Quasi-static compression and low-velocity impact behaviour of PVC foam cylinders: Experimental insights

Closed-cell PVC foams are widely used as core materials in lightweight sandwich structures; however, current characterisation methods rely on standard specimen geometries that are not suitable for minimally destructive or in-situ assessment. In this study, closed-cell PVC foam (Divinycell H-series) is examined under quasi-static compression and low-velocity impact using small cylindrical specimens (Ø20 mm, h/d = 1) designed to emulate minimally destructive coring, thereby enabling in-situ testing of installed panels with negligible disruption. Four densities (H100, H130, H200, H250) are tested at strain rates ranging 0.001–0.500 s ⁻¹ and impact energies of 8.33–25.00 J. Post-impact geometry and volume are quantified via surface-grid mapping and handheld 3D scanning, while scanning electron microscopy (SEM) is used to evaluate cell-level wall folding and lamination. The quasi-static responses follow the canonical elastic–plateau–densification sequence: density governs stiffness and strength, while strain rate acts as a secondary modifier whose influence intensifies with density. Under low-velocity impact, higher energy primarily extends displacement and accelerates densification in low-density grades, with only modest increases in peak stress; load–strain oscillations intensify with density, consistent with higher mechanical impedance. SEM reveals anisotropic cell morphology on planes parallel to the loading direction, along with impact-induced plastic flow, folding, and compaction of the cellular network. The results support the repeatability of the small-specimen approach and provide trends and reference data suitable for assessing and modelling sandwich-panel building components subjected to accidental impact and compression. By capturing both macroscopic responses and micromechanical deformation mechanisms, the dataset also establishes a foundation for future multiscale numerical modelling of impact behaviour, enabling more accurate predictive tools while reducing experimental cost and the need for destructive full-panel testing.