Co-authors​ :

 Georgios STAMOULIS (FRANCE), Nicolas CARRÈRE  

In recent years, adhesives have garnered increased attention in various industrial applications such as aeronautics, automobiles, and offshore projects. They are increasingly preferred over conventional assembly techniques like riveting and bolting due to several advantages, including more uniform stress distribution, the ability to join dissimilar materials, and overall weight reduction in structures. One crucial aspect is the calculation of the force at the break of adhesively bonded joints, which heavily relies on the critical Strain Energy Release Rate (SERR) [1]. The SERR represents the energy released when a crack propagates in a solid medium per unit area and is known to vary with the stress state at the crack tip [2]. However, most methodologies for measuring SERR are based on Linear Elastic Fracture Mechanics (LEFM), assuming both the adhesive and substrates are linear elastic. While this assumption may hold for substrates in most cases, it is less likely to be accurate for structural adhesives, which typically exhibit strong nonlinear and ductile behavior [3]. For instance, an analysis of SikaPower®-498 behavior in calculating SERR revealed minimal sensitivity to the mode ratio, except when approaching mode II [4]. Building upon the findings published in [4], this study investigates the fracture properties of a toughened epoxy adhesive (trade name withheld for confidentiality reasons) in the mixed mode I+II plane. The investigation employs (Tapered) Double Cantilever Beam ((T)DCB) and Mixed Mode Bending (MMB) tests at three loading speeds: 0.003, 0.3, and 30 mm/min. Initially, the use of LEFM theory indicated a significant dependency of the SERR on loading speed for all examined mode ratios, except for mode I [2]. Specifically, the SERR was found to increase with loading speed as the mode ratio approached mode II. In this work, the authors first identify the parameters of the Drucker-Prager plasticity criterion, as coded in the AbaqusTM software, to represent the mechanical behavior of the studied adhesive. This is based on monotonic tests conducted with the modified Arcan fixture at various loading angles. Subsequently, this behavior is integrated into Finite Element (FE) computations of the SERR, and the results are compared with the linear elastic assumption.