Real-time update of multi-state system reliability using prognostic data-driven techniques

Multi-state system (MSS) reliability has been an active topic of research in the recent past, given its relevance to modeling power systems and other mechanical equipment which can be classified to operate at different states of performance (more than two) ranging from completely functional to partial degradation (wear-out) to system failure. The assessment of MSS involves the use of the Markov chain to estimate the transient and steady state probabilities given the transition rates between different states and a user-defined threshold demand from the system which defines reliability. In the case of systems which are not new and have been tested in the past (similar units operated before), the transition rates can be estimated from the historical records of the data. However, for new systems or those with minimal past data, the estimates of the state transition rates can be uncertain, which undermines the value of modeling systems from a multi-state perspective. This study proposes a particle filter (PF) based technique that capitalizes on real-time sensor-based degradation data of the system to estimate the future trend of wear-out and estimates the transition rates and their confidence bounds, which get updated every time a new data is measured. This study is aimed at bringing together the fields of MSS and prognostics and health management (PHM) to model and estimate the reliability of engineering systems more accurately and dynamically, moving away from the traditional practice of relying on historical failure data. The proposed technique would be handy for condition monitoring, lifetime estimation and maintenance scheduling of new systems or imperfect repair systems where prior knowledge of state transition rates is seldom available.

[1]  Bhaskar Saha,et al.  Prognostics Methods for Battery Health Monitoring Using a Bayesian Framework , 2009, IEEE Transactions on Instrumentation and Measurement.

[2]  Markos V. Koutras,et al.  On a Markov chain approach for the study of reliability structures , 1996 .

[3]  M. Xie,et al.  Models and Analysis , 2004 .

[4]  Eric Moulines,et al.  Comparison of resampling schemes for particle filtering , 2005, ISPA 2005. Proceedings of the 4th International Symposium on Image and Signal Processing and Analysis, 2005..

[5]  P. Moral,et al.  Sequential Monte Carlo samplers , 2002, cond-mat/0212648.

[6]  Enrico Zio,et al.  Particle filtering prognostic estimation of the remaining useful life of nonlinear components , 2011, Reliab. Eng. Syst. Saf..

[7]  Gregory Levitin,et al.  Multi-State System Reliability - Assessment, Optimization and Applications , 2003, Series on Quality, Reliability and Engineering Statistics.

[8]  Li Jing,et al.  Multi-State System Reliability: A New and Systematic Review , 2012 .

[9]  Cher Ming Tan,et al.  A framework to practical predictive maintenance modeling for multi-state systems , 2008, Reliab. Eng. Syst. Saf..

[10]  Gregory Levitin,et al.  Optimization of imperfect preventive maintenance for multi-state systems , 2000, Reliab. Eng. Syst. Saf..

[11]  Hoang Pham,et al.  Reliability modeling of multi-state degraded systems with multi-competing failures and random shocks , 2005, IEEE Trans. Reliab..

[12]  Enrico Zio,et al.  A multi-state model for the reliability assessment of a distributed generation system via universal generating function , 2012, Reliab. Eng. Syst. Saf..