Date created: | Last Updated:

Category: Project

Description: Delamination onset and propagation represent one of the most safety-critical failure modes in structures made of FRP laminates. Due to the widespread use of laminated composites in primary and secondary structures, the assessment of fracture behavior of FRP and the development of reliable analytical and numerical formulations for damage prediction are increasingly important themes to be addressed. Experimental procedures for the determination of fracture toughness have been developed for quasi-static conditions; while simulations have mainly dealt with quasi-static conditions for benchmarking and with impacts for prediction, which are difficult to compare with experimental measurement. Nonetheless, most of the applications in which composites are adopted are characterized by dynamic conditions and loading-rate effects. Thus, the study of loading-rate effects on delamination onset and propagation assumes a significant importance, although it has not been addressed much in the past. Loading-rate effects on delamination were addressed in a threefold way in the present work. Experimental observations were carried out on UD specimens under mode I, mode II and mixed mode I-II loading and on specimens with a 0/90◦ interface at the delamination front under mode I loading. For all the different lay-ups and loading conditions, four different velocities have been tested: 1, 50, 250 and 500 [mm/min]. SEM analysis was also carried out on delaminated surfaces under mode I loading both in UD and in the 0/90◦ interface. In the range of velocities analyzed, no dependence seems to be present for mode I, both in the UD and in the 0/90◦ interface, for mode II and for mixed mode I-II; differences are in the order of experimental scatter. A code was developed in Mathematica environment for the analysis and post-processing of frames of delamination propagation gathered during experiments, allowing for the extraction of more information from experiments with the use of simple optical devices. Dynamic numerical simulations were conducted in the FEM software ABAQUS, modeling the experiments performed. For mode I, mode II and mixed mode I-II with UD specimens, 3D models have been developed with hexahedral brick elements and a single layer of cohesive elements. Variations have also been studied: the presence of a cohesive layer with randomly distributed interface strengths has been analyzed for mode I, while the influence of initial lever misalignment has been addressed for mixed mode. For mode I with specimens with 0/90◦ interface, a 3D model has been developed with shell elements and with five layers of cohesive elements and intra-laminar damage modeling with the use of the Hashin criterion. Numerical models show a clear dependence on loading rate for all loading modes and lay-ups, except mode II with UD specimen; the agreement with experiments is good for quasi-static conditions. Initial lever misalignment does not have any significant influence. A simulation has also been performed of the cooling process following curing for the 0/90◦ laminated plate from which specimens were cut. Stress concentrations can be observed due to the mismatch in orientation between plies. Analytical models for DCB (mode I), ENF (mode II) and MMB (mixed mode I-II) specimens have been formulated on the base of the Euler-Bernoulli and Timoshenko beam.