Abstract
Fatigue debond growth of bonded composite specimens in mode I, mode II, and mixed-mode I/II loadings were studied experimentally and numerically. For the mode I loading, test data sets were analyzed under the linear elasticity condition. For the mode II and mixed-mode loadings, specific test data processing was used to ensure that nonlinear effects that act outside of the cyclic loading range are excluded from the fatigue behavior assessment. An important finding under the mode II and mixed-mode loadings is that the actual loading ratios in the tests varied and were different from the applied displacement ratio. The experimental characterization of the fatigue debond behaviors was complemented with numerically determined correlations between the compliance and debond length under each individual loading. This correlation was used to determine an effective debond length from the compliance determined experimentally. According to the actual loading ratio variation, the fatigue growth behaviors were characterized using Paris’ law. Numerical fatigue analyses using two codes, Abaqus finite element (FE) and AFGROW, were carried out and validated using the developed test database. The details on an FE model setup and analysis strategy for a mixed-mode specimen are presented. Good agreement in the fatigue lives was obtained between the test and numerical analyses for the three types of loading. Also, good agreements in the load versus debond growth length relations were obtained between the test and FE analysis in the mode II and mixed-mode loadings. Reasons for the result discrepancy are discussed. The study shows that the presented analysis procedure is effective. The thorough description of both the used test data processing and numerical analysis procedure fills a knowledge gap in the available literature.