Continuum Compressive Damage Modelling in Composites Using Genetic Algorithms and Nonlocal Averaging.

Abstract

Continuum modelling of progressive damage in finite element analyses of fibre-reinforced polymers (FRPs) has become a popular tool because of its computational efficiency and ease of implementation. However, two of the major limitations are (i) mesh size and mesh orientation dependencies and (ii) the transparent determination of suitable input parameters. This study presents a combination of genetic algorithms (GA) with nonlocal continuum damage models to overcome these limitations. The use of GA provides an objective calibration process of input parameters, while nonlocal averaging of computed strain fields enables consistent damage evolution in FRPs irrespective of the underlying finite element mesh. The simulation of compact compression and open-hole compression tests on IM7/8552 carbon-fibre-reinforced polymers validates the calibration process and demonstrates the advantages of nonlocal damage modelling over conventional local approaches.