Probabilistic failure simulation of glass fibre reinforced weft-knitted thermoplastics

If textile-reinforced composites should be put into practice, the proof of load-bearing capacity has to be provided. For this purpose, deterministic overall safety proofs without material adapted partial safety factors are used which do not meet the textile-specific scatter characteristics and therefore neglect many advantages of those composites. Based on deterministic stress-based failure criteria and given loading conditions, a probabilistic methodology to predict the joint failure probability is proposed. The methodology is verified experimentally for hybrid yarn based textile thermoplastics. A statistical material characterisation using flat specimens, multi-axial tests with tube specimens and a Monte-Carlo simulation are used for that purpose. As a significant outcome of the study, the reliability of the failure prediction is significantly improved especially for multi-axial stress states with failure mode interaction. Based on the improved knowledge about the anisotropic material uncertainties, a more significant partial safety factor can be determined.