Co-authors​ :

 Ali SHIVAIE KOJOURI (BELGIUM), Javane KARAMI (BELGIUM), Kalliopi-Artemi KALTEREMIDOU (BELGIUM), Jialiang FAN (SWITZERLAND), Akash SHARMA (BELGIUM), Anastasios P. VASSILOPOULOS (SWITZERLAND), Véronique MICHAUD (SWITZERLAND), Wim VAN PAEPEGEM (BELGIUM), Danny VAN HEMELRIJCK  

Adhesive joints are used in different industries, including the wind turbine manufacturing sector, because of their advantages compared to traditional mechanical joining methods. By applying adhesives, there is no need to drill any holes; therefore, the structural integrity of the substrate materials remains intact. Moreover, compared to traditional mechanical joining methods, adhesive joints offer uniform stress distribution as well as the ability to join similar and dissimilar materials. In literature, there is a limited number of studies on the fracture behavior of thick adhesive joints. For instance, Miao et al. [1] showed both experimentally and numerically that the crack in thick adhesive joints loaded under pure mode I can initiate in the adhesive; however, the main tendency of the crack is to propagate towards the interface. Similarly, Fernandes et al. [2] experimentally revealed that the crack trajectory for a joint with a bond-line thickness of about 10 mm is not stable under mode I loading, and that the crack tends to propagate towards the interface from the middle of the specimen. This phenomenon is known as crack kinking and is depicted in Fig. 1. T-stress is known as constant stress acting parallel to the crack plane, and the positive T-stress can be a possible explanation for such behavior according to the finite element simulations provided in [3]. To the best of the authors' knowledge, there is no experimental study showing the influence of the T-stress on the crack path trajectory in thick adhesive joints. Consequently, in this study, pre-cracked thick adhesive joint samples manufactured by glass/epoxy composite adherends and a structural adhesive with a bond-line thickness of approximately 10 mm are experimentally tested under pure mode I loading. A DIC-assisted in-situ measurement technique [4] is used to determine the T-stress in the test samples, and the crack propagation angles are determined using the second-order crack kinking theory and compared to the experimental values. A good agreement is observed between the experimental and theoretical crack propagation angles. According to literature, the geometry is responsible for the positive T-stress and the crack kinking phenomenon [5], whereas the weak interface between the adhesive and the adherends is possibly not playing a major role. In order to validate this observation, two types of specimens, i.e. Compact Tension (CT) and Double Cantilever Beam (DCB), were manufactured out of PMMA and loaded under pure mode I. The same behavior, i.e. deviation of the crack from the middle of the specimen to the vertical path, is observed by changing the specimen geometry from CT to DCB specimens. It has to be noted that, in the PMMA specimens, no interface or voids were present in order to affect the stress field around the crack tip and cause the crack to deviate from the middle to a vertical path [5].