Curvature-driven modelling of human-mimetic surfaces for fabric-drape testing

Purpose This study establishes a computer modelling workflow to develop curvature-driven human-mimetic surfaces for enhanced fabric drape testing. Standard drape testing relies on flat disc-shaped supporting tables. These tables fail to replicate mechanical textile–body interactions governed by complex geometries of anatomical surfaces. The new approach aims to enhance deformation analysis by creating shapes representing curvature of the body while ensuring compatibility with existing protocols. Design/methodology/approach The four-stage workflow analysed curvature metrics (mean, Gaussian and principal) across 10 ASTM-compliant female avatars (sizes 2–20; 301.941 sampling points), classified surfaces into convex, saddle and concave types and synthesised parametric models using four curve forms, namely cubic Bézier, cubic B-spline, quadratic Bézier and quarter-ellipse. The generated surfaces were validated via seven non-parametric statistical indicators and evaluated in Clo3D simulations by comparing stress distributions and drape properties against flat discs. Findings Curvature analysis revealed that convex (45.4–49.7%) and saddle (48.6–52.7%) surfaces dominate anatomical topography. Quarter-ellipse models achieved superior curvature replication, up to 99.6% coverage. Simulations confirmed human-mimetic surfaces localise warp/weft stresses on protrusions and shear stresses in transitions, aligning with stress distribution in clothing, unlike uniform stress distribution on flat discs. An open-access library of 34 validated surfaces (17 convex–saddle pairs, diameters 20–857 mm) was published. Originality/value This is the first approach to reproduce curvature of the human body in supporting tables for drape testing, enabling anatomically driven fabric–body interaction modelling. The parametric workflow, curvature database and published library of 3D models bridge anthropometric diversity with textile performance analysis for both digital and physical testing.