Failure Mode Interaction in Fiber Reinforced Laminated Composites.

A novel computational modeling framework to predict the compressive strength of fiber reinforced polymer matrix composite (FRPC) laminates has been presented. The model development has been motivated by a set of experimental results on the compression response of two diferent FRPCs. The model accounts for failure mode interaction between kink-banding and interface fracture (or delamination), which are observed in the experimental results. To reduce the size of the computational model, those interfaces that are most susceptible to delamination are Lrst determined through a free-edge stress analysis. Furthermore, o-axis layers, which are passive in the failure process are represented through an equivalent homogenized model, but the microstructural features of the on-axis layers (zero plies) are retained in the computational model. The predictions of the model matched well with the experimental observations, and they were found to accurately account for failure mechanism interactions. Therefore, this model has the potential to replace the need to carry out large numbers of tests to obtain the compressive strength allowable for FRPC laminates, the latter allowable being an essential element in the design of lightweight FRPC aerostructures. Furthermore, the thesis presents a new computational model to predict bermatrix splitting failure, a failure mode that is frequently observed in in-plane tensile failure of FRPC's. By considering a single lamina, this failure mechanism was seamlessly modeled through the development of a continuum-decohesive nite element (CDFE). The CDFE was motivated by the variational multiscale cohesive method (VMCM) presented earlier by Rudraraju et al. (2010) at the University of Michigan. In the CDFE, the transition from a continuum to a non-continuum is modeled directly (physically) without resorting to enrichment of the shape functions of the element. Thus, the CDFE is a natural merger between cohesive elements and continuum elements. The predictions of the CDFE method were also found to be in very good agreement with corresponding experimental observations.PHDAerospace EngineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/97862/1/pavana_1.pd