• Non-uniform façade temperature gradients are modelled with soil–structure interaction. • Gradient shape governs damage: wide vertical profiles halve the ΔT needed for cracking. • Stiff sandy soils amplify restraint, while softer soils delay crack initiation. • Geometry matters: large openings and low vertical masonry ratios lead to more damage. • Uniform temperature, fully restrained models overpredict cracking by 15–20°C. Temperature effects are frequently cited as the cause of light cracking in masonry façades, yet most modelling studies idealise thermal loading as uniform steps and represent restraint as fully fixed, assumptions that tend to exaggerate damage. This work evaluates whether realistic, non-uniform temperature gradients, like those produced by shading and insolation, together with soil–structure interaction as the dominant restraint mechanism, can generate cracking patterns consistent with field observations. A coupled thermo-mechanical FEM model with a homogenised masonry continuum and tensile softening is employed; the façade–foundation–soil system is modelled explicitly, and damage is quantified using a crack-based index Ψ. A parametric campaign (1200 simulations) spans two façade typologies (clay masonry on unreinforced masonry foundations; calcium-silicate on reinforced concrete strips), three layered soils, 33 geometries, and multiple vertical and two-dimensional gradient shapes. The results indicate that gradient shape is decisive: widely distributed vertical gradients trigger visible damage (Ψ≥1) at roughly half the temperature differential required by more localised gradients, with visible damage becoming likely around ΔT≈20°C (warming) and ≈25°C (cooling) for the most adverse shapes. Restraint stiffness governs severity: stiffer sandy profiles increase tensile stresses and cracking, whereas softer profiles accommodate thermal movement; relative to uniform, fully restrained models, crack initiation is delayed by ∼15–20°C and cracking is less distributed. Geometric discontinuities also dominate sensitivity: larger/more openings and low vertical-masonry ratios promote earlier localisation, while overall length/height is secondary. Fragility-like curves provide thresholds useful for assessment and mitigation.
Korswagen et al. (Sun,) studied this question.