厚壁复合材料管纤维缠绕张力的神经网络设计方法

Approach to design tension of filament winding for thick composite pipes using a neural network

提出了一种纤维缠绕厚壁复合材料管的张力优化设计方法。介绍了纤维缠绕控制系统的工作原理,并讨论了缠绕厚壁复合材料管成型质量的影响因素。针对厚壁复合材料管纤维缠绕过程,利用弹性叠加原理建立了计算缠绕张力导致复合材料管残余内应力变化的模型和方法。分别比较了利用现有恒张力、恒力矩和锥度张力三种常规模式缠绕厚壁复合材料管的内应力分布特性。设计了一种独特的神经网络结构,并通过误差反向传播实现了对纤维缠绕张力的优化设计。以实验验证说明神经网络收敛优化过程的主要机制,结果表明,通过该神经网络优化的纤维缠绕张力能满足特殊内应力(如等应力)分布设计的需要。

A novel design approach was presented to optimizing the winding tension for thick composite pipes in this paper.The theories of filament winding machine control system were introduced,and the factors that influence the forming quality of thick composite pipes were discussed.An approach of calculating the residual inner stress and deformation in tension winding for thick fiber composite pipes according to superposition principle of the elastic theory was advanced.The inner stress distributions in the thick pipes wound with constant tension,constant torque and taper tension patterns separately were compared.A peculiar neural network structure was presented for optimizing the winding tension with error back propagation.Experiments were conducted to identify dominant mechanisms during neural network convergence process.The results show that the winding tension optimized by this network can meet the design requirement with special inner stress distributions such as isostress.