or you’re 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.
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.
12
The permutations of the three criteria and the three theorems are summarized in the following
table:
Table 1.1: Theorems of Plastic Analysis
Criterion Upperbound Lowerbound Unique Theorem
(Unsafe) Theorem (Safe) Theorem
Mechanism 𝜆 ≥ 𝜆𝐶
Equilibrium 𝜆 ≤ 𝜆𝐶 𝜆 = 𝜆𝐶
Yield
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.
13
1.3 BEAMS
Example 1 – Fixed-Fixed Beam with Point Load
For the following beam:
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, we’ll assume a collapse mode with the 3 plastic hinges at the
peak moment locations: A, B, and C.
14
Next, we impose a virtual rotation of θ to the plastic hinge at A (Fig. 1.15) and using the S =
Rθrule, relate all other displacements to it, and then apply the virtual work equation:
δWe = δWi
P(60) = MP (θA ) + MP (θC + 3θC ) + MP (3θB )
6Pθ = 8MP θ
8MP 8x60
P= = = 80 kN
6 6
ii. We need to check that this is the correct solution using the Uniqueness Theorem.
a. Equilibrium:
We’ll check that the height of the free BMD is 120 kNm as per the collapse BMD (Fig 1.17):
∑𝑀𝐴 = 0 ∴ 80 ∗ 6 − 8𝑉𝐵 = 0 ∴ 𝑉𝐵 = 60 𝑘𝑁
Thus, using a free body diagram of CB:
∑𝑀𝐶 = 0 ∴ 𝑀𝐶 − 2𝑉𝐵 = 0 ∴ 𝑀𝐶 = 120 𝑘𝑁𝑚
And so the applied load is in equilibrium with the free BMD of the collapse BMD.
b. Mechanism:
From the proposed collapse mode it is apparent that the beam is a mechanism.
c. Yield:
From the collapse BMD it can be seen that nowhere is Mp exceeded.
Thus the solution meets the three conditions and so, by the Uniqueness Theorem, is the correct
solution.
15
Example 2 – Propped Cantilever with Two Point Loads
i. For the following beam (Fig. 1.18), for a load factor of 2.0, find the required plastic moment
capacity:
ii. Check if the collapse load satisfy the Uniqueness Theorem.
Fig. 1.20: Mode 1 – Plastic Hinge at C Fig. 1.21: Mode 2– Plastic Hinge at D
Therefore, we analyse both and apply the Upperbound Theorem to find the design plastic
moment capacity.
Mode 1 – Plastic Hinge at C (Fig. 1.22):
Fig. 1.22: Plastic Hinge for Model I Fig. 1.23: Plastic Hinge for Model II
δWe = δWi
𝜃𝐶
150(2𝜃) + 60(𝜃) = MP (θA ) + MP (θC + )
2
5
360θ = M θ; 𝑀𝑃 = 144 kNm
2 P
16
Mode 2 – Plastic Hinge at D (Fig. 1.23):
δWe = δWi
150(2𝜃) + 60(4𝜃) = MP (θA ) + MP (θ𝐷 + 2𝜃𝐷 )
540θ = 4MP θ
𝑀𝑃 = 135 kNm
So by the application of the Upperbound theorem for the design plastic capacity, we choose
𝑀𝑃 = 144 𝑘𝑁𝑚 as the design moment and recognize Mode 1 to be the correct failure mode.
ii. We need to check that this is the correct solution using the Uniqueness Theorem.
1. Equilibrium:
Using the BMD at collapse, we’ll check that the height of the free BMD is that of the equivalent
simply-supported beam. Firstly the collapse BMD from Mode 1 is (Fig. 1.24):
Fig. 1.26
∑𝑀𝐴 = 0 ∴ 150 ∗ 2 + 60 ∗ 4 − 6𝑉𝐵 = 0 ∴ 𝑉𝐵 = 90 𝑘𝑁
∑𝐹𝑌 = 0 ∴ 150 + 60 − 90 − 𝑉𝐴 = 0 ∴ 𝑉𝐴 = 120 𝑘𝑁
17
Thus, using appropriate free body diagrams of AC and DB:
𝑀𝐶 = 120 ∗ 2 = 240 𝑘𝑁𝑚 𝑀𝐷 = 90 ∗ 2 = 180𝑘𝑁𝑚
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 MP = 144 kNm exceeded.
Lastly, we’ll 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 BMD for collapse
Mode 2 (Fig. 1.27):
From this it is apparent that Mode 2 is not the unique solution (since 𝑀𝐶 = 150 𝑘𝑁𝑚 > 𝑀𝑃 =
135 𝑘𝑁𝑚), and so the design plastic moment capacity must be 144 kNm as implied previously
from the Upperbound Theorem.
18
1.4. FRAMES
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 joint shown in Fig. 1.29, 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.
19
Example 4 – Frame
Find the collapse load in terms of the plastic moment capacity for the frame shown in Fig. 1.30.
Fig. 1.30
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.
Fig. 1.31
Notice that, as previously mentioned, we must take the plastic hinge at joint C to be in the
column which has the smaller 𝑀𝑃 . Applying the virtual work equation:
𝛿𝑊𝑒 = 𝛿𝑊𝑖
4 1
𝑊(3𝜃𝐸 ) + 𝑊(2𝜃𝐹 ) + 𝑊(𝜃𝐺 ) = 2𝑀𝑃 (𝜃𝐵 ) + 2𝑀𝑃 ( 𝜃𝐸 ) + 𝑀𝑃 ( 𝜃𝐶 )
3 3
6𝑊𝜃 = 5𝑀𝑃 𝜃
5
𝑊 = 𝑀𝑃
6
20
Sway Collapse Mode:
Fig. 1.32
Again notice how careful we are of the hinge location at joint C.
𝛿𝑊𝑒 = 𝛿𝑤𝑖
3
𝑊(9𝜃𝐵 ) = 2𝑀𝑃 (𝜃𝐵 ) + 𝑀𝑃 ( 𝜃𝐶 )
2
9𝑊𝜃 = 3.5𝑀𝑃 𝜃
7
𝑊= 𝑀
18 𝑃
Based on the above, the combined collapse mechanism is shown in Fig. 1.33 and try the
following:
Fig. 1.33:
21
Instead of using virtual work, we can combine the equations already found:
• External virtual work: Since all forces move through displacements:
𝛿𝑊𝑒 = 6𝑊𝜃 + 9𝑊𝜃 = 15𝑊𝜃
• 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:
𝛿𝑊𝑖 = 5𝑀𝑃 𝜃 + 3.5𝑀𝑃 𝜃 − 2𝑀𝑃 𝜃𝐻𝑖𝑛𝑔𝑒 𝐵−𝐵 − 2𝑀𝑃 𝜃𝐻𝑖𝑛𝑔𝑒 𝐵−𝑆 = 4.5𝑀𝑃
Thus we have:
𝛿𝑊𝑒 = 𝛿𝑊𝑖
15𝑊𝜃 = 4.5𝑀𝑃 𝜃
3
𝑊= 𝑀
10 𝑃
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 (Fig.
1.34):
Fig. 1.34:
𝛿𝑊𝑒 = 𝛿𝑊𝑖
4 3 1
𝑊(9𝜃𝐵 ) + 𝑊(3𝜃𝐸 ) + 𝑊(2𝜃𝐹 ) + 𝑊(𝜃𝐺 ) = 2𝑀𝑃 ( 𝜃𝐸 ) + 𝑀𝑃 ( 𝜃𝐶 + 𝜃𝐶 )
3 2 3
15𝑊𝜃 = 4.5𝑀𝑃 𝜃
3
𝑊= 𝑀
10 𝑃
Check for the three conditions, recognizing that 𝑀𝑃 = 3.33𝑊
1. Equilibrium:
We start by determining the reactions:
22
Fig. 1.35: Fig. 1.36:
𝑀𝑃 3.33𝑊
∴ 𝐻𝐷 = = = 0.55𝑊
6 6
Since there is a plastic hinge at E of value 2𝑀𝑃 = 2 ∗ (3.33𝑊) = 6.67𝑊 we have equilibrium.
2. Mechanism:
3. Yield:
To verify yield we draw the collapse BMD from the reactions:
Fig. 1.37:
From the diagram we see that there are no moments greater than 2𝑀𝑝 = 6.67𝑊 in members
AB and BC, and no moments greater than 𝑀𝑝 = 3.33𝑊 in member CD.
23