基于BP神经网络的航空电子设备安装架结构优化  

Structure optimization of airborne equipment avionics rack based on BP neural network

作　　者：张敬东[1] 向国齐[1] 起雪梅[1] 殷国富[2] 

机构地区：[1]攀枝花学院交通与汽车工程学院,攀枝花617000 [2]四川大学制造科学与工程学院,成都610065

出　　处：《现代制造工程》2014年第6期125-130,共6页Modern Manufacturing Engineering

基　　金：四川省科技支撑计划资助项目(2011GZX0065)

摘　　要：航空机载设备的可靠性决定飞行器的稳定安全性能,其结构的优化设计是一个重要研究课题。建立了航空电子设备安装架的有限元动力学分析模型,根据正交试验设计方法选取设计参数样本空间,利用计算模态分析方法获得了每个样本的性能参数,采用神经网络和遗传算法相结合的优化方法,数值求解了优化命题。结果表明,优化后的结构具有更佳的动力学性能、质量更轻,验证了该优化方法的高效性、可行性。The reliability of airborne equipment decides the stability and safety performance of the vehicle ,the structure optimiza-tion design of airborne equipment is an important research subject .The finite element dynamic analysis model of airborne equip-ment avionics rack is established by using ANSYS software ,appropriate design parameter samples are selected by orthogonal ex-perimental design theory ,then the response samples are obtained from numerical simulations ,the BP neural network metamodel is constructed and treated as the objective performance functions ,the optimization model is formed and ready to be solved by GA . The results show that the optimized structure has better dynamic performance ,and this methodology is efficiency and feasibility .

关 键 词：神经网络 有限元法 计算模态 遗传算法 结构优化 

分 类 号：TH113.1[机械工程—机械设计及理论]