基于遗传算法和BP神经网络的圆柱壳大开孔接管结构优化研究

应用神经网络模拟GB150-2011《压力容器》标准中圆柱壳径向接管开孔结构参数与强度关系,以开孔接管结构用材最省为优化目标,采用遗传算法对圆柱壳大开孔接管结构进行优化设计研究.经神经网络拟合后,标准算图数据信息归一于一个三层的BP网络并应用于遗传算法程序中,与反复查图和有限元优化算法相比,此优化算法更节省设计时间.经验证,优化后的结构符合强度要求,优化结果满足工程设计需要.

Optimization of cylindrical shell with large opening nozzle based on BP neural network and genetic algorithm

This paper, according to the national pressure vessel standard GB150-2011, simulates the relationship between the structure parameters and the strength of the cylindrical shell with radial opening by neural network, which aims to minimize the material of the structure by using the genetic algorithm so as to optimize the design of large opening cylinder with nozzle. After the simulation of the neural network, the standard rendering data normalized to a three layer BP network and applied to the genetic algorithm, and the optimization algorithm is saving time if compared to repeatedly reading design and ANSYS optimization design. The strength of the optimized structure is analyzed by ANSYS and the optimized resulting structure meets the requirements of engineering design. This method has certain guiding significance for the optimization design of structure parameters of the special pressure equipment.