| description abstract | Design optimization of composite structures is a challenging task due to the large dimensionality of the design space. In addition to the geometric variables (e.g., thickness of each component), the composite layup (the fiber orientation of each layer) also needs to be considered as design variables in optimization. However, the existing optimization methods are inefficient when applied to the multicomponent, multilayer composite structures. The low efficiency is caused by the high dimensionality of the design space and the inherent shortcomings in the existing design representation methods. In this work, two existing composite layup representation methods are investigated to discuss the root cause of the low efficiency. Furthermore, a new structural equation modeling (SEM)-based strategy is proposed to reduce the dimensionality of the design space. This strategy also helps the designers identify the loading mode of each component of the structural system. This strategy is tested in two scenarios of engineering optimization: (1) the direct multidisciplinary design optimization (DMDO), and (2) the metamodeling-based optimization. The proposed methods are compared with the traditional methods on two engineering design problems. It is observed that the design representation methods have a strong impact on the optimization results. The two case studies also demonstrate the effectiveness of the proposed strategy. Furthermore, recommendations are made on the selection of optimization methods for the design of composite structures. | |
