A new algorithm for prognostics using Subset Simulation

• Published in: Reliability engineering & system safety Vol. 168; pp. 189 - 199
• Main Authors: Chiachío, Manuel, Chiachío, Juan, Sankararaman, Shankar, Goebel, Kai, Andrews, John
• Format: Journal Article
• Language: English
• Published: Barking Elsevier Ltd 01.12.2017; Elsevier BV
• Subjects: Algorithms; Carbon fiber reinforced plastics; Carbon fibers; Computational efficiency; Computer applications; Computer simulation; Computing time; Crack propagation; Data processing; Fatigue failure; Predictions; Prognostics; Propagation; Rare events; Reliability engineering; Simulation; Stochastic modeling; Structural damage; Subset Simulation; Subset Simulation; Rare events; Stochastic modeling; Prognostics
• ISSN: 0951-8320; 1879-0836; 1879-0836
• DOI: 10.1016/j.ress.2017.05.042

•A new algorithm based on Subset Simulation is provided for general prognostics.•The Subset Simulation method is used to obtain efficiency for rare events.•A simulated example and a challenging case study are used to demonstrate its efficacy.•Discussion is provided through comparison with a standard prognostics algorithm. This work presents an efficient computational framework for prognostics by combining the particle filter-based prognostics principles with the technique of Subset Simulation, first developed in S.K. Au and J.L. Beck [Probabilistic Engrg. Mech., 16 (2001), pp. 263-277], which has been named PFP-SubSim. The idea behind PFP-SubSim algorithm is to split the multi-step-ahead predicted trajectories into multiple branches of selected samples at various stages of the process, which correspond to increasingly closer approximations of the critical threshold. Following theoretical development, discussion and an illustrative example to demonstrate its efficacy, we report on experience using the algorithm for making predictions for the end-of-life and remaining useful life in the challenging application of fatigue damage propagation of carbon-fibre composite coupons using structural health monitoring data. Results show that PFP-SubSim algorithm outperforms the traditional particle filter-based prognostics approach in terms of computational efficiency, while achieving the same, or better, measure of accuracy in the prognostics estimates. It is also shown that PFP-SubSim algorithm gets its highest efficiency when dealing with rare-event simulation.