列车轴温监测数据软测量方法

谢国; 张永艳; 上官安琪; 杜许龙; 黑新宏; 高橋聖; 望月寛

doi:10.19818/j.cnki.1671-1637.2018.06.011

列车轴温监测数据软测量方法

doi: 10.19818/j.cnki.1671-1637.2018.06.011

1.
西安理工大学陕西省复杂系统控制与智能信息处理重点实验室, 陕西西安 710048
2.
日本大学应用情报工学科, 千叶船桥 274-8501

基金项目:

国家重点研究发展计划项目 2017YFB1201500

国家自然科学基金项目 61873201

国家自然科学基金项目 U1534208

国家自然科学基金项目 61773313

陕西省重点研究发展计划项目 2018GY-139

详细信息

作者简介:
谢国(1982-), 男, 湖北宜昌人, 西安理工大学教授, 工学博士, 从事轨道交通系统研究

中图分类号: U270.11
计量
- 文章访问数: 551
- HTML全文浏览量: 96
- PDF下载量: 508
- 被引次数: 0
出版历程
- 收稿日期: 2018-06-11
- 刊出日期: 2018-12-25

Soft measurement method for temperature monitoring data of train axle

1.
Shaanxi Key Laboratory of Complex System Control and Intelligent Information Processing, Xi'an University of Technology, Xi'an 710048, Shaanxi, China
2.
Department of Computer Engineering, Nihon University, Funabashi 274-8501, Chiba, Japan

More Information

Author Bio:
XIE Guo(1982-), male, professor, PhD, guoxie@xaut.edu.cn.

摘要

摘要: 为解决监测数据缺失导致的轴温监测系统误诊和漏诊率较高的问题, 提出了一种基于数据特征分析的轴温监测数据软测量方法; 通过轴温监测点的布局与相关性分析, 确定了监测数据软测量的源数据范围; 采用自组织特征映射算法, 通过对源数据归一化、优胜区域定义与隶属度优化, 实现了轴温数据本征维数确定与数据聚类; 引入多维尺度分析方法, 通过数据间距的相似性量化与距离矩阵特征值分解, 实现了轴温数据的类内降维; 采用多维尺度分析方法对类间降维数据再次降维, 提出了一种分步式降维方法, 构建了信息量最大化与计算量最小化的平衡策略; 采用深度学习栈式自编码器方法提取类间降维数据的内部特征, 构建了缺失轴温数据的软测量模型。研究结果表明: 基于降维数据的软测量方法的时间效率比基于原始数据的软测量方法高14.25%;2种方法的精度相当, 当一维数据缺失时, 数据软测量的平均精度可达99.83%;当二维数据缺失时, 平均精度可达99.75%;当三或四维数据缺失时, 平均精度均可达99.16%;在满足最大允许误差2.5%、误差容忍度1.0%条件的情况下, 针对任意缺失维度不高于四维的情况, 提出的方法可有效地实现高精度与高效率的缺失数据恢复。
- 高速列车 /
- 软测量 /
- 轴温监测 /
- 分步式降维 /
- 聚类 /
- 深度学习
Abstract: To solve the high misdiagnosis and missing report rates of existing axle temperature monitoring system caused by data missing, a soft measurement method based on the characteristic analysis of axle temperature monitoring data was proposed.By analyzing the layout and correlation of axle temperature monitoring points, the source data range of soft measurement for monitoring data was determined.By using the self-organizing feature mapping algorithm, the intrinsic dimension of axle temperature data was determined.The data were clustered through source data normalization, winning region definition and membership degree optimization.By introducing the multi-dimensional scaling analysis method, the dimension reduction of axle temperature data in the internal class was realized through data space quantization and similarityas well as eigenvalue decomposition of the distance matrix. The multi-dimensional scaling analysis method was reused to reduce the dimension of data between external classes.A step-bystep dimension reduction method was proposed, and an equilibrium strategy was built to maximize the amount of information and minimize the calculations.The stacked auto encoder of a deep learning method was utilized to extract the internal characteristics of data between external classes, and a soft measurement model of axle temperature missing data was constructed.Research result shows that the time efficiency of the soft measurement method based on the dimension reduced data is 14.25% higher than that based on original data.The accuracies of the two methods maintain a similar level.When one-dimensional data is missed, the average accuracy of soft measurement data can reach 99.83%.When two-dimensional data is missed, the average accuracy can achieve 99.75%.When three-or four-dimensional data is missed, the average accuracy can arrive at 99.16%.Under the conditions that the maximum allowable error and error tolerance are 2.5% and 1.0%, respectively, as long as the dimension of missing data is not higher than four, the method can effectively recover the missing data with high accuracy and efficiency.
- high-speed train /
- soft measurement /
- axle temperature monitoring /
- step-by-step dimension reduction /
- clustering /
- deep learning

HTML全文

图 1 高速列车结构

Figure 1. Structure of high-speed train

下载: 全尺寸图片幻灯片

图 2 轴温软测量流程

Figure 2. Flow of soft measurement of axle temperature

下载: 全尺寸图片幻灯片

图 3 SOFM算法流程

Figure 3. Flow of SOFM algorithm

下载: 全尺寸图片幻灯片

图 4 轴温数据类别统计

Figure 4. Class statistics of axle temperature data

下载: 全尺寸图片幻灯片

图 5 类间降维后的维数概率

Figure 5. Dimension percentages after reduction dimension between external classes

下载: 全尺寸图片幻灯片

图 6 不同训练数据比例下的平均相对误差

Figure 6. Average relative errors with different ratios of training data

下载: 全尺寸图片幻灯片

图 7 TC转向架1的二轴轴温估计

Figure 7. Temperature estimation of second axle of TC bogie 1

下载: 全尺寸图片幻灯片

图 8 BC转向架2的四轴轴温估计

Figure 8. Temperature estimation of fourth axle of BC bogie 2

下载: 全尺寸图片幻灯片

图 9 平均相对误差对比

Figure 9. Comparison of average relative errors

下载: 全尺寸图片幻灯片

图 10 估计时间的频次对比

Figure 10. Frequency comparison of estimation times

下载: 全尺寸图片幻灯片

图 11 两种方法的估计时间异常

Figure 11. Prediction time abnormals of two methods

下载: 全尺寸图片幻灯片

图 12 两种方法的相对误差异常

Figure 12. Relative error abnormals of two methods

下载: 全尺寸图片幻灯片

表 1 一维缺失数据的降维效果评价

Table 1. Performance assessment of dimensionality reduction with one-dimensional data missing

下载: 导出CSV

表 2 多维缺失数据软测量的平均相对误差

Table 2. Average relative errors of soft measurement with multi-dimensional data missing %

下载: 导出CSV

表 3 多维缺失数据的最大无效软测量比例

Table 3. Maximum invalid soft measurement ratios with multi-dimensional data missing %

下载: 导出CSV

参考文献(30)

[1]	哈大雷, 王乾, 蒋涛, 等. 新型轴温监测系统在高速动车组上的应用[J]. 大连交通大学学报, 2013, 34 (1): 89-94. doi: 10.3969/j.issn.1673-9590.2013.01.019 HA Da-lei, WANG Qian, JIANG Tao, et al. Application of new axle temperature monitoring system in high speed EMU[J]. Journal of Dalian Jiaotong University, 2013, 34 (1): 89-94. (in Chinese). doi: 10.3969/j.issn.1673-9590.2013.01.019
[2]	韩烨, 刘志刚, 耿肖. 基于HOG特征与二维Gabor小波变换的高铁接触网支撑装置耳片断裂故障检测[J]. 铁道学报, 2017, 39 (2): 52-57. doi: 10.3969/j.issn.1001-8360.2017.02.008 HAN YE, LIU Zhi-gang, GENG Xiao, et al. Fracture detection of ear pieces in catenary support devices of highspeed railway based on HOG features and two-dimensional Gabor transform[J]. Journal of the China Railway Society, 2017, 39 (2): 52-57. (in Chinese). doi: 10.3969/j.issn.1001-8360.2017.02.008
[3]	CAO Yuan, MA Lian-chuan, ZHANG Yu-zhuo. Application of fuzzy predictive control technology in automatic train operation[J]. Cluster Computing, 2018 (1): 1-10.
[4]	CAO Yuan, MA Lian-chuan, XIAO Shuo, et al. Standard analysis for transfer delay in CTCS-3[J]. Chinese Journal of Electronics, 2017, 26 (5): 1057-1063. doi: 10.1049/cje.2017.08.024
[5]	文成林, 吕菲亚, 包哲静, 等. 基于数据驱动的微小故障诊断方法综述[J]. 自动化学报, 2016, 42 (9): 1285-1299. https://www.cnki.com.cn/Article/CJFDTOTAL-MOTO201609001.htm WEN Cheng-lin, LYU Fei-ya, BAO Zhe-jin, et al. A review of data driven-based incipient fault diagnosis[J]. Acta Automatica Sinica, 2016, 42 (9): 1285-1299. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-MOTO201609001.htm
[6]	KIM T Y, YOON S J, PARK Y L. Soft inflatable sensing modules for safe and interactive robots[J]. IEEE Robotics and Automation Letters, 2018, 3 (4): 3216-3223. doi: 10.1109/LRA.2018.2850971
[7]	DWIVWDI A, RAMAKRISHNAN A, REDDY A, et al. Design, modeling, and validation of a soft magnetic 3-D force sensor[J]. IEEE Sensors Journal, 2018, 18 (9): 3852-3863. doi: 10.1109/JSEN.2018.2814839
[8]	DIN S, XU W L, CHENG L K, et al. A stretchable multimodal sensor for soft robotic applications[J]. IEEE Sensors Journal, 2017, 17 (17): 5678-5686. doi: 10.1109/JSEN.2017.2726099
[9]	吴链. 三维插值方法在湖南省气温精细化预报中的应用与检验[J]. 气象与减灾研究, 2017, 40 (1): 30-35. https://www.cnki.com.cn/Article/CJFDTOTAL-HXQO201701005.htm WU Lian. Application and verification of a three-dimensional interpolation method for refined temperature forecast in Hunan Province[J]. Meteorology and Disaster Reduction Research, 2017, 40 (1): 30-35. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-HXQO201701005.htm
[10]	MA Wei-gang, TAN Si-yu, HEI Xin-hong, et al. A prediction method based on stepwise regression analysis for train axle temperature[C]//IEEE. The 12th International Conference on Computational Intelligence and Security. New York: IEEE, 2017: 386-390.
[11]	LIU Qi-yang. High-speed train axle temperature monitoring system based on switched ethernet[J]. Procedia Computer Science, 2017, 107: 70-74. doi: 10.1016/j.procs.2017.03.058
[12]	张雨晨. 基于改进的SVM和t-SNE高速列车走行部故障诊断[D]. 成都: 西南交通大学, 2016. ZHANG Yu-chen. Research on the high-speed train running gear fault diagnosis with improved SVM and t-SNE[D]. Chengdu: Southwest Jiaotong University, 2016. (in Chinese).
[13]	侯贺. 缺失数据处理方法的研究及其在软测量技术中的应用[D]. 沈阳: 东北大学, 2011. HOU He. Research on missing data and application in softsensing[D]. Shenyang: Northeastern University, 2011. (in Chinese).
[14]	张煜东, 霍元铠, 吴乐南, 等. 降维技术与方法综述[J]. 四川兵工学报, 2010, 31 (10): 1-7. https://www.cnki.com.cn/Article/CJFDTOTAL-CUXI201010001.htm ZHANG Yu-dong, HUO Yuan-kai, WU Le-nan, et al. The summary of dimension reduction techniques and methods[J]. Journal of Sichuan Ordnance, 2010, 31 (10): 1-7. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-CUXI201010001.htm
[15]	POCHET N, DE SMET F, SUYKENS J A K, et al. Systematic benchmarking of microarray data classification: assessing the role of non-linearity and dimensionality reduction[J]. Bioinformatics, 2004, 20 (17): 3185-3195. doi: 10.1093/bioinformatics/bth383
[16]	SOLBERGT R, SONESSON A K, WOOLLIAMS J A, et al. Reducing dimensionality for prediction of genome-wide breeding values[J]. Genetics Selection Evolution, 2009, 41 (1): 1-8. doi: 10.1186/1297-9686-41-1
[17]	SHI Ming-wang. Application of BP neural net work and PCA in prediction of plant diseases control[J]. Advanced Materials Research, 2011, 219/220: 742-745. doi: 10.4028/www.scientific.net/AMR.219-220.742
[18]	SHEN Y, STEVEN X D, NAKIE A, et al. On PCA-based fault diagnosis techniques[C]//IEEE. The 3th International Conference on Control and Fault-Tolerant Systems. New York: IEEE, 2010: 179-184.
[19]	LI Xue-mei, DING Li-xing, LYU Jin-hu, et al. A novel hybrid approach of KPCA and SVM for building cooling load prediction[C]//IEEE. The 3th International Conference on Knowledge Discovery and Data Mining. New York: IEEE, 2010: 522-526.
[20]	ZHANG Ying-wei, CHI Ma. Fault diagnosis of nonlinear processes using multiscale KPCA and multiscale KPLS[J]. Chemical Engineering Science, 2011, 66 (1): 64-72. doi: 10.1016/j.ces.2010.10.008
[21]	张宏达, 王晓丹, 徐海龙. 一种LDA与SVM混合的多类分类方法[J]. 控制与决策, 2009, 24 (11): 1723-1728. doi: 10.3321/j.issn:1001-0920.2009.11.024 ZHANG Hond-da, WANG Xiao-dan, XU Hai-long. Hybrid multi-class classification approach based on LDA and SVM[J]. Control and Decision, 2009, 24 (11): 1723-1728. (in Chinese). doi: 10.3321/j.issn:1001-0920.2009.11.024
[22]	闫德勤, 刘胜兰, 李燕燕, 等. 一种基于稀疏嵌入分析的降维方法[J]. 自动化学报, 2011, 37 (11): 1306-1312. https://www.cnki.com.cn/Article/CJFDTOTAL-MOTO201111006.htm YAN De-qin, LIU Sheng-lan, LI Yan-yan, et al. An embedding dimension reduction algorithm based on sparse analysis[J]. Acta Automatica Sinica, 2011, 37 (11): 1306-1312. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-MOTO201111006.htm
[23]	唐朝辉, 朱清新, 洪朝群, 等. 基于自编码器及超图学习的多标签特征提取[J]. 自动化学报, 2016, 42 (7): 1014-1021. https://www.cnki.com.cn/Article/CJFDTOTAL-MOTO201607004.htm TANG Chao-hui, ZHU Qing-xin, HONG Chao-qun, et al. Multi-label feature selection with autoencoders and hypergraph learning[J]. Acta Automatica Sinica, 2016, 42 (7): 1014-1021. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-MOTO201607004.htm
[24]	吴晓婷, 闫德勤. 数据降维方法分析与研究[J]. 计算机应用研究, 2009, 26 (8): 2832-2835. doi: 10.3969/j.issn.1001-3695.2009.08.008 WU Xiao-ting, YAN De-qin. Analysis and research on method of data dimensionality reduction[J]. Application Research of Computers, 2009, 26 (8): 2832-2835. (in Chinese). doi: 10.3969/j.issn.1001-3695.2009.08.008
[25]	高宇. 基于SOM神经网络的风电电子装置故障诊断[J]. 电力系统及其自动化学报, 2010, 22 (3): 142-145. https://www.cnki.com.cn/Article/CJFDTOTAL-DLZD201003028.htm GAO Yu. Fault diagnosis of wind turbine power electronic devices based on SOM neural network[J]. Proceedings of the CSU-EPSA, 2010, 22 (3): 142-145. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-DLZD201003028.htm
[26]	胡友涛, 胡昌华, 孔祥玉, 等. 基于WSVR和FCM聚类的实时寿命预测方法[J]. 自动化学报, 2012, 38 (3): 331-340. https://www.cnki.com.cn/Article/CJFDTOTAL-MOTO201203004.htm HU You-tao, HU Chang-hua, KONG Xiang-yu, et al. Realtime lifetime prediction method based on wavelet support vector regression and fuzzy c-means clustering[J]. Acta Automatica Sinica, 2012, 38 (3): 331-340. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-MOTO201203004.htm
[27]	ZHAO Hua-sheng, JIN Long, HUANG Ying, et al. An application of ensemble prediction for typhoon intensity based on MDS and PSO-ANN[C]//IEEE. Proceedings of the 5th International Joint Conference on Computational Sciences and Optimization. New York: IEEE, 2012: 885-888.
[28]	WEI He-wen, PENG Rong, WAN Qun, et al. Multidimensional scaling analysis for passive moving target localization with TDOA and FDOA measurements[J]. IEEE Transactions on Signal Processing, 2010, 58 (3): 1677-1688.
[29]	ZHANG Yan, ZHANG Er-hu, CHEN Wan-jun. Deep neural network for halftone image classification based on sparse auto-encoder[J]. Engineering Applications of Artificial Intelligence, 2016, 50: 245-255.
[30]	SUN Wen-jun, SHAO Si-yu, ZHAO Rui, et al. A sparse auto-encoder-based deep neural network approach for induction motor faults classification[J]. Measurement, 2016, 89: 171-178.