Research on Engine Remaining Useful Life Prediction Based on Oil Spectrum Analysis and Particle Filtering
SUN Lei1, JIA Yun-xian1, CAI Li-ying2, LIN Guo-yu1, ZHAO Jin-song1, 3
1. Equipment Command and Management Department, Ordnance Engineering College, Shijiazhuang 050003, China 2. Ordnance Technique Research Institution, Shijiazhuang 050003, China 3. Military Transportation Academy, Tianjin 300161, China
Abstract：The spectrometric oil analysis(SOA) is an important technique for machine state monitoring, fault diagnosis and prognosis, and SOA based remaining useful life(RUL) prediction has an advantage of finding out the optimal maintenance strategy for machine system. Because the complexity of machine system, its health state degradation process can’t be simply characterized by linear model, while particle filtering(PF) possesses obvious advantages over traditional Kalman filtering for dealing non-linear and non-Gaussian system, the PF approach was applied to state forecasting by SOA, and the RUL prediction technique based on SOA and PF algorithm is proposed. In the prediction model, according to the estimating result of system’s posterior probability, its prior probability distribution is realized, and the multi-step ahead prediction model based on PF algorithm is established. Finally, the practical SOA data of some engine was analyzed and forecasted by the above method, and the forecasting result was compared with that of traditional Kalman filtering method. The result fully shows the superiority and effectivity of the new method.
Key words：Oil spectrometric analysis;Particle filtering;Engine;Remaining useful life prediction
孙 磊1, 贾云献1, 蔡丽影2, 林国语1, 赵劲松1,3 . 基于油液光谱分析和粒子滤波的发动机剩余寿命预测研究 [J]. 光谱学与光谱分析, 2013, 33(09): 2478-2482.
SUN Lei1, JIA Yun-xian1, CAI Li-ying2, LIN Guo-yu1, ZHAO Jin-song1, 3 . Research on Engine Remaining Useful Life Prediction Based on Oil Spectrum Analysis and Particle Filtering. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2013, 33(09): 2478-2482.
[1] Anderson D P. Proceedings of the International Conference on Condition Monitoring. 1987, Swansea: 222. [2] Wang W, Hussin B. Journal of the Operational Research Society, 2009, 60, 789. [3] Ray A, Tangirala S. IEEE Transactions on Control Systems Technology, 1996, 4(4): 443. [4] Swanson D C. Proceedings of IEEE Aerospace Conference. Big Sky, Mantana: IEEE, 2001(3): 2971. [5] Willett P K, Kirubarajan T. Proceedings of SPIE. Orlando: SPIE, 2002: 1. [6] Arulampalam M S, Maskell S, Gordon N, et al. IEEE Transactions on Signal Processing 2002, 50(3): 174. [7] Gordon N J, Salmond D J, Smith A F M. IEEE Proceedings F: Radar and Signal Processing, 1993, 140(2): 107. [8] Rao C R, Pathak P K，Koltchinskii V I. Journal of Statistical Planning and Inference，1997, 64(2): 257. [9] Marseguerra M, Zio E. LiLoLe-Verlag GmbH (Publ. Co. Ltd. ). 2002.
