基于PSO-BP神经网络的油船中部结构优化

摘要

The design variables are determined by sensitivity analysis. Then the optimum design of large oil tanker mid structure is carried out by taking hold section structure weight as the objective function, and taking rule's requirements of the plate thickness and stress as the constraint conditions. The BP neural network model based on particle swarm optimization is built, which is used to determine the relationship between stress and design variables in place of finite element analysis. The optimized structure weight decreased by 4.2%. The finite element analysis results show that the optimized structure is satis-fied with the requirements of the rule.The PSO-BP neural network model is feasible in the optimization design of the ship structure.%以舱段质量为目标函数,以相关规范要求的板厚及应力为约束条件,通过灵敏度分析确定设计变量,对油船中部结构优化.构建基于粒子群优化的BP神经网络模型,并代替有限元分析确定应力与设计变量之间关系,从而对舱段进行结构优化.优化后舱段质量降低了4.2%,优化后的有限元分析结果表明满足规范要求,PSO-BP神经网络模型在船舶结构优化设计中具有可行性.