Question 3 of 7: Plastic Design of the Rigid Frame and its Welded Knee
Nivaar worked solution (AI-drafted; not reviewed by a licensed engineer)
Notes on this paper
Paper format. 98-Civ-B2 Advanced Structural Design, National Exams May 2016 — 3 hours, closed book (design handbooks permitted). Seven design questions of equal value; any five constitute a complete paper, and all seven are solved here.
Given. An L-shaped rigid frame, built in at both ends, whose member strengths are prescribed in the ratio 1 : 1.5.
Quantity
Value
Beam AB
8 m, built in at A, plastic capacity Mp
Column BC
10 m, built in at C, plastic capacity 1.5Mp
Factored vertical loads
525 kN at mid-span of AB; 750 kN at joint B
Factored horizontal load
150 kN at mid-height of BC
Lateral support
at A, mid-span, B, mid-height of BC and C
Steel grade
Fy = 350 MPa
Find. (a) the plastic moment the frame actually needs and rolled sections that supply it, and (b) a welded knee connection at B that can deliver the beam's plastic moment into the column.
Figure 3 as drawn on the exam sheet (built in at A and at C, rigid corner at B), and beneath it the governing collapse mechanism. The 500 kN load sits over the joint and does no work in this mechanism, but it dominates the column axial force.
Approach. Count the redundants, enumerate the independent mechanisms and every combination of them, take the largest required Mp, then confirm the answer with a lower-bound moment field obtained by stepping the load up hinge by hinge. Select sections on strength and on the Class 1 requirement that plastic design imposes, and size the connection on the beam's own capacity rather than on the demand.
Part (a) — establish the degree of redundancy. Two built-in supports give six reaction components against three equations of equilibrium, so the frame is three times redundant and a complete mechanism needs four hinges. A partial mechanism, in which only part of the frame moves, may need fewer — and that is what happens here.
Beam mechanism. Joint B cannot move vertically (the column is axially rigid) and cannot move horizontally (the beam is axially rigid and A is built in), so hinges at A, at mid-span and at B alone form a valid mechanism. Note that the hinge at B forms in the beam, the weaker of the two members meeting there. With a virtual rotation $\theta$ at each end and $2\theta$ at mid-span, and mid-span deflection $4\theta$: $$525(4\theta) = M_{p}\theta + M_{p}(2\theta) + M_{p}\theta \Rightarrow M_{p} = \frac{2100}{4} = 525\ \text{kN.m}$$ The 750 kN load at B does no work because B does not move.
Column mechanism. Hinges at C, at the 150 kN load point and at B, with the load point displacing $5\theta$, give $$150(5\theta) = 1.5M_{p}\theta + 1.5M_{p}(2\theta) + M_{p}\theta \Rightarrow M_{p} = \frac{750}{5.5} = 136.4\ \text{kN.m}$$ far less demanding than the beam mechanism, because the column is both stronger and more lightly loaded.
Test every combination, and watch the sign at B. Let the column-mechanism rotations be $r$ times the beam-mechanism rotations. At the shared hinge B the beam mechanism puts tension on the outside of the knee while the column mechanism puts it on the inside, so the two rotations subtract and the hinge partially cancels: $$M_{p}(r) = \frac{2100 + 750r}{3 + |1-r| + 4.5r}$$ This is a maximum at $r = 0$. Combining the mechanisms therefore lowers the required capacity (380 kN.m at $r = 1$), the opposite of the usual portal-frame result, so $$\boxed{M_{p} = 525\ \text{kN.m}\quad\text{and}\quad 1.5M_{p} = 787.5\ \text{kN.m}}$$
Confirm with a lower bound. Because the collapse mechanism is partial, the column moments are not fixed by statics and must be traced. Stepping the factored loads up on an elastic-plastic model, hinges appear at A at load factor 0.849, at mid-span at 0.898 and at B at 1.0000 — collapse landing exactly on the factored load, which confirms the mechanism analysis. The moment field at collapse is $-525$ at A, $+525$ at mid-span, $-525$ at B, $+103$ at mid-height of the column and $-19$ at C, every value within capacity, so the answer is exact rather than merely an upper bound.
Extract the member forces the sections must carry. From the beam free body with $M = -525$ at A and $+525$ at mid-span, $V_{A} = (525 + 525)/4 = 262.5$ kN, so the beam delivers $525 - 262.5 = 262.5$ kN to the joint and the column axial force is $262.5 + 750 = 1012.5$ kN. The column shear is $(103.1 + 525)/5 = 125.6$ kN in the upper half and $-24.4$ kN in the lower half, the difference being the 150 kN applied load.
Select the beam. Strength requires $Z_{x}\ge M_{p}/(\phi F_{y}) = 525\times10^{6}/(0.9\times350) = 1.667\times10^{6}$ mm$^{3}$. W460 by 82 gives $Z_{x} = 1.810\times10^{6}$ mm$^{3}$ and $\phi M_{p} = 570$ kN.m, is Class 1 on both flange ($b/2t = 5.97$) and web ($h/w = 43.2$), and satisfies the Cl 13.7 limit on unbraced length next to a plastic hinge: with double curvature ($\kappa = +1$) $L_{cr} = r_{y}(25\,000 + 15\,000\kappa)/F_{y} = 42.4\times40\,000/350 = 4847$ mm against the 4000 mm actually provided.
Select the column, where the section class governs. Strength alone needs only $Z_{x}\ge2.50\times10^{6}$ mm$^{3}$, which W360 by 147 satisfies at $\phi M_{p} = 885$ kN.m. But plastic design requires Class 1 sections throughout, and the Class 1 flange limit is $b/2t\le145/\sqrt{350} = 7.75$: W360 by 147 is 9.34 (Class 3), W360 by 162 is 8.51 and even W360 by 179 is 7.80, all Class 2. The lightest Class 1 shape in the series is W360 by 196 at $b/2t = 7.14$, giving $\phi M_{p} = 1199$ kN.m. It is 52 per cent stronger than required, and the surplus is the price of the class requirement, not of the analysis.
Part (b) — set the design force for the knee at B. A connection at a plastic hinge is sized on the capacity actually installed, not on the demand, so that the hinge forms in the member and not in the joint. With $\phi M_{p} = 570$ kN.m for the beam, each flange delivers $$T_{f} = \frac{570\times10^{6}}{d - t} = \frac{570\times10^{6}}{460 - 16} = 1284\ \text{kN}$$
Weld the beam to the column. Complete-joint-penetration groove welds from each beam flange to the column flange develop the flange directly and are the standard detail for a moment knee. The web carries only the 262.5 kN beam shear: two 8 mm fillets along the 428 mm web depth give $2\times428\times1.244 = 1065$ kN using $V_{r} = 0.67\phi_{w}A_{w}X_{u}$ with $\phi_{w} = 0.67$, so 8 mm is ample and is set by the minimum size for a 16 mm flange rather than by strength.
Provide continuity stiffeners. Opposite each beam flange the column needs a stiffener to carry the 1284 kN across the column web. Taking the stiffeners alone, $A_{st}\ge1284\times10^{3}/(0.9\times350) = 4077$ mm$^{2}$; a pair of 150 mm by 16 mm plates gives 4800 mm$^{2}$ with $b/t = 9.4$ inside the $200/\sqrt{F_{y}} = 10.7$ limit.
Check the panel zone and add the diagonal. The shear carried across the joint panel is the flange force less the column shear, $1284 - 126 = 1159$ kN, against a web resistance of $$V_{r} = 0.55\phi F_{y}d_{c}w = 0.55\times0.9\times350\times372\times16.4 = 1057\ \text{kN}$$ The web is 10 per cent short, so a diagonal stiffener across the panel takes the balance. The diagonal is $\sqrt{460^{2} + 372^{2}} = 592$ mm long with $\cos\theta = 372/592 = 0.629$, so it must carry $(1159-1057)/0.629 = 162$ kN, needing $162\times10^{3}/(0.9\times350) = 513$ mm$^{2}$. A pair of 100 mm by 10 mm plates (2000 mm$^{2}$) is the smallest practical size and is adopted.
Required plastic moment as the column-mechanism rotation is blended into the beam mechanism. The curve falls away from r = 0, so no combination beats the pure beam mechanism — the opposite of the usual portal-frame result.
The welded knee at B. Sizing the connection on the beam's own plastic capacity (570 kN.m, not the 525 kN.m demanded) is what makes the panel zone deficient and forces the diagonal stiffener.