作者：赵飞1, 宁静1,2, 方明宽1, 陈春俊1
作者单位：1. 西南交通大学机械工程学院，四川 成都 610031;
2. 轨道交通运维技术与装备四川省重点实验室，四川 成都 610031
关键词：高速列车;蛇行失稳;小幅蛇行;互信息分析;HHT
摘要：
近年来，人工智能技术广泛应用于高速列车运行状态的识别。用作列车运行状态识别的机器学习和深度学习模型需要大量的数据，然而在实测数据中获取的蛇行失稳数据样本少且不均衡，再者，由于实测数据获取成本高，实际采集数据困难，难以满足各种运行条件（稳定、小幅蛇行失稳和大幅蛇行失稳状态）。为解决以上问题，通过SIMPACK软件建立高速列车动力学仿真模型，模拟出车辆的各种运行状态数据。对仿真数据和实测数据进行互信息分析和希尔伯特-黄变换（HHT）分析，发现仿真数据在时域特征和频域特征上与实测数据高度相似。把仿真数据作为训练集，用来训练高速列车运行状态识别的人工智能算法模型，再用实测数据作为测试集进行验证。实验结果表明：仿真数据用于机器学习和深度学习模型中都能得到很好的结果，说明仿真数据可以作为训练集用于复杂的深度学习模型中。
Identification of high-speed trains hunting states based on dynamic simulation data
ZHAO Fei1, NING Jing1,2, FANG Mingkuan1, CHEN Chunjun1
1. School of Mechanical Engineering, Southwest Jiaotong University, Chengdu 610031, China;
2. Technology and Equipment of Rail Transit Operation and Maintenance Key Laboratory of Sichuan Province, Chengdu 610031, China
Abstract: In recent years, artificial intelligence technology has been widely used to identify the operation states of high-speed trains. Machine learning and deep learning models for trains operation states recognition need a large amount of data, but the hunting instability data samples obtained from the measured data are small and unbalanced. Moreover, due to the high cost of acquiring the measured data, it is difficult to collect the actual data, and it is difficult to meet various operation states (stability, small amplitude hunting instability and large amplitude hunting instability). In order to solve the above problems, this paper establishes the high-speed train dynamics simulation model through SIMPACK software, and simulates various operation states data of the vehicle. Through mutual information analysis and Hilbert Huang transform (HHT) analysis, it is found that the simulation data are highly similar to the measured data in time domain and frequency domain. The simulation data is used as training set to train the artificial intelligence algorithm model of operation states recognition of high-speed trains, and then the measured data are used as test set for verification. The experimental results show that the simulation data can be used in both machine learning and deep learning models, which shows that simulation data can be used as training set in complex deep learning models.
Keywords: high-speed trains;hunting instability;small amplitude hunting;mutual information analysis;HHT
2021, 47(1):120-126 收稿日期: 2020-09-08;收到修改稿日期: 2020-10-13
基金项目: 国家自然科学基金项目（51975486，51975487）；四川省科技计划资助（2020JDTD0012）
作者简介: 赵飞(1992-),男,四川达州市人,硕士研究生,专业方向为智能化状态监测及故障诊断
参考文献
[1] 王福天. 车辆系统动力学[M]. 北京: 中国铁道出版社, 1994.
[2] 孙丽霞. 高速列车横向运动稳定性和动态脱轨理论分析及评价方法研究[D]. 北京: 中国铁道科学研究院, 2014.
[3] 宋欣武. 转向架蛇行失稳监测的横向位移识别方法研究[D]. 成都: 西南交通大学, 2017.
[4] ZENG Y C, ZHANG W H, SONG D L. A new strategy for hunting alarm and stability evaluation for railway vehicles based on nonlinear dynamics analysis[J]. Proceedings of the Institution of Mechanical Engineers Part F Journal of Rail and Rapid Transit, 2020, 234(1): 54-64
[5] YE Y G, NING J. Small-amplitude hunting diagnosis method for high-speed trains based on the bogie frame’s lateral-longitudinal-vertical data fusion, independent mode function reconstruction and linear local tangent space alignment[J]. Proceedings of the Institution of Mechanical Engineers Part F Journal of Rail and Rapid Transit, 2019, 233(10): 1050-1067
[6] SUN J F, Meli E, CAI W B, et al. A signal analysis based hunting instability detection methodology for high-speed railway vehicles[J]. Vehicle System Dynamics, 2020(6): 1-23
[7] 崔万里, 宁静, 种传杰, 等. 基于EMD-ISOMAP高速列车小幅蛇行异常特征提取[J]. 中国测试, 2016, 42(12): 105-110
[8] 叶运广, 宁静, 种传杰, 等. 高速列车转向架蛇行失稳的MEEMD-LSSVM预测模型[J]. 铁道学报, 2018(1): 38-43
[9] 高学军. 铁道客车系统横向运动对称不对称分岔行为与混沌研究[D]. 成都: 西南交通大学, 2010.
[10] 董浩. 铁道车辆运动稳定性及分岔类型研究[D]. 成都: 西南交通大学, 2014.
[11] 王靖铭,宁静,赵飞, 等. 不同频率轨道激励下列车横向半主动控制研究[J]. 中国测试, 2020, 46(8): 116-120
[12] 翟婉明. 车辆-轨道耦合动力学(上册)[M]. 北京: 科学出版社, 2015.
[13] 邹祥, 黄炜. HHT以及边际谱在信号分析中的研究[C]//中国西部青年通信学术会议, 2009.
[14] 冉伟, 宁静, 陈杨, 等. 基于EEMD-SVD-LTSA的高速列车蛇行演变特征提取框架[J]. 电子测量技术, 2019, 42(5): 1-5
[15] 王晓东,宁静,陈春俊. 1D CNN和LSTM高速列车横向稳定性状态识别研究[J]. 中国测试, 2020, 46(11): 25-30
