March 25, 2015: Invited Speaker Seminar: Integrated Prognostics and Uncertainty Quantification for Equipment Health Management

Dr. Fuqiong Zhao

Wednesday, March 25, 2015 at 2:30 p.m.
Room EV003.309

Abstract

Reliability program has been well designed and integrated into the product realization process, unfortunately, unexpected failures still happen to engineering equipment. These unexpected failures tremendously impact the three key elements in production competitiveness: safety, cost and quality. In particular, the impact will be even more catastrophic in safety- critical applications, e.g., aerospace, process industry, nuclear industry.  The technology of prognostics and health management (PHM) aims at developing advanced algorithms to detect faults and to predict remaining useful life (RUL) of the equipment to allow for proactive actions before failures so that the adverse failure impact is minimized. Thanks to sensing technology, it is possible to monitor the equipment by collecting various condition monitoring (CM) data. However, accurate RUL prediction is firmly challenged by accurate prognostic model development and inherent uncertainty quantification. This seminar will present research work that exploited some challenging issues in existing prognostic approaches.

The research work consists of three parts. In the first part, an integrated prognostics method is proposed for RUL prediction of a gear that is subject to fatigue fracture under constant loading and time-varying loading conditions. The physical models and CM data are integrated in a way that model parameters are updated in light of CM data. The model update process is cast into a Bayesian framework which is naturally amenable to uncertainty quantification. The RUL prediction gets improved because of the improved accuracy and reduced uncertainty in model parameters. The proposed integrated prognostics method is more advantageous than physics-based methods in that the model parameters are able to be adjusted for a specific unit in a specific working condition, and more advantageous than data-driven methods in that massive data trending is not necessary. The second part investigates the stochastic collocation method based on polynomial chaos expansion (PCE) to improve computational efficiency of uncertainty propagation in integrated prognostics. The results showed that PCE is able to significantly accelerate both RUL distribution computation and model update process, while achieving satisfactory approximation accuracy. These desired properties make it promising in a real time prognostic environment. The third part presents the integrated prognostics method for RUL prediction under shock degradation, which contains a discontinuity caused by sudden damage accumulation. This discontinuity is accommodated by introducing damage initiation time as a new uncertainty source. A virtual continuous degradation path is generated by searching an appropriate damage initiation time. In addition, future research directions are discussed toward the integrated prognostics methods for complex physical systems, efficient uncertainty quantification and experimental validation.

Biography                                                                                                                                                                  

Ms. Fuqiong Zhao got her Bachelor’s and Master’s Degrees in Computational Mathematics from Shandong University, China and she is expected to obtain her PhD Degree in Engineering Management from University of Alberta, Canada. Her research interests include prognostics, uncertainty quantification, finite element modeling, system dynamics modeling, fault diagnostics, signal processing and scientific computing. Her research papers have appeared in top-tier journals and conferences. She has participated in several research projects that are relevant to reliability and availability of engineering systems. She has been a research intern in Rolls-Royce Canada, Montreal under MITACS project for nine months researching the potential capability of acoustic technology for equipment health management. She was selected as one of the four finalists in 2014 Quality Control and Reliability Engineering Best Student Paper Competition organized by IIE Industrial and Systems Engineering Research Conference (ISERC), Montreal. She was the recipient of Queen Elizabeth II Graduate Scholarship in University of Alberta for her excellence in both academic and research achievements.