Abstract
Carbon fiber reinforced plastic (CFRP) laminated structures have been widely used in modern engineering due to their excellent material properties, especially in the aerospace and shipping industries. This requires a high-accuracy finite element model of CFRP laminated structures. However, it is difficult to master the mechanical properties of CFRP structures comprehensively and accurately due to influences from multiple aspects, such as dispersion of material properties, uncertainty of manufacturing technologies, etc. Therefore, a finite element model modification method of CFRP laminated structures based on correlation analysis and an approximate model was proposed. Aiming at minimizing the difference between the analysis model and the measured inherent frequency, the proposed method improves the finite element modeling accuracy of CFRP laminated structures, by iterative optimization based on a global optimization algorithm. In order to solve the problem of high spatial dimension and slow searching in modification of CFRP laminated structure models, the Pearson correlation analysis method was used to screen the material parameters which exert significant impacts on frequency characteristics to reconstruct the searching space. Based on significance parameters, an approximate response model of the CFRP laminated structural system was established. Meanwhile, the modeling accuracy of different orders of response surface models (RSM) and a radial basis function (RBF) neural network model was analyzed, and the best approximate modeling scheme was obtained. The approximate model was updated based on the multi-island genetic algorithm (MIGA) to modify the finite element model of the CFRP laminated structure model. The maximum error and mean error of the updated model are 1.47% and 0.45%. It was proved that the material parameters modified by the proposed method are applicable to simulation analysis of the CFRP laminated structure.