Experimental study and numerical simulation of adhesively bonded timber-concrete composite panels: bending behavior, adhesive shear and peel stress distributions

Abstract

This study investigated the structural response of timber-plain concrete panels bonded with epoxy and PUR adhesives under a four-point bending load, both experimentally and numerically. Tests evaluating epoxy- and PUR-bonded glulam-plain concrete panels measured ultimate load (Fult), effective bending stiffness (EIeff), and mid-span deflection (∆ult). Shear stress within the adhesives was monitored using fiber optic sensors. Numerical simulations estimated Fult and EIeff with errors of 5% and 14%, respectively. These discrepancies may stem from assuming defect-free wood (e.g., no knots), idealized boundary conditions, uniform adhesive thickness assumptions, and unmeasured mechanical and damage properties of wood and concrete (Detailed error source analysis is available in Section 4). Qualitative validation involved comparing sensor-recorded shear stress with simulation results. A parametric study evaluated the effects of adhesive thickness (0.5, 1, and 3 mm), concrete-wood depth ratio (50/85, 85/85, and 150/85 mm/mm), wood type (spruce (Picea), beech (Fagus), and Azobe (Lophira)), concrete strength class (C12/15 and C30/37), and span length (2, 4, and 8 m) on the bending behavior of the panel, and the shear and peel stress distribution within the adhesive bond line. Shear stress prevailed over peel stress in the adhesives, with peel stress approaching shear stress as span length increased (max. τa / max. σa ∼ 3.5–7). The ultimate load typically resulted in concrete compression damage and wood fiber damage, influenced by the concrete strength class and wood type. Higher depth ratios led to tension damage in concrete, while adhesive thickness had a minimal impact on stress distribution and failure modes.