Reliability Assessment Method of the System with Multiple Degradation Parameters Based on SVDD and SVR

Abstract:

This paper presents a reliability assessment method for the system with multiple degradation parameters based on SVDD and SVR. The method transforms different performance parameters of the system to a distance measure by SVDD, uses SVR to analyze and predict the distribution parameters of SVDD distance, then create the overall system performance degradation assessment model. Through vehicle tracking experiment, collecting different degradation parameters of the diesel engine, a reliability curve stands for the whole system degradation is achieved. Compared to traditional reliability assessment method, the new method combines the various parameters so that more effectively describes the diesel engine performance degradation, and is more instructive for the maintenance of the vehicle.

• Periodical: Applied Mechanics and Materials (Volumes 80-81)
• Pages: 1187-1192
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.80-81.1187
• Online since: July 2011
• Authors: Hua Cong, Guang Ping Wu, Fu Zhou Feng, Guo Qiang Rao
• Keywords: Multi-Parameter, Performance Degradation, Reliability, SVDD, SVR

References

[1] Ramirez J G, Gore W L and Johnston G. New Methods for Modeling Reliability Using Degradation Data. Statistics Data Analysis and Data Mining, 2001. 263-226.

[2] WANGPeng, COIT David. Reliability Prediction Based on Degradation Modeling for Systems with Multiple Degradation Measures [C]/Proceedings of the 2004 Reliability & Maintainability Symposium(RAMS). Los Angeles, CA, 2004: 302—307.

DOI: 10.1109/rams.2004.1285465

[3] Jun Fang, Xing Wei, Lixia Fan. Reliability Evaluation and Application Based on Multiple Performance Degradation [J]. Equipment Environmental Engineering: 2008, 5(5): 29-32. In Chinese.

[4] Fuzhi Dong, Yunchun Jiang, Baogang Gu. Reliability Evaluation Method for System with Multiple Performance Degradation Parameters based on Weight Allotment [J]. Missile Test Techonolygy: 2010, 1: 63-65. In Chinese.

[5] Vapnik N. The nature of statistical learning theory[M]. New York：Springer Press, (2000).

[6] Mei Yang, Yu Sun, Bo Kong. Research on Parameters Selection Method of SVR Model [J]. Computer Era: 2009, 11: 53-55. In Chinese.

[7] Mitchell T M. Machine Learning [M]. Mechanism Industry Press, Beijing, 2006: 166-167.

[8] Xing Jin, Yanji Hong. Assessment Methods of System Reliability[M]. National Defense Industry Press, Beijing. 2005: 33-74. In Chinese.