Torsional Wind Response in Asymmetrical Timber Buildings - A Parametric Study of Plan Irregularity in Mid-Rise Structures

Abstract

Wind loading often governs the lateral response of mid-rise timber buildings and can be critical for torsion, since the structure may rotate in addition to swaying. This thesis applies a parametric modal study of rectangular and L-shaped timber floor plans to identify when torsion governs the fundamental mode, and which stabilizer layouts most effectively reduce torsional sensitivity. The study is limited to the chosen investigated plan sizes and the structural configuration is based on a 6×6 m column grid with constant span lengths, rigid diaphragm action, and stabilizing systems modelled using CLT shear walls and (where applicable) a core, whose lengths, and positions are varied parametrically. Effects such as openings/discontinuities, height/vertical irregularities, connection flexibility, additional bracing systems, and explicit wind-response/comfort checks are outside the scope. Across both geometries, torsion is governed by the combined effect of (i) eccentricity, e, between the center of mass (CM) and the center of rigidity/rotation (CR), expressed as the normalized measure e/D, where D is the plan diagonal, and (ii) the torsional resistance provided by stabilizer lever arms, represented by the normalized torsional stiffness R =qKθ/(Kx + Ky). Here, Kθ is the torsional stiffness about CR, while (Kx and (Ky are the total lateral bending stiffnesses resisting sway in the global x- and y-directions. For rectangular plans, torsion becomes consistently likely once eccentricity is high. In the compiled results, configurations with a normalized eccentricity over the diagonal of the building plan (D), e/D ≥ 0.16 fall in the torsion-dominated region, while configurations with sufficiently high normalized torsional stiffness (R) (about R ≥ 13.5 m) remain translation-dominated. The most efficient torsion-reducing measures in the rectangular study were therefore avoiding stabilizer asymmetry that shifts CR (especially off-centre core placement) and increasing lever arms by placing stabilizers toward façades/corners. For L-shaped plans, torsion sensitivity is generally higher because geometric effects make low eccentricity harder to achieve in practice, so robustness relies more strongly on torsional resistance. In the combined L-shape summary, the key stiffness thresholds are Rcrit,1=10.0 m and Rcrit,2=28.11 m, with corresponding boundary ratios (R/(e/D)) of roughly 128 and 184. Practically, configurations below the lower stiffness level are consistently torsion-prone, whereas for moderate eccentricities, maintaining R above the upper level is associated with translational behaviour. The most effective measures in the L-shape study were moving stabilizers toward the plan corners and avoiding pronounced directional stiffness imbalance, which was shown to broaden the range of torsion-dominated configurations. Overall, the analyses indicate that the most efficient design takes are (1) controlling eccentricity by limiting CR shifts (dominant for rectangles), and (2) maximizing stabilizer lever arms/torsional resistance (dominant for L-shapes). Configurations combining high eccentricity with low torsional resistance are the most torsion-sensitive and should be prioritized for detailed wind-serviceability verification.