Abstract
When monitoring complex engineering systems, sensors often measure mixtures of signals that are unique to individual components (component signals). However, isolating component signals directly from sensor signals can be a challenge. As an example, in vibration monitoring of a rotating machine, if different components generate vibration signals at similar frequencies, they cannot be distinguished using traditional spectrum analysis (non-inseparable). However, developing degradation signals from component signals is important to monitor the deterioration of crucial components and to predict their residual lifetimes. This article proposes a simultaneous signal separation and prognostics framework for multi-component systems with non-inseparable component signals. In the signal separation stage, an Independent Component Analysis (ICA) algorithm is used to isolate component signals from mixed sensor signals, and an online amplitude recovery procedure is used to recover amplitude information that is lost after applying the ICA. In the prognostics stage, an adaptive prognostics method to model component degradation signals as continuous stochastic processes is used to predict the residual lifetimes of individual components. A case study is presented that investigates the performance of the signal separation stage and that of the final residual-life prediction under different conditions. The simulation results show a reasonable robustness of the methodology.
Additional information
Notes on contributors
Li Hao
Li Hao received a B.S. degree in Automotive Engineering from Tsinghua University, Beijing, China, in 2009 and an M.S. degree in Statistics from the Georgia Institute of Technology in 2012. Currently, she is a Ph.D. student at the H. Milton Stewart School of Industrial and Systems Engineering, Georgia Institute of Technology. Her research interests are focused on the prognostics and degradation-based control of complex systems.
Nagi Gebraeel
Nagi Gebraeel received M.S. and Ph.D. degrees in Industrial Engineering from Purdue University, West Lafayette, Indiana, in 1998 and 2003, respectively. Currently, he is an Associate Professor at the H. Milton Stewart School of Industrial and Systems Engineering, Georgia Institute of Technology. His research focuses on improving the accuracy of predicting unexpected failures of engineering systems by leveraging sensor-based data streams. His major research interests are in the areas of degradation modeling and sensor-based prognostics, reliability engineering, and maintenance operations and logistics. He is a member of IIE and INFORMS.
Jianjun Shi
Jianjun Shi received B.S. and M.S. degrees in Electrical Engineering from the Beijing Institute of Technology, Beijing, China, in 1984 and 1987, respectively, and a Ph.D. degree in Mechanical Engineering from the University of Michigan, Ann Arbor, in 1992. Currently, he is the Carolyn J. Stewart Chair Professor in the H. Milton Stewart School of Industrial and Systems Engineering, Georgia Institute of Technology. His research interests include the fusion of advanced statistical and domain knowledge to develop methodologies for modeling, monitoring, diagnosis, and control of complex manufacturing systems. He is a Fellow of the Institute of Industrial Engineers, a Fellow of the American Society of Mechanical Engineers, a Fellow of The Institute for Operations Research and the Management Sciences, and a member of ASQ, SME, and ASA.