Question 8 of 8: Deriving the Equilibrium Equations and the Stiffness Matrix for a Gable Frame
Nivaar worked solution (AI-drafted; not reviewed by a licensed engineer)
Notes on this paper
Paper format. National Examinations, May 2016 — 98-Civ-B1 Advanced Structural Analysis, three hours, closed book (an approved Sharp or Casio calculator is permitted). Questions 1 and 2 are compulsory at 8 marks each; the candidate then answers two of Questions 3, 4 or 5 and two of Questions 6, 7 or 8 at 21 marks each, so six questions make a complete paper of 100 marks. All eight questions are worked here, because this set is a study resource rather than a sat examination.
Reference texts. R. C. Hibbeler, Structural Analysis, 10th ed. (Ch. 6 influence lines, Ch. 9 virtual work and Castigliano, Ch. 10 force method, Ch. 11 slope-deflection, Ch. 12 moment distribution, Ch. 16 stiffness method for frames); A. Kassimali, Structural Analysis, 6th ed.; A. Ghali, A. M. Neville & T. G. Brown, Structural Analysis: A Unified Classical and Matrix Approach, 7th ed.; J. C. McCormac, Structural Analysis: Using Classical and Matrix Methods, 4th ed. Canadian design context for the same structures: CSA S16, CSA A23.3 and the National Building Code of Canada 2020.
Check: sign conventions used throughout. Slope-deflection and moment-distribution end moments \(M_{ij}\) and joint rotations \(\theta\) are clockwise positive (Hibbeler), and the chord rotation \(\psi_{ij}\) is positive when the member chord rotates clockwise. Internal bending moments are plotted sagging positive, converted from the end moments by \(M_{\text{sag}}(i)=M_{ij}\) at the near end and \(M_{\text{sag}}(j)=-M_{ji}\) at the far end. Truss bar forces are tension positive.
Question 8: Deriving the Equilibrium Equations and the Stiffness Matrix for a Gable Frame (21 marks)
Find. (a) the translation equilibrium equation at joint 3, (b) the two joint moment equilibrium equations, and (c) the terms of \([K]\) and \(\{P\}\). The equations are not to be solved.
Question 8 — symmetric gable frame, fully built in at joint 1 and on a horizontal roller at joint 3; 9.6 kN/m acts normal to the left rafter over its whole length. δ is the translation of joint 3, taken positive to the right.
Approach. Express the sway kinematics in terms of the single parameter \(\delta\), write the four end moments in the paper’s counter-clockwise-positive convention, then form one equation by virtual work on the sway mechanism and two by joint moment balance.
Sway kinematics of the gable. Neglecting axial strain, both rafters keep their length. Joint 1 is fixed and joint 3 has \(v_3 = 0,\ u_3 = \delta\). Writing the two inextensibility conditions,$$3u_2 + 4v_2 = 0,\qquad 3(\delta-u_2) - 4(0-v_2) = 0$$and solving the pair,$$u_2 = \boxed{\tfrac{\delta}{2}},\qquad v_2 = \boxed{-\tfrac{3\delta}{8}}$$so pushing the roller to the right lifts nothing — the apex moves right by half the roller movement and drops by three-eighths of it. Everything that follows hangs on these two numbers.
Chord rotations. The transverse component of the relative end displacement of each rafter, divided by its length, gives (counter-clockwise positive)$$\psi_{12} = -0.125\,\delta,\qquad \psi_{23} = +0.125\,\delta$$equal and opposite, as symmetry of the mechanism requires.
Fixed-end moments of the normal load. A load acting perpendicular to a member is a uniform transverse load on that member, so the standard formula applies with the member’s own length:$$\mathrm{FEM}_{12} = +\frac{wL^{2}}{12} = +\frac{9.6(5)^{2}}{12} = +20.0\ \text{kN}\cdot\text{m},\qquad \mathrm{FEM}_{21} = -20.0\ \text{kN}\cdot\text{m}$$in the counter-clockwise-positive convention the question specifies. No projection on to the horizontal is involved, because the load is normal to the member and not gravity-directed.
End moments with \(k = EI/L = EI/5\). With \(\theta_1 = 0\),$$M_{12}=2k(\theta_2+0.375\delta)+20.0,\qquad M_{21}=2k(2\theta_2+0.375\delta)-20.0$$$$M_{23}=2k(2\theta_2+\theta_3-0.375\delta),\qquad M_{32}=2k(2\theta_3+\theta_2-0.375\delta)$$
Part (b): moment equilibrium at joint 2. The apex is a rigid joint carrying no applied couple, so$$M_{21}+M_{23}=0 \;\Longrightarrow\; 2k(4\theta_2+\theta_3) = 20.0$$$$\boxed{8k\,\theta_2 + 2k\,\theta_3 = 20.0}$$The \(\delta\) terms cancel identically. That cancellation is a property to be verified, not assumed: it happens here because both rafters have the same length and the same \(EI\), so their equal and opposite chord rotations contribute equal and opposite \(3k\psi\) terms at the apex.
Part (b): moment equilibrium at joint 3. Joint 3 is a hinge on a roller, so the moment there is zero:$$M_{32}=0 \;\Longrightarrow\; \boxed{2k\,\theta_2 + 4k\,\theta_3 - 0.75k\,\delta = 0}$$
Part (a): the translation equation by virtual work. Impose the sway mechanism with \(\delta^{*} = 1\); then \(\psi^{*}_{12}=-0.125,\ \psi^{*}_{23}=+0.125\), and the mid-point of the loaded rafter moves \((0.25,\,-0.1875)\). The resultant of the normal load is \(9.6(5) = 48\) kN acting in the direction \((0.6,\,-0.8)\), so the external virtual work is$$W^{*} = 48\big[(0.6)(0.25)+(-0.8)(-0.1875)\big] = 48(0.3125) = 15.0\ \text{kN}\cdot\text{m}$$and the virtual-work statement \(\sum (M_{ij}+M_{ji})\psi^{*}_{ij} + W^{*} = 0\) becomes$$-0.125\,(M_{12}+M_{21}) + 0.125\,(M_{23}+M_{32}) + 15.0 = 0$$Substituting the end moments and collecting terms,$$\boxed{0.375k\,\delta - 0.75k\,\theta_3 = 15.0}$$The rotation \(\theta_2\) drops out of the translation equation, mirroring the way \(\delta\) dropped out of the joint-2 equation — the two cancellations are the same symmetry seen from opposite sides, and together they guarantee that \(K_{12} = K_{21} = 0\).
Part (c): assemble the matrix. With \(k = EI/5\), so that \(0.375k = 0.075EI\), \(0.75k = 0.15EI\), \(8k = 1.6EI\), \(4k = 0.8EI\) and \(2k = 0.4EI\), the three equations in the order (translation, joint 2, joint 3) are$$EI\begin{bmatrix} 0.075 & 0 & -0.150 \\ 0 & 1.600 & 0.400 \\ -0.150 & 0.400 & 0.800 \end{bmatrix}\begin{Bmatrix} \delta \\ \theta_2 \\ \theta_3 \end{Bmatrix} = \begin{Bmatrix} 15.0 \\ 20.0 \\ 0 \end{Bmatrix}$$or, entry by entry,$$\boxed{K_{11}=0.075EI,\; K_{12}=K_{21}=0,\; K_{13}=K_{31}=-0.150EI}$$$$\boxed{K_{22}=1.600EI,\; K_{23}=K_{32}=0.400EI,\; K_{33}=0.800EI}$$$$\boxed{\{P\} = \{15.0,\;20.0,\;0\}^{\mathsf{T}}\ \text{kN}\cdot\text{m}}$$As instructed, the equations are left unsolved.
Check the assembly without solving it. Three tests cost nothing. First, \([K]\) must be symmetric, and it is — an unsymmetric result would mean the translation row had been scaled differently from the moment rows. Second, every rotational term must be recognisable as \(4EI/L\), \(2EI/L\) or a sum of them: \(K_{22} = 8k = 4EI/L + 4EI/L\) is the two rafters meeting at the apex, and \(K_{23} = 2k = 2EI/L\) is the carry-over along member 2–3. Third, \(P_2 = 20.0\) is exactly the fixed-end moment \(wL^{2}/12\) released at the apex, and \(P_3 = 0\) because a roller cannot apply a couple.
Check: sign convention translated from the textbook form. The question prescribes counter-clockwise-positive rotations, whereas the standard slope-deflection presentation in Hibbeler is clockwise positive. The equations above have been written throughout in the paper’s convention, which flips the sign of every \(\theta\), every \(\psi\) and every fixed-end moment together; the stiffness terms themselves are unaffected, and only the signs in \(\{P\}\) and the off-diagonal \(\delta\) coupling would change if the other convention were used. A candidate who states the convention and uses it consistently loses nothing.