接触网系统各部件内部应力概率分布规律

采用有限元并结合改进算法的径向基神经网络RBF,判别接触网系统各部件内部应力的随机分布规律。以兰新线武嘉段接触网腕臂为算例,把安装角度、冰负载、承力索覆冰风负载、接触线覆冰风负载等参量作为基本变量,按各变量的分布规律随机地产生50组样本数据,通过有限元计算出每一组参数对应的内部最大应力,依据随机抽样规律抽取40组,按经验分布计算出对应的应力分布函数值。利用RBF进行模式识别得到应力符合威布尔分布,并用回归分析计算出正态分布、指数分布、威布尔分布、对数正态分布二个变量的相关系数。从相关系数和拟合图可以看出,应力符合威布尔分布最好,服从指数分布的可能性最小,RBF识别的结果与回归分析的结果是一致的。

The Probability Distribution Law of the Internal Stress on Each Components of Catenary System

Random distribution law of internal stress on catenary parts was judged by combining finite element and improved arithmetic Radial Basis Neural Networks(RBFNN).Taking catenary wrist arm of Lanxin railway line Wujia section as an example and taking setting angle,ice load,ice and wind load of holding force rope,ice and wind load of catenary wire as main variable,fifty groups of random samples were obtained based on statistical law of each variable.The maximum stress of each group in the dangerous section was calculated by finite element.Function value of experience distributing was calculated by taking out forty groups based on random sample law.Weibull distribution law was acquired by RBF pattern recognition and correlation coefficient of normal distribution,logarithm normal distribution,Weibull distribution and exponential distribution were calculated.The result of correlation coefficient and fitting figure shows that stress distribution accords best with Weibull distribution and accords the least possibly with exponential distribution.Result of RBF is consistent with the result of recursive analysis.