1B Structural Mechanics 2001/2 C—O Handout 4 Plastic Theory S.D.G., January 31, 2002126 1B Structural Mechanics 2001/2 4.1 Background and Introduction 4.1.1 History In the 1930's, experiments showed that elastic analysis was not suitable for estimating the col- lapse load of structures. A new method was developed that instead concentrated on the final failure mode — this method is known as plastic theory. Much of the work on plasticity theory was carried out in this Department by a team led by Lord Baker, and some of the Department's buildings are very early examples of the application of plasticity theory to design. Research on using plastic theory continues in the Department to the present day. 4.1.2 Material Model Plasticity theory was originally developed for the design of steel structures. Although it is rou- tinely applied to the analysis of structures made from other materials, we will initially consider a simple material model that is suitable for steel. Ina tension test, the stress strain curve for mild stee! is approximately: In simple plasticity theory, this is idealised as either rigid-perfectly plastic, or elastic-perfectly plasticHandout 4. Plastic Theory 127 rigid-perfectly plastic elastic-perfectly plastic The main assumptions are: 1. Neglect work-hardening; 2. The material is ductile; 3. The behaviour is identical in tension and compression. The rigid-perfectly plastic model is often used for simplicity. It neglects elastic deforma on, 4.1.3 Example: failure of a simply-supported beam ‘An experiment shows that as the load is increased, collapse occurs by the formation of a kink in the beam, known as a plastic hinge. Initially the behaviour is elastic: Eventually further curvature becomes concentrated under the load, at the plastic hinge:128 1B Structural Mechanics 2001/2 w To understand the collapse, we need to examine what happens on the cross section where the plastic hinge forms. 4.1.4 Stresses on a rectangular cross-section as a plastic hinge forms We make two assumptions about the behaviour of the beam: (1) Plane sections remain plane (the ‘railway track’ analogy). Even when the section becomes plastic this remains a good assumption. It implies that the strains on the cross-section must vary linearly with curvature. If y is the distance from the neutral axis, and x is the local curvature, the strain is given by The stresses are related to the strains by eSey For a simple rectangular section, this gives the behaviour shown below:Handout 4, Plastic Theory 129 cross-section strain stress comments (constant) aba & strain Ee stress (b) 2 z (c) @, . 4 We can calculate a moment-curvature relationship for each of the regimes (a)~(d): (a) Linear elastic range. From elastic bending theory (1A) where J = second moment of area of section 8 for a rectangular section1B Structural Mechanics 2001/2 where ymax = distance from the neutral axis to the extreme fibre d ‘ =$ fora rectangular section (b) At first yield (= M,) This is often written as My = Z,Gy, where Z, is the elastic section modulus, For a rectangular section, The curvature at this point is “EL Ey = 2 fora rectangular section (d) Fully plastic (M = M,) We wish to calculate the moment on the section when it becomes fully plastic (as the curvature becomes very large)Handout 4, Plastic Theory 131 cross-section strain stress equivalent (constant) forces rb g, The fully plastic moment My is in equilibrium with the resultant forces Phis is often written as My = Z,G,, where Zp is the plastic section modulus. For a rectangular section, (c) Partially plastic It can be shown that for a rectangular section in the partially plastic range 1-38) M=M,132 1B Structural Mechanics 2001/2 Moment-Curvature plot ‘The results (a)-(d) can be plotted to show the formation of a plastic hinge in a rectangular cross- section: Moment (M4) Curvature (K) If we had instead assumed that the material was rigid-perfectly plastic, there would have been no deforation at all until the entire section was plastic, giving a moment-curvature plot: Moment (M) My Curvature (x) 4.1.5 Example: failure of a simply-supported beam revisited Now we know the moment-curvature relationship in the hinge, we can examine the collapse of the beam more carefully:Handout 4, Plastic Theory 133, Elastic- Rigid- perfectly perfectly plastic plastic >=——s =F? ~~ s If collapse occurs when W = Weaitapses Bending- Load moment diagram BM.134 1B Structural Mechanics 2001/2 4.2 Plastic section modulus This section will show how the calculation carried out to find the plastic moment of a rectangular beam can be generalised to almost any cross-section. The only assumption that we shall make is that there is an axis of symmetry of the cross-section in the direction of the applied loads. We shall again aim to write the plastic moment as M, = Zp), where Zp is the plastic section modulus, which only depends on the cross-section. cross-section stresses forces 5, iy a stress The total tensile force is given by The total compressive force is given by In the absence of any overall axial force (the usual assumption), these forces must balance ‘Thus the neutral axis for fully plastic bending is the equal-area axis. ‘The plastic moment is given by My yoydA + —oydA tensile [compressive =| plada and hence the plastic section modulus is given by Z, Tocnl?l#4Handout 4. Plastic Theory 135 In practice, sections can usually be split into a number of simple regions, and the formula for plastic section modulus reduces to Zy= Adel regions where: A, is the area of a region yielded fully in tension, or fully in compression; |jy| is the distance from the equal area axis to the centroid of A, For our example 4.2.1 Example Find Z, and My for the following section, given 6, = 250N/mm* 50 a 7 = fa All dimensions in mm alfa s | ss I Locate position of equal area axis X-X Divide the tension zone into simple shapes and identify their centroids w.rt. X-X axis. Repeat for the compression zone136 1B Structural Mechanics 2001/2 Evaluate Zp = ZregionsArltrls Mp = Zp 4.2.2. Summary of plastic section modulus 1. 2 43 In the limit, the whole section is yielded either in tension of compression. . If there is no axial force, longitudinal equilibrium forces the neutral axis to be at the equal area axis, Note that this may not be the centroid of the section — when a beam yields in bending, the neutral axis may move. The plastic section modulus is given by Zp = freaion|)|4A. Note that this is not the first moment of area, as all of the beam makes a positive contribution. In practice, Z, is more easily calculated by splitting the beam into a number of simple regions which are fully yielded er tension or compression, and applying Zp = Sregions Ar|Vr|- Try Question 4, Examples Sheet 2/4 Plastic collapse for determinate beams An alternative way of calculating the collapse load of a beam is to consider an energy balance of the beam as it fails. The key idea is that during plastic collapse, the energy dissipated as the beam yields must balance the work done by the load. 4.3.1 Example 1 Consider our earlier simple example: fHandout 4. Plastic Theory 137 Neglecting elastic deformation (or assuming rigid-perfectly plastic material), during collapse the beam will look like: Compatibility Work done by external load Energy dissipated by plastic hinge Calculate plastic collapse load — equate work done and energy dissipated: Note that we have calculate the same collapse load as before without using any statement of equilibrium. 4.3.2 Example 2 Find the collapse load of the simply supported beam shown below. The plastic moment of the beam is Mp.138 1B Structural Mechanics 2001/2 _ if f. For this example, we might consider that two failure mechanisms were possible, either a pla tic hinge would form at B, or it would form at C. We will examine the effect of these assumptions on the estimate of collapse load. Mechanism (1): plastic hinge at C Collapse mechanism: Compatibility Work done by external load Energy dissipated by plastic hinge Calculate plastic collapse loadHandout 4. Plastic Theory 139 Mechanism (2): plastic hinge at B Collapse mechanism: Compatibility Work done by external load Energy dissipated by plastic hinge Calculate plastic collapse load140 1B Structural Mechanics 2001/2 Choosing correct collapse mechanism As W increases from zero, the beam will collapse when W reaches W;, and the failure mechanism will be mechanism (1). The load will never reach W, and failure mechanism (2) will never occur. 4.3.3, Example 3 — Distributed loading In the previous examples it was clear that hinges would only form where a point load was applied. With distributed loading, it is not clear where the hinge will form. Consider the following beam, with plastic moment M, A G mnt length Initial estimate For simplicity, initially assume that the hinge forms in the centre. Collapse mechanism: Compatibility Work done by external load Energy dissipated by plastic hinge Calculate plastic collapse loadHandout 4. Plastic Theory 141 More careful analysis Consider a general collapse mechanism, where the hinge forms at some unknown position x. B x QI) Compatibility Work done by external load in HB in CH in total142 IB Structural Mechanics 2001/2 Energy dissipated by plastic hinge Calculate plastic collapse load Choose correct collapse mechanism — to find the actual collapse load, and the actual posi- tion of the hinge, we require the lowest possible value of wyyech. We can find the minimum by considering when dWmech/dx = 0. In this case, however, the expression for Wiech has all the terms involving x on the bottom, and hence it is easier to find the position of the maximum of 1/Wzcis (the same thing), by considering when d(1/wmecn) /dx = 0. Substitute back to give the critical load we Note the relatively small improvement in the estimated collapse load compared with the initial estimate, w1 = 36M)/9L?. oe Shoot Al Try Questions 5 and 6, Examples Sheet 2/4)Handout 4, Plastic Theory 143 4.4 Upper Bound Theorem of Plasticity An estimate of the plastic collapse load, Wmech, calculated for any| arbitrary compatible mechanism by equating the work done by the applied load, and the plastic energy dissipated, will either be greater: than, or equal to, the actual collapse load W. Winecn 2 We General methodology 1. Postulate compatible collapse mechanisms; 2. Evaluate Wmech for each; 3. Select the collapse mechanism that gives the lowest upper bound on W- Note that we have only used two of the three basic principles of structural analysis. We have to postulate a compatible collapse mechanism, and we use a material law to find the energy dissipated, but we have not considered equilibrium. Later, we shall also see the Lower Bound theorem of Plasticity, that considers only equilib- rium and a material law, but neglects compatibility. A formal proof of the bound theorems of plasticity will be given at the end of the lectures on plastic theory, 4.5 Plastic Collapse of Statically Indeterminate Beams ‘We saw in Handout 1 that an elastic analysis of a statically indeterminate structure is harder than the analysis of a determinate structure: it was necessary to simultaneously solve equations of equilibrium and compatibility. For a plastic collapse analysis, however, it is no more difficult to analyse an indeterminate structure than a determinate one. The only distinction is that failure must occur at more than one point. A statically determinate beam can collapse when one hinge forms. A statically indeterminate beam will require more than one hinge. 4.5.1 Example — Propped Cantilever |"144 1B Structural Mechanics 2001/2 For this example, two hinges are required for collapse to occur. It is evident that the only sensible places that they may occur are under the load or at the root, and hence we can be confident that our upper bound analysis in this case will give the correct collapse load. Collapse mechanism |" k—12 12 41 Compatibility Work done by external load Energy dissipated by plastic hinge Calculate plastic collapse load 4.5.2, Example — Multi-span Beam Find the collapse load W, of the following structure:Handout 4. Plastic Theory 145 Collapse mechanism 1 Collapse mechanism 2 These are the only two sensible mechanisms — for uniform structures subjected to point loads, hinges will only form at the load or support points. For instance, the third mechanism shown below has the same work done by the load, but more energy dissipated. Collapse mechanism 3146 1B Structural Mechanics 2001/2 Ws Ws BNL ER Choose correct collapse mechanism By the upper-bound theorem, the actual collapse load will be the lowest collapse load for any collapse mechanism. Because there are no other sensible mechanisms, W> must be the actual collapse load. 8M, We = Wp Handout 4. Plastic Theory 153 202 | 302 _ Sway mechanism Combined mechanism Sway mechanism is most critical, and the load factor at collapse is 1.5.154 1B Structural Mechanics 2001/2 Example - pitched roof portal frame Draw an interaction diagram for the following structure, which has uniform plastic moment. 8 9 = ‘The same mechanisms are possible for this structure as the normal portal frame, but now the beam mechanism has to push out one of the legs: J a A — a \’ ‘HH Hy Beam mechanism 1 coal — > J #2 Sway mechanism Combined mechanism Beam mechanism 2Handout 4. Plastic Theory Beam mechanism 1 Beam mechanism 2 155156 1B Structural Mechanics 2001/2 ‘Sway mechanism Combined mechanismHandout 4. Plastic Theory 157 Interaction Diagram SE | Try Questions 9 and 10, Examples Sheet 2/4 Le 4.7 Yield Line analysis of slabs and plates 4.7.1 Introduction So far, plasticity theory has been applied to ID elements such as beams and frames. In this sec- tion, we will extend the same ideas to 2D plate structures. An elastic analysis of these structures is very difficult, but a plastic analysis just follows the principles we have seen before. The main difficulty is to ensure that we use a compatible collapse mechanism.158 1B Structural Mechanics 2001/2 Simple extension of beam example to plate Notation A Cc B Edges Yield lines Free edge Simply-supported edge Clamped edge Sagging Hogging A c B Axes of rotationHandout 4. Plastic Theory 159 Calculation of the plastic moment of a yield line Rectangular beam Section of plate For a plate made from e.g. steel, the moment capacity per unit length will therefore be m=So, Many plates are in fact made of reinforced concrete, and are then referred to as slabs. Con- crete slabs do fail by yield lines, but the energy absorbed is all in the steel reinforcement, and the above calculation for moment capacity is not valid, Calculating the plastic moment for re- inforced concrete beams and slabs will be covered later in the course. However, one important point is that, depending on the layout of the steel, the plastic moment capacity m can be different in sagging or hogging, and can also be different in different directions. The 1B course will always assume that the value of m is isotropic, apart from one examples paper question, Assumptions of yield line theory 1. Collapse is due to ductile flexural failure, and is not due to other modes such as shear failure, 2. In-plane forces are ignored. 3. The deformations are small compared to the overall dimensions of the structure.160 1B Structural Mechanics 2001/2 4.7.2. Compatible Yield Line Patterns An essential requirement of an upper-bound collapse analysis is that the failure mechanism is compatible. A few simple guidelines help to ensure that a chosen yield line pattern is compatible. 1. The yield lines divide the slab into several rigid regions, which must remain planar. Each region must have a unique axis about which it rotates. 2. The yield lines must be straight between intersections. 3. A yield line between two rigid regions must pass through the intersection of the axis of rotation of those regions If these rules are broken, the chosen collapse mechanism will be incompatible, for instance it may require some of the rigid regions to twist. A good way of visualising a collapse mechanism is to imagine it projected into 3D. Could you make a simple cardboard model of the collapsed structure? Example collapse mechanisms for uniformly distributed loading x, yete, are variable parametersHandout 4, Plastic Theory 161 e | column Example collapse mechanisms for point loads K 4 single point load skew bridge with axle loading (2 point loads) 4.7.3. Work equation An upper bound for the collapse load of a slab can be calculated in the same way as for beam structures, by assuming a compatible collapse mechanism, and equating the work done and the energy dissipated during collapse.162 1B Structural Mechanics 2001/2 Energy dissipated: For a hinge between two rigid regions, i and j, the total plastic moment in the hinge will be Mp = mij where: m is the plastic moment per unit length of the slab; hy is the length of the hinge. ‘The energy dissipated in this hinge will therefore be: E.D. in one hinge = m1, 0; where: 03; is the rotation of the hinge between rigid regions i and j. The total energy dissipated will therefore be: toral ED.= Y mlyj6,; ai finges A principal difficulty with yield line analysis is finding a compatible set of hinge rotations, 6;;. However, because we assume that all rotations are small, the rotations of each rigid region can be plotted as vectors (using a right-hand screw-rule), and the difference between these vectors is the rotation along the hinges. Work done by loads: The work done by a point load W displacing a distance 6 is still W8. The work done by a uniform pressure appears to be much harder to calculate. Formally it could be written as W.D. = J p8(x,y) dA, where 8(x,y) is the displacement at a point on the slab. An alternative would be to find the centroid of each rigid region, as the average distance moved by the load on that region. A much simpler alternative, however, is to simplify the expression for work done to W.D = pV, where V = J'8(x,y) dA is the volume swept out by the collapsing slab. V can be easily calculated, as most collapse mechanisms can be split into roof sections, and pyramid sections. roof section pyramid sectionHandout 4, Plastic Theory 163 4.7.4 Example — Balcony Estimate the collapse load W to cause the collapse of the baleony shown below, which has a moment capacity per unit length m. Vector diagram of rotation164 1B Structural Mechanics 2001/2 Energy dissipated Hinge 12 (hinge between rigid regions (1) and (2)) Hinge | (hinge between support and rigid region (1)) Total Work equation 4.7.5 Example — ply-supported slab Estimate the uniform pressure p required to cause the collapse of the simply-supported slab shown below, which has a moment capacity per unit length m. 2bHandout 4. Plastic Theory 165 Vector diagram of rotation Energy dissipated Work done166 1B Structural Mechanics 2001/2 Work equation Optimized analysis An optimized analysis using Dr Middleton’s “Cobras’ programmes, which allows the geometry of the collapse mechanism to vary, and also tries different collapse mechanisms, finds that the ‘true’ collapse load to be p = 14.1m/L?, for a collapse mechanism similar to the one we analyzed, but with the central hinge line slightly shorter. a — 7 [ty Questions 1,2 and 3, Examples Sheet 2/5 pencmneenis| 4.8 Slip plane analysis of continua 4.8.1 Introduction As a final example of an upper bound method, we will look at the failure of rigid-plastic blocks of material (a continuum), for example in the figure below:Handout 4. Plastic Theory 167 | force Fn length rigid-plastic continuum Applications Soil mechanics: Design of foundations, embankments etc. Metal-forming: Extrusion, indentation etc. Assumptions Rigid-plastic material We shall assume that the material will only fail in shear at a shearing stress k, and is otherwise rigid (This is an example of the Tresca yield criterion, which we will examine in more detail later in the course) rigid-plastic Small deformations ‘The usual assumption that the geometry remains essentially unchanged All the examples in this course will be 2D plane strain — we will take a unit depth into the page, and assume no deformation out of the plane, but the method is equally applicable to more complex geometries in 3D.168 IB Structural Mechanics 2001/2 Methodology ‘We will again use the upper-bound work method, where we postulate a compatible mechanism, and equate the work done by the loads (or possibly self-weight, for e.g. the collapse of an em- bankment), with the energy dissipated in the material during an incremental deformation. For slip plane mechanisms (and also for plastic hinges and yield lines), work methods are often written in books using velocities and equating rates of work, and rates of deformation. However we will continue to consider small deformations — our formulation is identical to a rate of work formulation where everything is multiplied by some small time 8. 4.8.2. Energy dissipated in shearing ‘The energy dis as shown below jpated in slip plane mechanisms is dissipated in narrow bands of intense shearing, Shear strain in the shear zone Engineering shear strain is measured as the change in an angle that was originally a right-angle in the material, the angle -y shown above. Stresses and strains will be discussed in more detail later in the course.Handout 4, Plastic Theory 169 Energy dissipated per unit volume of shear zone Total energy dissipated Note that the energy dissipated does not depend on the thickness of the shear zone — we shall assume that itis infinitesimally thin. 4.8.3 Slip-circle mechanism ‘One way that the example shown in Section 4.8.1 might fail is by the formation of a slip circle, which is certainly a compatible mechanism:170 1B Structural Mechanics 2001/2 F, funit length | rigid block Collapse mechanism x rigid-plastic continuum eae Consider a unit depth into the page. Work done by load Energy dissipated along slip circle Area of slip circle Relative displacement along slip circle Total energy dissipated Work equationHandout 4. Plastic Theory 71 4.8.4 Triangular blocks mechanism Another class of mechanisms commonly considered is where the slip planes split the material into rigid triangular blocks. In this case, it is necessary to use a displacement diagram to ensure that the mechanism is compatible. Consider a mechanism for the indentation problem where all of the blocks are equilateral triangles. F funit length | Displacement diagram — Assume the indenter moves down a distance 6. Relative displacement along slip planes Consider unit depth into the paper172 1B Structural Mechanics 2001/2 _ Work done by load Energy dissipated along slip planes Work equation 4.8.5 Estimate the force/unit length required to indent a rigid-plastic material with yield stress in shear k between two long rigid anvils of breadth b Forming Process F funit length F hunit lengthHandout 4. Plastic Theory 173 4.8.6 Mechanism 1 F, funit length F, junit length Displacement diagram — Assume anvils move down/up by 5174 1B Structural Mechanics 2001/2 Work equation (Considering unit depth into the paper) 4.8.7 Mechanism 2 F, Juni length Fy Junit lengthHandout 4, Plastic Theory 175 Displacement diagram — Assume anvils move down/up by 5 ‘A new feature of this mechanism is that it includes sliding between the anvil and the material (interface 12) — and we don’t know the properties of this interface. However, we can consider two limiting situations: Zero friction no energy is dissipated in the interface; Infinite friction material directly next to the anvil will shear as a normal slip plane. These limiting cases give us some idea of the effect of friction on the forming process. Work equation (i) Zero interface friction (considering unit depth into the paper)176 1B Structural Mechanics 2001/2 Work equation (ii) Infinite interface friction (considering unit depth into the paper) For different thicknesses of material, these upper-bounds are summarized on the following ‘non-dimensional graph: A Fike Try Questions 4,5 and 6, Examples Sheet 2/5|Handout 4. Plastic Theory 7 4.9 Lower-Bound Analysis 4.9.1 Intoduction So far, our plasticity theory has only considered work equations for compatible collapse mech- anisms (ignoring equilibrium), and this gives upper bounds on collapse loads. However, there is another side to plasticity theory. This Section will show that an alternative method of con- sidering equilibrium without considering compatibility allows us to find a lower-bound on the collapse load, 4.9.2. Collapse of a propped cantilever _— In Section 4.1 we looked in detail at the collapse of a statically determinate beam. Here we will look at the more complex series of events that occurs during the collapse of a statically indeterminate structure. We shall assume that the moment-curvature relationship is a simplified elastic-plastic relationship, for simplicity (see Section 4.1.4 for a more detailed analysis) Initial Calculations We shall make the structure statically determinate by adding a pin at the left hand support, and hence splitting the response in two.178 1B Structural Mechanics 2001/2 Particular Equilibrium Solution in equilibrium with the applied load \" A Bending moment State of Self-Stress in equilibrium with zero applied load A Bending moment sy B a ‘The Particular Equilibrium Solution must be present for equilibrium with the applied load. There is also an unknown amount of the State of Self Stress — but we cannot find the magnitude (M) by equilibrium alone. Moments — For any value of M, the moment at the root is: ‘The moment at the centre is: Data Book o= we 1" T6EL Ml = SEI Before Yield The structure is elastic, and we can use compatibility to find M, the magnitude of the state of self-stressHandout 4, Plastic Theory 179 Compatibility 0, Bending moment First Plastic Hinge Forms The maximum moment is at the root of the cantilever. A plastic hinge will first form here when Although a plastic hinge has formed, this does not lead to collapse. As we know from the upper-bound method, a propped cantilever cannot collapse until two plastic hinges have formed. Single Plastic Hinge We can no longer use compatibility to find M, the magnitude of the state of self-stress. Because a plastic hinge has formed, itis quite possible for a kink to appear here, and hence all we can say is that 0; > 6. We can, however, use our knowledge of plasticity to find M. As a plastic hinge has formed, we know that180 1B Structural Mechanics 2001/2 Bending moment The plastic hinge prevents any increase in the magnitude of the state of self-stress. The load can keep increasing, however, but the load is taken by increasing the particular solution only. Final Collapse Once the bending moment at the centre also reaches M,, the structure can carry no more load, and the final collapse load is reached. Bending moment What load corresponds to this collapse load? ‘There are two approaches: (i) consider the Particular Equilibrium Solution and State of Self-Stress (we'll do this later); (ii) Consider free- body diagrams of the sections of the cantilever as it collapses — we know the moment at the centre is +Mp, and at the root is ~My.Handout 4. Plastic Theory 181 Vertical forces on 4 material around B: ‘The exact sequence of events leading to collapse depends on the initial conditions — we have assumed that the structure was initially stress-free. If it wasn’t, due to e.g. settlement of supports, the sequence would change, and for instance the first hinge to form may be the one in the centre. However, the final collapse state would be the same whatever the initial conditions. 4.9.3 Lower-bound theorem of plasticity If a set of internal stresses can be found in the structure that are in equilibrium with an applied load Weqyl, and nowhere violate the yield| condition, then the applied load will be less than, or equal to, the actual collapse load We. Wequit S We A proof of this theorem will be given at the end of the handout. We can use the lower-bound theorem to avoid having to consider the detailed sequence of events leading to collapse. For beam structures, all we need to do is to find a set of moments in equilibrium with the applied load that nowhere exceed the plastic moment M,. We must have the Particular Equilibrium Solution, but we can add as much or as little of the State of Self-Stress to give the largest possible load. We don’t have to consider compatibility at all — kinks will form as necessary at plastic hinges182 1B Structural Mechanics 2001/2 Equilibrium systems for the propped cantilever All the systems below are in equilibrium with the applied load, but only one is the optimum lower-bound solution. We will use the Particular Equilibrium Solution and State of Self-Stress shown in Section 4.9.2. Equations 4.1 and 4.2 then give the moments at the root and the centre in terms of the magnitude of the Self-Stress, M_ The moment at the centre is: 1. Choose M = 0. A Bending moment yield limit If this solution is about to violate the yield condition (at the centre): 2. Choose M = W1/10.Handout 4. Plastic Theory 183 A Bending moment 4 LLLLLLLLLL, yield limit SR If this solution is about to violate the yield condition (at the centre): 3. Choose M = W1/6. A Bending moment yield limit + If this solution is about to violate the yield condition (at the centre and root simultane- ously): 4. Choose M }W1/16 (clastic solution).184 1B Structural Mechanics 2001/2 A Bending moment yield limit | SSSA If this solution is about to violate the yield condition (at root): 5. Choose M = W1/4. A Bending moment yield limit If this solution is about to violate the yield condition: All of the solutions given are in equilibrium, and do not violate yield, and are therefore Jower-bounds on the true collapse load W. Wi = Ws < Wa < Wy < Wy = WeHandout 4, Plastic Theory 185 Ws is the largest (optimum) lower-bound solution, which equals the smallest (optimum) upper-bound solution from a collapse analysis (see Section 4.5.1), and is therefore the true col- lapse load. Although the lower-bound method can be applied to many different systems, e.g. frames or slabs, equilibrium for these systems can be difficult. We will instead only concentrate on multi- span beams, where a Particular Equilibrium Solution, and States of Self-Stress, are very easy to calculate, 4.9.4 Useful equilibrium results It is helpful to remember some equilibrium results for a simply-supported beam of length J (al- though these can, of course, easily be calculated from first principles). The moment sign con- vention is given in the Structures Data Book, and all forces are applied vertically down. * When a central point load of magnitude W is applied, the largest magnitude bending mo- ment is M = ~W1/4 at the centre of the beam, and the bending moment diagram is piece- wise linear. ‘* When a point load of magnitude W is applied at a distance a from a support, the largest magnitude bending moment is M = —Wa(!—a)/I at the point where the load is applied, and the bending moment diagram is piecewise linear. © When a distributed load of magnitude w per unit length is applied to the entire beam, the bending moment at a distance x from a support is M = —wx(! ~x)/2, and the largest magnitude bending moment is M = —wi?/8. * When a clockwise couple C is applied at a support, the bending moment at a distance x from the other support is M = Cx/I. Try Question 7, Examples Sheet 2/5 4.9.5 Example — two-span beam For the structure shown below, use the lower-bound theorem to find a UB (Universal Beam) that would be suitable for the entire span AE. The loads shown include a safety factor. The beam is to be made from steel with yield stress 6, = 300 N/mm’, ie kN }” kN ‘We will make this determinate by adding pins at the joints,186 1B Structural Mechanics 2001/2 Particular Equilibrium Solution 400 KN }” kN Bending moment A. B State of Self-Stress 1 +3 Bending momentHandout 4. Plastic Theory 187 State of Self-Stress 2 ue Bending moment A B c D E All possible equilibrium systems must contain the Particular Equilibrium Solution, but can in- clude any amount of the two States of Self-Stress. General Bending Moment Diagram 400 kN 300 kN Bending moment Optimum solution ‘We want to design a beam that is as small as possible, while still finding a set of moments where \M| -100 -200 | -300 | -400 id Again, because this BMD: (i) is in equilibrium with the applied load; (ii) nowhere exceeds the yield limit, then this is a safe design, by the lower-bound theorem. 4.9.7 Example — two-span beam 3 In some ways, the previous structure was a foolish design, as the strength was reduced at a point where there was a peak in the bending moment (at C). A better design would reduce the strength alittle further along, e.g. at F, below.192 1B Structural Mechanics 2001/2 400 kN 300 kN 3m. 3m 7imim 2m A B Che D E Bending moment Because C is now full-strength, we can design the left-hand part of the beam (AF) as for the first design. Hence we choose My = Mz = 300 kNm, giving |M|max as 300 KNm. We can now investigate FE. The minimum moment will occur at D Mp for FE must be at least this — we also need to check the moment at F. We have now ensured that |M| < 300 KNm in AF, and |M| < 150 kNm in FE, For AF we require Z, > 1000 cm’, and can choose UB 356 x 171 x 57 as before. For FE we require Zp > 500 cm’, and from the structures data book, we can choose The final bending moment diagram is shown below:Handout 4. Plastic Theory _ 193 4 Bending moment (kNm) 300 - 2004+} 100 0 -100 44 -2004 1 -3004 Despite all of this optimization, it is likely that the initial design would be the cheapest, because of the simplicity in construction, 4.9.8 Example — A three-span beam Ithas been suggested that a suitable design for the three-span beam shown below it to have My = {32:16:32} KNm in the three spans {AB:BC;CD}. Use the Lower-Bound Theorem to show that this is a safe design 10 kNim. Make the structure statically determinate by adding pins over the supports, and calculate the Particular Equilibrium Solution and the States of Self-Stress.194 1B Structural Mechanics 2001/2 _ Particular Equilibrium Solution 10 kN/m Bending moment State of Self-Stress 1 M My Bending moment State of Self-Stress 2 My M, Bending momentHandout 4, Plastic Theory 195 General Bending Moment Diagram All possible equilibrium systems must contain the Particular Equilibrium Solution, but can include any amount of the two States of Self-Stress. We are not trying to find an optimum design, simply to show that the one we are given is suitable. Hence, try making M; and Mz as large as possible without violating the yield criterion, and check that the resulting moments elsewhere in the beam are suitable. ‘The central span is easy to check. The minimum moment will occur in the centre of the beam. ‘The outer spans are more difficult. We can no longer say that the minimum moment will occur in the centre of the beam. However, itis straightforward to find an expression for the moment. The minimum will occur where dM /ds196 1B Structural Mechanics 2001/2 ‘Thus the moments everywhere are less than the plastic moments, and the design is safe. The chosen bending moment diagram, which shows this clearly, is below. ‘a. Bending moment (kNm) 32 4 164 ALLL LLL Ltt tts 0 T T > A B cS D -165 SN] 32 x We could have chosen M; and Mp to be slightly smaller, and still have found a bending moment diagram that didn’t violate yield. However, there is no need — a single valid equilibrium state shows that the design is safe. Try Questions 8, 9 and 10, Examples Sheet 2/5| 4.9.9 Justification for Elastic Analysis We have seen for a number of examples that an elastic analysis is not suitable for finding the collapse load of a structure. Indeed, the basic initial assumptions that a structure in unstressed in its initial state are probably not valid. Despite this, elastic analysis is still commonly used — largely because it is straightforward to write a computer program to do the analysis. The Lower-Bound theorem provides a justification for the elastic designer. The elastic so- lution is certainly an equilibrium solution (that also satisfies various compatibility constraints). ‘And an equilibrium solution that does not violate yield is safe — hence the elastic design is safe, even though it is unlikely to be an optimum design.Handout 4, Plastic Theory 197 4.10 Summary of Plasticity 4.10.1 Assumptions For both lower and upper-bound method: Ductility Itis important that the complete failure mechanism is able to form before the material loses load-bearing capacity at any point, We have often assumed that the material is rigid- perfectly plastic. Small deformations Although plastic deformation may be large compared with elastic defor- mation, we still assume that the deformations are small compared with the overall dimen- sions of the structure. Ignore other failure modes We have assumed that failure will not occur due to e.g. buckling. 4.10.2 Bound theorems We will only prove the bound theorems for beam structures with point loads applied, for brevity. Our proof will also take a few (intuitively reasonable) shortcuts, such as assuming that there is only one collapse load for any chosen load distribution. A more general proof will be found in Calladine's ‘Plasticity for Engineers’ Lower-bound theorem If a set of internal stresses can be found in the structure that are in equilibrium with an applied load W,, and nowhere violate the yield condition, then the applied load will be less than, or equal to, the! actual collapse load We Consider a complete collapse solution, where loads W¢ are displacing by 8 at positions i, and moments MF have developed at critical points j in the beam, where kinks @; occur. The loads WE are in equilibrium with the moments ‘M§, and the displacements 3; are compatible with the rotations 8). We can thus write a statement of virtual work: Iwe Now consider another set of loads where W/ = BWE. Also consider a set of moments MY that are everywhere less than or equal to the plastic moment, and are in equilibrium with the new loads W} (note that MF are not necessarily BMF — we are happy with any set of equilibrium MF; (43)198 1B Structural Mechanics 2001/2 moments, not just a set that are scaled from the collapse solution). We can also write, by virtual work, using the new equilibrium system, and the original compatible system: Lwis =D Mijo; (4) Hence BL ws, LMjo; =BL Mie; 46) ‘We now explore every possible critical point in the beam j. There are three possibilities for the complete collapse solution: rMo; (45) 4.3 and 45 give 1. A ‘positive’ plastic hinge has formed, 8; > 0, M° = Mp, and because M' < Mp, MF; > iO. 79) 2. A ‘negative’ plastic hinge has formed, 8; <0, MC = MSO; > M‘0;. ‘Mp, and because M! > —Mp, 3. No plastic hinge has formed, there is no kink, ®; = 0, and hence M°0; = M50j. Overall, for every possible critical point, M5@; > M¥8;, and so DMje 2 LMjo, 47) Thus, substituting into 4.6, B < 1. Thus, if a set of moments can be found in equilibrium with some applied load that nowhere violate yield, it would be necessary to multiply those loads by 1/B > | to cause collapse, proving the lower-bound theorem, Upper-bound theorem ‘An estimate of the plastic collapse load, Wiech, calculated for any arbitrary compatible mechanism by equating the work done by the applied load, and the plastic energy dissipated, will either be greater than, or equal to, the actual collapse load We. Consider that we have a complete collapse solution, where there are loads W/° at positions i on the beam, and moments MF have developed at critical points j. The loads we are in equilibrium with the moments M¢. Consider a postulated collapse mechanism — a compatible set of displacements, 6/7 and hinge rotations @¥ that may or may not be the correct collapse mechanism. By virtual work we weal = Yacoy as)Handout 4. Plastic Theory __ 19 Now consider doing a work calculation to find a load factor ‘yfor this assumed mechanism. Each assumed hinge does work M,|@4|, giving: YLWES! = Mylo4 | 49 Atevery section, ~My < ME < Mp, and hence MSOY < Mp|0Y|. Thus cou y Myo < EM/I07 | 4.10) Substituting 4.8 and 4.9 into 4.10 gives y> 1. Thus, the collapse load calculated from a work calculation with a compatible collapse mechanism will either be high, or correct, proving the upper-bound theorem.