Failure prediction for fiber reinforced polymer composites based on virtual experimental tests

A novel virtual experimental testing methodology that utilizes virtual microstructures representative of genuine fiber distribution, is devised to micromechanically simulate and scrutinize the inter-fiber failure of unidirectional (UD) carbon fiber reinforced polymer (CFRP) composites. The experimentally identified nonlinear behavior of the matrix is incorporated by utilizing the Drucker-Prager plasticity model. The interaction behavior between fiber-matrix interfaces, including delamination and friction, is represented by means of a cohesive interaction approach. The results of virtual testing demonstrate that interfacial parameters and the epoxy properties exert considerable influences on the predicted macroscopic responses under individual loading cases. Additionally, the interface and epoxy material properties are inversely determined through virtual testing analysis assisted by experimental results. Furthermore, the validated virtual testing method is utilized to obtain the inter-fiber failure envelope of the composites. The predictions are then compared with those theoretically derived from classical failure criteria. The current comparative studies indicate that the existing classical failure criterion exhibits limited predictive ability with computational data.