A novel parallel method for layup optimization of composite structures with ply drop-offs

Abstract

This paper presents a novel parallel optimization method for obtaining blended layups of composite laminates which closely match target lamination parameters. Firstly, a guide-based adaptive genetic algorithm (GAGA) which stochastically searches the layups is developed. Then the parallel optimization method DLBB-GAGA is developed by combining GAGA and a dummy layerwise branch and bound method (DLBB) which performs a logic-based search in a parallel computation, during which optimization information is shared between the two methods. The combination of these two different methods gives the parallel DLBB-GAGA method the advantages of both, enhancing the searching ability for the blended layup optimization. The superiority of the parallel DLBB-GAGA method is demonstrated through comparisons between the three methods, and it is concluded that the method is particularly effective for practical design where more layup design constraints are considered.