Statistical analysis and modeling of wrinkle defects in continuous-fiber composite forming processes

Wrinkles are detrimental to the performance of composite components. The prediction of wrinkling during forming is challenging because of complex material interactions. Current computational modeling methods are capable of wrinkle prediction but only for one given set of parameters per simulation cycle. To improve the prediction of wrinkles, this article presents a theory that correlates wrinkle formation with factors such as composite layup and part dimension. This theory allows one to perform design of experiments for physical or virtual forming tests and obtain binomial probability distributions capable of wrinkle prediction. The theory is validated with forming experiment data and demonstrates its potential in wrinkle prediction.