动车组转向架构架的结构可靠性研究  

Research on structural reliability of bogie frame for electric multiple units trains

作　　者：孟建军[1,2,3] 赵文涛 MENG Jian-jun;ZHAO Wen-tao(Institute of Mechatronics Technology,Lanzhou Jiaotong University,Lanzhou 730070,China;Gansu Provincial Logistics and Transportation Equipment Informatization Engineering Technology Research Center,Lanzhou 730070,China;Gansu Logistics and Transportation Equipment Industry Technology Center,Lanzhou 730070,China)

机构地区：[1]兰州交通大学机电技术研究所,兰州730070 [2]甘肃省物流及运输装备信息化工程技术研究中心,兰州730070 [3]甘肃省物流与运输装备行业技术中心,兰州730070

出　　处：《安全与环境学报》2022年第2期649-656,共8页Journal of Safety and Environment

基　　金：国家自然科学基金项目(62063013,72061021)。

摘　　要：动车组转向架的关键部件构架是整个转向架的安装基础,为了验证构架的结构是否满足安全要求,保障转向架的安全稳定运行,提出了一种基于替代模型与主动学习策略的结构可靠性研究方法。分析构架承受的载荷及约束情况,找到不同工况中承受Von Mises应力最大的工况条件。采用前馈多层感知器网络估算转向架构架的极限状态函数,基于蒙特卡洛模拟和多基因遗传编程构造替代模型进行结构可靠性研究。为了提高模型的精度,采用改进的有效全局优化算法寻找次要点,扩充替代模型的训练集,进而对替代模型进行迭代更新。结果表明,在较危险的工况条件下,转向架构架可靠性为98.71%,满足安全要求。所提方法通过减少数值模型的调用次数大大减少了计算时间,通过与其他研究方法相比较验证了其具有较高的精确性,该方法适用于无法得到准确显式极限状态方程的复杂系统,以及需要进行快速可靠性计算的产品设计优化过程。The key component frame of the Electric Multiple Units Train(EMU) bogie is the installation foundation of the entire bogie, and its structural reliability research is related to the safe and stable operation of the whole vehicle. This article takes the CRH5 G high-cold and sand-resistant EMU bogie frame currently operating on the Lanxin line as the research object, analyzes its load and constraints, and finds the working condition that bears the largest Von Mises stress in different working conditions. Five random variables of the secondary suspension lateral load, secondary suspension vertical load, thrust baffle lateral load, elastic modulus, and Poisson’s ratio are selected as input variables. The extended algorithm of a neural network is used to feed forward the multi-layer perception network to estimate the limit state function of the bogie frame, and the structural reliability of the frame is studied based on Monte Carlo simulation and multi-gene genetic programming to construct an alternative model. To improve the accuracy of the model, an improved effective global optimization algorithm is used to find the minor points, and the training set of the replacement model is expanded to update the replacement model iteratively. The study shows that the reliability of the bogie frame under more dangerous working conditions is 98.71%, which meets the safety requirements. The structural reliability research method proposed in this paper effectively reduces the calculation time by reducing the number of calls to the numerical model. Compared with other structural reliability research methods, the proposed method in this paper is more accurate and can be applied to unavailable. Complicated systems with accurate and explicit limit state equations and product design optimization processes that require fast reliability calculations.

关 键 词：安全工程 转向架构架 可靠性 遗传编程 主动学习策略 蒙特卡洛模拟 

分 类 号：X951[环境科学与工程—安全科学]