Optimization design for laminated composite structure based on kriging model

An effective optimization method is developed for laminated composite structures using two-level optimization strategy based on Kriging model and genetic algorithm (GA). Firstly, the design of experiment (DOE) technique is used to create sample points and MSC.Nastran is employed to obtain the response (minimum weight subjected to bulking and strength constraints) of each sample point. Based on sample points and the corresponding responses, the Kriging model is formulated. Secondly, GA is performed to obtain the best thickness by optimizing the Kriging model as objective function. Then, the best stacking sequence is obtained basing on lamination parameters using GA. This paper takes a Z shape composite stiffened plate as example to verify the feasibility of the method above. The results illustrate that it can significantly save computational costs and can greatly improve the optimization efficiency.