An A-STAR algorithm for thick composite laminate design

Abstract

This paper describes a new approach for the design of optimum laminate stacking sequences. This technique uses an incremental procedure that combines implicit ternary decision trees with an informed search method based on an A-STAR algorithm capable of providing highly accurate and efficient solutions in terms of laminate computations. The method is especially suitable for thick composite laminate designs. Unlike other A-STAR implementations, in this case the main challenge lies in defining the heuristic function since the goal node is not known beforehand (blind search). To establish both the accuracy (minimum number of plies) and the efficiency of the new method, the technique was validated using a battery of Monte-Carlo simulations for the design of carbon fiber reinforced plastic plates subjected to a generic load case scenario and to some manufacturing constraints. This novel technology was then applied to the design of thick laminates, specifically in order to obtain the optimal solution (minimum laminate thickness) using the least number of optimization iterations.