A PCA Based Fault Detection and Isolation Method for Train Locating System

Abstract:

Fault tolerance is crucial to the operating safety and performance of train locating system. Based on the requirements of reliability and safety for train locating, the fault characteristics of location measuring sensors are analyzed. Based on the structure of the train locating system, the fault-tolerant design of the system is given with the location filtering module for case, in which six fault detectors are employed to determine the configuration of the module. Then a PCA based fault detection and isolation method is proposed with Hawkins T2 statistics and the corresponding control limit. By dynamically adjusting the efficiency factors, fault could be detected and isolated as prior defined isolating strategies, and then the fault tolerant performance will be guaranteed. Simulation results demonstrate the high fault tolerant ability of the proposed approach and certain practical application value.

• Periodical: Applied Mechanics and Materials (Volumes 20-23)
• Pages: 688-693
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.20-23.688
• Online since: January 2010
• Authors: Jiang Liu, Bai Gen Cai, Tao Tang, Jian Wang
• Keywords: Fault Detection, Fault Isolation, Multi-Sensor Fusion, Principal Component Analysis (PCA), Train Locating

References

[1] A Polivka, A Filip. Satellite-Based Positioning for CBTC[C]. Proceeding of second international conference Reliability, safety and diagnostics of transport structures and means 2005, University of Pardubice, Czech Republic, 7-8 July (2005).

[2] Petr Ernest, Roman Mazl, Libor Preucil. Train Locator Using Inertial Sensors and Odometer[C]. Proceedings of IEEE Intelligent Vehicles Symposium 2004, June 14-17, 2004: 860-865.

DOI: 10.1109/ivs.2004.1336497

[3] Roman Mazl, Libor Preucil. Sensor data fusion for inertial navigation of trains in GPS-dark areas[C]. Proceeding of the IEEE Intelligent Vehicles Symposium 2003, June 2003: 345-350.

DOI: 10.1109/ivs.2003.1212934

[4] Antonella Albanese, Livio Marradi. The RUNE project: The Integrity Performances of GNSS-Based Railway User Navigation Equipment[C]. Proceedings of Joint Rail Conference 2005, Pueblo, Colorado, March 16-18, 2005: 211-218.

DOI: 10.1109/rrcon.2005.186082

[5] Yang Bo, Gao She-sheng et al. New Method for Fault Detection and Isolation in IMU/GPS Integrated Navigation System for Vehicle[J]. Systems Engineering, 2005, 23(2): 112-116.

[6] S Kiriczi, E Schnieder. Possibilities of Failure Detection and Identification (FDI) in a Train Localization System. Proceedings of the ASME/IEEE Joint Railroad Conference, 1994: 141-145.

DOI: 10.1109/rrcon.1994.289032

[7] Zhou Dong-hua, Ye Zhong-yin. Modern Fault Diagnosis and Fault-tolerant Control[M]. Beijing: Tsinghua University Press, 2000: 4-10, 46-56.

[8] Fu Xiao, Shengwei Wang, Xinhua Xu, Gaoming Ge. An isolation enhanced PCA method with expert-based multivariate decoupling for sensor FDD in air-conditioning systems [J]. Applied Thermal Engineering, Volume 29, Issue 4, March 2009: 712-72.

DOI: 10.1016/j.applthermaleng.2008.03.046

[9] Huang Xiao-bin, Niu Zheng, Niu Yu-guang et al. Behavior Analysis of Sensor Fault Detection Using PCA Approach[J]. Chinese Journal of Sensors and Actuators, 2003, 4: 419-423.

[10] Qiu Tian, Ding Yan-jun, Wu Zhan-song. Sensor fault detection statistics based on principal component analysis[J]. Journal of Tsinghua University (Sci & Tech), 2006, 46(8): 1447-1450.

[11] Wang Hai-qing, Yu Shi-ming. Selection of Number of Principal Conponents based on Fault Diagnosis Performance Optimization[J]. Journal of Chemical Industry and Engineering, 2004, 55(2): 214-219.