基于遗传算法-BP神经网络的动车组列车轮对磨耗模型
高明亮1邵俊捷1常振臣1王连富1刘德权1牛振虎2陈之恒2
EMU Train Wheelset Abrasion Model Based on Genetic Algorithm-BP Neural Network
GAO MingliangSHAO JunjieCHANG ZhenchenWANG LianfuLIU DequanNIU ZhenhuCHEN Zhiheng
-
作者信息:1.中车长春轨道客车股份有限公司检修研发部， 130062, 长春；
2.西南交通大学牵引动力国家重点实验室， 610031， 成都
-
Affiliation:Overhaul R & D Department, CRRC Changchun Railway Vehicles Co., Ltd., 130062, Changchun, China
-
关键词:
-
Key words:
-
DOI:10.16037/j.1007-869x.2022.06.013
-
中图分类号/CLCN:U270.331+.1
-
栏目/Col:研究报告
摘要:
鉴于动车组运行过程中轮径磨耗及轮缘厚度磨耗对于列车的平稳、安全运行的重要影响，利用相关性算法确定了轮对磨耗的影响因素，并在传统BP神经网络的基础上采用GA(遗传)算法对其初始权重和阈值进行了优化，构建出GA-BP神经网络模型。输入某动车组列车的历史数据，对该模型进行训练，轮径磨耗预测准确率达到了95.29%，平均误差为0.212 mm；轮缘厚度磨耗预测准确率达到91.76%，平均误差为0.052 mm。证实了此模型在轮对磨耗预测方面的可用性。
Abstracts:
In view of the important influence of wheel abrasion and wheel rim thickness abrasion on smooth and safe operation of the train, wheelset abrasion influencing factors are determined by using correlation algorithm. By adopting GA (genetic algorithm) on the basis of the conventional BP neural network, the initial weights and thresholds of the factors are optimized, establishing the GA-BP neural network model. By imputing historical data of certain EMU train and training the model, wheel diameter abrasion prediction accuracy has reached 95.29%, and the average error is 0.212 mm; wheel rim thickness abrasion prediction accuracy has reached 91.76%, and the average error is 0.052 mm. Applicability of the model in the aspect of wheelset abrasion prediction is proved.
引文 / Ref:
高明亮, 邵俊捷, 常振臣. 基于遗传算法-BP神经网络的动车组列车轮对磨耗模型. 城市轨道交通研究, 2022, 25(6): 65.
GAO Mingliang, SHAO Junjie, CHANG Zhenchen. EMU train wheelset abrasion model based on genetic algorithm -BP neural network. Urban Mass Transit, 2022, 25(6): 65.
GAO Mingliang, SHAO Junjie, CHANG Zhenchen. EMU train wheelset abrasion model based on genetic algorithm -BP neural network. Urban Mass Transit, 2022, 25(6): 65.
- 上一篇: 高铁物流运输模式及其可行性
- 下一篇: 我国Ⅰ型大城市的城市轨道交通线网规划思考
过刊浏览
资料下载
友情链接