基于先验知识的弓网接触电阻预测模型精度提升方法研究

doi:10.19595/j.cnki.1000-6753.tces.231065

摘要
图/表
参考文献
相关文章 (15)

全文: PDF (1473 KB)
输出: BibTeX | EndNote (RIS)

摘要

高速列车的运行实践表明,随着运行速度提高,受电弓与接触导线分离的可能性就越大,弓网电弧就越容易产生。导致电弧产生的因素很多,但最终都归结到滑板与接触网的接触电阻上面。通过滑动电接触实验机,研究了波动载荷、滑动速度和接触电流对接触电阻的影响,并结合表面形貌分析了磨损机制、电弧放电与接触电阻演变规律之间的关系。随后为了预测不同工况下的接触电阻,建立了RBF神经网络回归模型,通过在模型训练中融入先验知识和采用改进的食肉植物优化算法（ICPA）优化RBF神经网络超参数提升弓网接触电阻预测模型的精度。有、无先验知识的ICPA-RBF模型预测性能对比仿真结果表明,两类先验知识分别有助于提高模型的收敛速度和预测精度,最后采用假设检验的方法验证了模型的有效性。

	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
	时光
	陈翼喆
	李莹
	张国威

关键词 ：接触电阻, 先验知识, RBF神经网络, 食肉植物算法, 假设检验

Abstract：

High-speed trains acquire electrical energy through sliding contact between the pantograph and catenary, and good current-collection performance between the pantograph and the contact wire is a necessary condition for the safe and stable operation of the train. With the development of electric locomotives towards high-speed and heavy-load, the probability of electric arc generation increases. Numerous factors contribute to the generation of electric arcs, but these ultimately stem from the contact resistance between the slider and the contact wire. Therefore, it is very important to establish an accurate prediction model for contact resistance based on the analysis of its characteristics. However, in previous research, the analysis and modeling of contact resistance characteristics were isolated from each other. This paper attempts to effectively combine the two, applying the conclusions obtained from the mechanism analysis as prior knowledge to the modeling process of contact resistance, and exploring methods to further improve the accuracy of the contact resistance prediction model.
First, the universal conclusions in previous research on electrical contact were summarized: (1) As the contact pressure increases, the contact resistance decreases. (2) Temperature rising significantly affects contact material properties, wear mechanisms, and wear surface morphology, and can be considered as an independent variable. Considering the temperature of the contact core of the friction pair is not easy to measure and the conductive current loss has a close correlation with temperature rising, the conductive current loss as an independent variable can be used to replace the temperature rising. Subsequently, through the sliding electrical contact experiment machine, the effects of fluctuating load, sliding speed, and contact current on contact resistance were studied, and the relationship between wear mechanism, electric arc discharge, and contact resistance evolution law was analyzed in combination with surface morphology. The following conclusions were drawn: (1) As the fluctuating load increases, the stability of electrical contact weakens, and it mainly has a negative effect on contact resistance. (2) As the contact current increases, the current lubrication effect strengthens, which is conducive to the reduction of contact resistance. (3) The increase of speed weakens the stability of contact, the contact surface deteriorates, and the contact resistance increases. Finally, based on the aforementioned two conclusions, the partial derivative of contact resistance with respect to contact load is taken as a non-positive number, and the square of the loss current which is used to replace the temperature rising should be considered as an independent model input variable. They are called prior knowledge (I) and (II) respectively and applied to process of establishing the RBF network model of contact resistance. The improved Carnivorous Plant Optimization Algorithm (ICPA) is used to optimize the hyperparameters of RBF network to improve the accuracy of the contact resistance prediction model.
Simulation results indicate that prior knowledge (I) can confine the direction of weight adjustment of the RBF model, reduce the search space of weights, help to prevent overfitting, and enhance the convergence speed of the ICPA-RBF model, but it does not improve the final approximation accuracy. The MSE index and determination coefficient R² index of the RBF network without prior knowledge (II) are inferior to those of the RBF network with prior knowledge (II). The RBF network which adds the square of the loss current as an input variable considers the energy of electric arc discharge indirectly, thus providing a more comprehensive description of the contact surface state and stronger performance robustness of the prediction model. Finally, in order to further verify the generalization ability and effectiveness of the model, hypothesis testing is conducted under other working conditions. The results not only demonstrate the effectiveness of the contact resistance prediction model but also reflect the effectiveness of prior knowledge (II) in improving the generalization ability and enhancing the performance robustness of the model.

Key words： Contact resistance prior knowledge RBF neural networks carnivorous plant algorithm hypothesis testing

PACS:

TM501

基金资助:

国家自然科学基金（51477071）和辽宁省教育厅面上项目（LJKMZ20220675）资助

通讯作者: 陈翼喆男,1998年生,硕士研究生,研究方向为电接触理论及其应用、非线性系统。E-mail：779461819@qq.com

作者简介: 时光, 男,1981年生,博士,副教授,研究方向为电接触理论及其应用、非线性系统。E-mail：shiguang@cqu.edu.cn

引用本文:

时光, 陈翼喆, 李莹, 张国威. 基于先验知识的弓网接触电阻预测模型精度提升方法研究[J]. 电工技术学报, 0, (): 9035-35. Shi Guang, Chen Yizhe, Li Ying, Zhang Guowei. Research on Accuracy Improvement Method of Pantograph Contact Resistance Prediction Model Based on Prior Knowledge. Transactions of China Electrotechnical Society, 0, (): 9035-35.

链接本文:

https://dgjsxb.ces-transaction.com/CN/10.19595/j.cnki.1000-6753.tces.231065 https://dgjsxb.ces-transaction.com/CN/Y0/V/I/9035

[1] 张冬冬, 李小波, 张浩等. 地铁列车弓网接触电阻数学模型研究[J]. 智能计算机与应用, 2022, 12(02): 165-169.
Zhang Dongdong, Li Xiaobo, Zhang Hao, et al.Investigation on the mathematical model of contact resistance of pantograph-catenary in metro vehicle[J]. Intelligent Computer and Applications, 2022, 12(02): 165-169.
[2] 王英, 刘志刚, 范福强等. 弓网电弧模型及其电气特性的研究进展[J]. 铁道学报, 2013, 35(08): 35-43.
Wang Ying, Liu Zhigang, Fan Fuqiang, et al.Review of research development of pantograph-catenary arc model and electrical characteristics[J]. Journal of the China Railway Society, 2013, 35(8): 35-43.
[3] 吴积钦. 弓网系统电弧的产生及其影响[J]. 电气化铁道, 2008(02): 27-29.
Wu Jiqin.Occurrence of arc in pantograph and overhead contact system and its interference[J]. Electric Railway, 2008, 19(2): 27-29.
[4] Liu Jianqiang, Yang Qilin and Zheng T. Q. Harmonic analysis of traction networks based on the CRH380 series EMUs accident[C]//2012 IEEE Transportation Electrification Conference and Expo (ITEC), Dearborn, MI, USA, 2012: 1-6.
[5] 陈忠华, 李兵红, 陈明阳, 平宇, 郭凤仪. 弓网滑动电接触电磁热力耦合效应研究进展[J]. 电工技术学报, 2023, 38(10): 2777-2793.
Chen Zhonghua, Li Binghong, Chen Mingyang, Ping Yu, Guo Fengyi.Research advances in electrical-magnetic-thermal-mechanical coupling effects of electric contact between pantograph and catenary[J]. Transactions of China Electrotechnical Society, 2023, 38(10): 2777-2793.
[6] Wang Wangang, Dong Anping, Wu Guangning, et al.Study on characterization of electrial contact between pantograph and catenary[C]//2011 IEEE 57th Holm Conference on Electrical Contacts (Holm), Minneapolis, USA, 2011: 1-6.
[7] 郭凤仪, 马同立, 陈忠华, 等. 不同载流条件下滑动电接触特性[J]. 电工技术学报, 2009, 24(12): 18-23.
Guo Fengyi, Ma Tongli, Chen Zhonghua, et al.Characteristics of the sliding electric contact under different currents[J]. Transactions of China Electrotechnical Society, 2009, 24(12): 18-23.
[8] 王万岗, 梁明, 吴广宁, 等. 弓网系统静态接触电阻特性研究[J]. 高压电器, 2012, 48(1): 30-34.
Wang Wangang, Liang Mingliang, Wu Guangning, et al.Characteristics of static contact resistance between pantograph and catenary[J]. High Voltage Apparatus, 2012, 48(1): 30-34.
[9] 陈忠华, 石英龙, 时光, 等. 受电弓滑板与接触网导线接触电阻计算模型[J]. 电工技术学报, 2013, 28(5): 188-195.
Chen Zhonghua, Shi Yinglong, Shi Guang, et al.Calculation model of the contact resistance between pantograph slide and contact wire[J]. Transactions of China Electrotechnical Society, 2013, 28(5): 188-195.
[10] 李春茂, 朱宁俊, 吴广宁, 高国强, 吴杰. 弓网系统动态接触电阻数学模型的研究[J]. 高电压技术, 2015, 41(11): 3554-3560.
Li Chunmao, Zhu Ningjun, Wu Guangning, Gao Guoqiang, Wu Jie.Investigation on the mathematical model of dynamic contact resistance of pantograph-catenary system[J]. High Voltage Engineering, 2015, 41(11): 3554-3560.
[11] Watanabe Y.High-speed sliding characteristics of Cu-Sn-based composite materials containing lamellar solid lubricants by contact resistance studies[J]. Wear, 2008, 264(7-8): 624-631.
[12] 刘方林. 电气化铁路动态弓网接触电阻研究[J].电气技术, 2018, 19(09): 69-72.
Liu Fanglin.Study on dynamic contact resistance between pantograph and catenary in electrified railway[J]. Electrical Engineering, 2018, 19(09): 69-72.
[13] D Paulmier, M El Mansori, H Zaidi. Study of magnetized or electrical sliding contact of a steel XC48 graphite couple[J]. Wear, 1997, 203-204: 148-154.
[14] Z L Hu, Z H Chen, J T Xia.Study on surface film in the wear of electrographite brushes against copper commutators for variable current and humidity[J]. Wear, 2008, 264: 11-17.
[15] G R Hric and W G Odendaal. Improving Start-Up Contact Distribution Between Railgun Armature and Rails[C]//IEEE Transactions on Plasma Science, 2016, vol.44, no.7: 1202-1207.
[16] 陈忠华, 王一帆, 杨彩红, 时光, 回立川. 波动载荷下弓网动态接触电阻特性研究[J]. 高压电器, 2019, 55(10): 148-155+163.
Chen Zhonghua, Wang Yifan, Yang Caihong, Shi Guang, Hui Lichuan. Study on dynamic contact resistance of pantograph catenary under fluctuating load[J]. High Voltage Apparatus, 2019, 55(10): 148-155+163.
[17] 王英, 刘志刚, 黄可,等. 计及机械和电气特性的弓网表面热流分析和计算[J] .铁道学报, 2014, (7): 36-43.
Wang Ying, Liu Zhigang, Huang Ke, et al.Pantograph-catenary surface heat flow analysis and calculations based on mechanical and electrical characteristics[J]. Journal of the China Railway Society, 2014, (7): 36-43.
[18] T. Tamai and K. Tsuchiya. Contact resistance characteristics at low temperature[C]//IEEE Transactions on Components, Hybrids, and Manufacturing Technology. 1978, 1, 54-58.
[19] Liu Xinlong, Cai Zhenbing, He Jifan, et al. Effect of elevated temperature on fretting wear under electric contact[J]. Wear, 2017, 376-377: 643-655.
[20] 郭凤仪, 王喜利, 王智勇等. 弓网电弧辐射电场噪声实验研究[J]. 电工技术学报, 2015, 30(14): 220-225.
Guo Fengyi, Wang Xili, Wang Zhiyong, et al.Research on radiated electric field noise of pantograph arc[J]. Transactions of China Electrotechnical Society, 2015, 30(14): 220-225.
[21] Feng Chenzefang, Lin Xinxin, Xu Yixin and Zhu Fulong. Investigation on Contact Resistance of Connector Based on FEM[C]//2020 IEEE 22nd Electronics Packaging Technology Conference (EPTC) Singapore, Singapore, 2020: 403-405.
[22] 李白, 鲁军勇, 谭赛, 张永胜, 蔡喜元. 高速滑动电接触电枢表面动态磨损过程研究[J]. 电工技术学报, 2023, 38(01): 131-139.
Li Bai, Lu Junyong, Tan Sai, Zhang Yongsheng, Cai Xiyuan.Research on dynamic wear process of armature surface in high-speed sliding electric contact[J]. Transactions of China Electrotechnical Society, 2023, 38(01): 131-139.
[23] 王增基, 陈立学, 尤彭昊等. 考虑速度趋肤效应与接触电阻影响的枢轨界面电流分布特性[J]. 电工技术学报, 2022, 37(19): 5003-5010.
Wang Zengji, Chen Lixue, You Penghao, et al.Current distribution characteristics of armature-rail interface under velocity skin effect and contact resistance[J]. Transactions of China Electrotechnical Society, 2022, 37(19): 5003-5010.
[24] Boyer L.Contact resistance calculations: generalizations of Greenwood's formula including interface films[J]. IEEE Transactions on Components & Packaging Technologies, 2001, 24(1): 50-58.
[25] 时光, 陈忠华, 郭凤仪, 回立川, 党伟. 波动载荷下弓网接触电阻特性及建模研究[J]. 电工技术学报, 2019, 34(11): 2287-2295.
Shi Guang, Chen Zhonghua, Guo Fengyi, et al.Research on characteristic of the contact resistance of pantograph-catenary under load fluctuation condition[J]. Transactions of China Electrotechnical Society, 2019, 34(11): 2287-2295.
[26] 张鹏, 张峻霞, 刘瑞恒, Ahmed Mohamed Moneeb Elsabbagh. 基于GA-RBF神经网络和sEMG的下肢动作识别方法研究[J]. 中国生物医学工程学报, 2022, 41(01): 41-47.
Zhang Peng, Zhang Junxia, Liu Ruiheng, Ahmed Mohamed Moneeb Elsabbagh. The study of lower limb motion recognition method based on GA-RBF neural network and s EMG signals[J]. Chinese Journal of Biomedical Engineering, 2022, 41(01): 41-47.
[27] 张伟, 黄卫民. 基于SAPSO算法的RBF神经网络设计[J]. 控制与决策, 2021, 36(09): 2305-2312.
Zhang Wei, Huang Weimin.Design of RBF neural network based on SAPSO algorithm[J]. Control and Decision, 2021, 36(9): 2305-2312.
[28] 蒋慧平, 董霖. 钢铝复合轨/受电靴摩擦热与接触电阻热耦合的温度场有限元模拟研究[J]. 润滑与密封, 2010, 35(1): 45-48.
Jiang Huiping, Dong Lin.FE simulation research of temperature field couple under contact resistor-friction thermal between aluminum-stainless steel composite conductor rail and collector shoe[J]. Lubrication Engineering, 2010, 35(1): 45-48.
[29] Ong K M, Ong P, Sia C K, A carnivorous plant algorithm for solving global optimization problems[J]. Applied Soft Computing, 2020, 98(April): 106833.
[30] 田红亮, 朱大林, 秦虹玲. 两弹性接触粗糙低速滑动表面温升的分形模型[J]. 三峡大学学报, 2010, 32(2): 65-71.
Tian Hongliang, Zhu Dalin, Qin Hongling.Fractal model of temperature rise between two elastic contact rough slow sliding surfaces[J]. Journal of China Three Gorges University, 2010, 32(2): 65-71.
[31] 郭凤仪, 姜国强, 赵汝彬等.基于相对稳定系数的滑动电接触特性[J]. 中国电机工程学报, 2009, 29(36): 113-119.
Guo Fengyi, Jiang Guoqiang, Zhao Rubin, et al.Sliding electrical contact characteristics based on relative stability coefficients[J]. Proceedings of the CSEE, 2009, 29(36): 113-119.
[32] 黄鹤, 李文龙, 杨澜, 王会峰, 高涛, 陈婷. ICPA-LQR优化的两轮平衡机器人自稳定与轨迹跟踪PID控制器设计[J]. 哈尔滨工业大学学报: 1-12.
Huang He, LiWenlong, Yang Lan, Wang Huifeng, Gao Tao, Chen Ting. Design of PID controller for self-stabilization and trajectory tracking of two-wheel balanced robot with ICPA-LQR[J]. Journal of Harbin Institute of Technology: 1-12.
[33] Feng Junhong, Zhang Jie, Zhu Xiaoshu, et al.A novel chaos optimization algorithm[J]. Multimedia Tools and Applications, 2017, 76(16): 17405-17436.
[34] 杨虎, 刘琼荪, 钟波. 数理统计[M]. 北京: 高等教育出版社, 2004.