A Rational Approach to Characterize Fracture Properties of Composites to Support a Cohesive Zone Material Model

Composites are replacing conventional materials in aerospace applications due to their light weight, non-corrosiveness, and high specific strength. This thesis aims to characterize the input data for IM7-8552 unidirectional composite to support MAT213, an orthotropic elasto-plastic damage material model and MAT_186, a mixed mode cohesive zone model used to model delamination. MAT_213 in conjunction with MAT_186 can be used to predict the behavior of composite under crush and impact loads including delamination. MAT_213 requires twelve sets of stress-strain curves, direction-dependent material constants, and flow rule coefficients as input. All the necessary inputs are obtained through the post-processing of a total of twelve distinct quasi-static and room temperature (QS-RT) experiments. MAT_186 is driven by a set of Mode I and Mode II fracture parameters and traction separation laws, a constitutive law that derives the relationship between stresses and relative displacements at integration points of cohesive elements. Obtaining cohesive law parameters experimentally is a tedious process as it requires close monitoring of the crack length during the test, which is a difficult task to achieve with accuracy even after using sophisticated equipment such as Digital Image Correlation (DIC). In this thesis, a numerical inverse analysis method to precisely predict these parameters by using finite element analysis with cohesive zone modeling and response surface methodology (RSM) is proposed. Three steps comprise RSM. The process in Step 1 involves calculating the root mean square error between the finite element and experimental load-displacement curves to produce the response surface. In step 2, the response surface is fitted with a second-order polynomial using the Levenberg-Marquardt algorithm. In step 3, an optimization problem is solved by minimizing the fitted function to find the optimum cohesive zone parameters. Finally, the obtained input for MAT_213 and MAT_186 material models is validated by performing a quasi-isotropic tension test simulation.