From low-velocity transverse impact towards failure: Multi-site damage pathways in GFRP-to-aluminum adhesive joints

Abstract

Impact loading represents a critical condition affecting the damage tolerance and residual strength of adhesive-bonded composite joints in advanced engineering structures. However, developing a prediction core capable of estimating damage distribution and severity after impact is essential for future condition monitoring of bonded joints. This study investigates the post-impact behavior of glass fiber–reinforced polymer (GFRP) to 2024 aluminum alloy single-lap joints subjected to transverse impacts using drop-weight tests. Specimens were categorized into non-impacted and impacted groups at energy levels of 1 J, 1.5 J, 2 J, and 2.5 J, and subsequently tested under quasi-static lap shear loading to assess residual strength. The originality of this work lies in: (i) developing a coupled numerical framework that integrates an extended three-dimensional Hashin (3D) criterion with a cohesive zone model (CZM) within a user-defined VUMAT to simultaneously capture intralaminar composite damage and adhesive failure under both impact and post-impact conditions, and (ii) establishing a quantitative correlation between impact energy, through-thickness damage modes, and residual joint strength in a hybrid GFRP–aluminum configuration, which has not been previously reported. Results indicate that impact energies between 1 J and 2.5 J lead to strength reductions ranging from 6.5% to 45%, with a transition from cohesive to mixed adhesive/cohesive failure accompanied by fiber breakage. Fiber tension and matrix compression were identified as the dominant damage mechanisms, extending from the impacted surface to the composite–adhesive interface. At the highest impact energy (2.5 J), combined fiber tension damage in the composite and the formation of a circular-shaped damage zone in the adhesive layer caused the maximum strength reduction. Damage within the composite–adhesive interface and adhesive layer was confirmed through failure mode analysis and thermal imaging. To quantify the predictive capability of the proposed numerical framework, the integrated 3D Hashin–CZM model achieved an average prediction error of 3.2% in estimating the post-impact residual lap shear strength across impacted specimens. This represents an average improvement of approximately 4.6% compared with a conventional model employing a 2D Hashin failure formulation, demonstrating the enhanced accuracy and reliability of the adopted multi-site damage modeling approach.