Deviation-based calibration for progressive damage analysis in pultruded glass fiber reinforced composites

This paper presents a systemic calibration methodology to efficiently simulate progressive damage evolution in four different pultruded glass fiber reinforced polymer (GFRP) composites using the strain-based COMposite DAMage Model (CODAM2) in the commercial finite element software LS-DYNA. In particular, Compact Tension (CT), scaled-up CT, and wide CT tests are simulated to find the best set of input parameters by considering four distinct indicators obtained from experimental and numerical load vs displacement data. By combining these indicators into a physically meaningful equivalent deviation value via a linear weighted-sum method, the results show that the most suited input damage variables yield physically accurate crack length predictions which underlines the robustness and accuracy of the proposed method. Furthermore, it is shown that the incorporation of bi-linear softening laws improves CODAM2 simulation results by up to 90%, however it also increases the number of parameters to be calibrated.