摘要
结合UIC615-4标准,对某转向架构架在各工况下进行了有限元强度分析,获得了最危险工况下构架危险点的应力响应.通过拉丁超立方抽样方法与试验设计方法获取构架危险点的应力与随机变量的仿真样本,并利用BP神经网络的非线性映射功能训练得到应力响应关于随机变量的显式表达式.结合随机摄动技术、一次二阶矩法及矩阵微分法,对构架的可靠性及可靠性灵敏度进行了分析,所得结果对转向架构架的工况和载荷优化具有一定的指导意义.
The strength analysis of a bogie frame under all load conditions was carried out according to the UIC615-4 code,and the stress response of the dangerous point under the most dangerous load condition was obtained.The simulation samples of random parameters and the stress of the dangerous point were obtained by the Latin hypercube sampling method and experiment design,and the explicit expression of the stress response about random variabes was obtained by the nonlinear mapping function training of a BP neurd network.The reliability and reliability sensitivity of the bogie frame were analyzed with the random perturbation technique,first order second moment method and matrix differential method.The work may provide guidance to the load conditions optimization.
出处
《东北大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2012年第5期711-714,共4页
Journal of Northeastern University(Natural Science)
基金
长江学者和创新团队发展计划项目(IRT0816)
国家科技重大专项(2010ZX04014-014)
国家自然科学基金资助项目(50875039)
国家"十一五"科技支撑计划项目(2009BAG12A02-A07-2)