应用人工神经网络算法对固体火箭发动机复合材料壳体裙连接区结构的5个参数进行了研究。文中以发动机壳体在400t轴压载荷下裙外缠绕层最大纤维主应变作为目标函数。基于计算机数值仿真技术的实验设计(DOE)对5个参数进行了优化设计。利用另一组27个测试样本集对网络的有效性进行了校核。最后,获得了目标函数随设计变量的曲线关系,并对参数进行了灵敏度分析。

Five parameters of lap structure between skirt and composites material cases of SRM were investigated using Artificial Neural Network algorithm.The objective function employed in the present work was the max strain along the fiber in the fibrous composites layers which outside the skirt under the case of SRM bearing the 400 t thrust loading.Simulation-based of design of experiment (DOE) was applied to simulate the five parameters.A sample of 27 cases, different from the previous ones, was used to validate the ANN.Finally, the objective function vs.design variables curves were plotted, and the parameters sensitivity analysis were carried out.