关键词: 复合材料耐压舱, 环肋加强区, 径向基神经网络近似模型, 遗传算法

Abstract: The composite material can greatly reduce the weight of the pressure hull, which has the great significance for increasing the payload of the underwater vehicle and improving its endurance. The composite material has the designable characteristic, so the ‘inverted T’-stiffened area of the pressure hull combined with the ribs is considered. The ultimate buckling pressure and the structure failure mode of the pressure hull are studied as the number of the composite material layers is changed. Taking the number of layers of the ring-stiffened area, the connection area and the coverage area as design variables, the sample points are selected in the design variable space by using the Latin hypercube design of experiment and conduct finite element structure analysis in order to obtain the response of each sample point. Combined with the multi-island genetic algorithm, the optimal design scheme of the composite ring-stiffened pressure hull is obtained by establishing the relationship between the number of layers and the ultimate buckling pressure, the Tsai-Wu failure index and the weight. The results show that the optimized pressure hull can meet the design requirements, improve the utilization ratio of structure material and reduce the optimization time.

Key words: composite pressure hull, ring-stiffened area, radial basis function neural network, genetic algorithm