电工技术学报  2021, Vol. 36 Issue (zk1): 31-39    DOI: 10.19595/j.cnki.1000-6753.tces.L90002
电工理论与新技术 |
一种用于钢轨电位评估的回归建模方法
张征, 杨少兵, 吴命利, 叶晶晶
北京交通大学电气工程学院 北京 100044
A Method of Regression Modeling for Rail Potential Evaluation
Zhang Zheng, Yang Shaobing, Wu Mingli, Ye Jingjing
School of Electrical Engineering Beijing Jiaotong University Beijing 100044 China
全文: PDF (1142 KB)   HTML
输出: BibTeX | EndNote (RIS)      
摘要 钢轨电位是地铁安全运行指标之一,评估钢轨电位水平一直是供电仿真计算中的一项难题,传统的供电仿真算法在定量计算钢轨电位时存在较大偏差,无法应用至实际行车组织中。该文首先建立钢轨回流系统的等效电路模型,通过分析供电系统运行特点,将钢轨电位数学模型分解为常量参数与变量因子,进而给出钢轨电位与变量因子间的关联关系式。根据所得关联关系式,分析钢轨电位回归建模方法的可行性与适用性,并基于变电所实测数据及BP神经网络算法,建立变电所的钢轨电位回归模型,结果表明,模型误差满足实际工程需要。最后给出模型在运营场景时的实际应用示例,示例体现出模型能够很好地融入现有的供电仿真系统中。
服务
把本文推荐给朋友
加入我的书架
加入引用管理器
E-mail Alert
RSS
作者相关文章
张征
杨少兵
吴命利
叶晶晶
关键词 钢轨电位非线性参数回归分析BP神经网络    
Abstract:Rail potential is one of the safety operation indexes of metro. The evaluation of rail potential level has always been a difficult problem in calculation of power supply simulation. There exists a large error when apply the traditional algorithm in power supply simulation quantitatively. Therefore, it cannot be applied to the actual train timetable. Firstly, this paper established the equivalent diagram model for rail reflux system; then through analyzing the characteristics of power supply system, the mathematical model of rail potential was divided into constant parameters and variable factors, and furthermore the relation between rail potential and variable factors was given. Based on the relation, the feasibility and applicability of regression modeling for rail potential was analyzed. The regression model of rail potential of substation was established with actual measured data and BP neural network algorithm. The result shows that error of model meets the needs of practical engineering. In addition, a practical application example of the model in operation scenario is given, which shows that the model can be well embedded in current system of power supply simulation system.
Key wordsRail potential    nonlinear parameters    regression analysis    BP neural network   
收稿日期: 2020-02-28     
PACS: TM922  
基金资助:中车南京蒲镇车辆有限公司技术开发资助项目(2020K115)
通讯作者: 杨少兵 男,1972年生,教授,博士生导师,研究方向为电力系统仿真。E-mail: shbyang@bjtu.edu.cn   
作者简介: 张 征 男,1997年生,硕士研究生,研究方向为牵引供电仿真。E-mail: 18121548@bjtu.edu.cn
引用本文:   
张征, 杨少兵, 吴命利, 叶晶晶. 一种用于钢轨电位评估的回归建模方法[J]. 电工技术学报, 2021, 36(zk1): 31-39. Zhang Zheng, Yang Shaobing, Wu Mingli, Ye Jingjing. A Method of Regression Modeling for Rail Potential Evaluation. Transactions of China Electrotechnical Society, 2021, 36(zk1): 31-39.
链接本文:  
https://dgjsxb.ces-transaction.com/CN/10.19595/j.cnki.1000-6753.tces.L90002          https://dgjsxb.ces-transaction.com/CN/Y2021/V36/Izk1/31