My research interests are in the areas of manufacturing system modeling, optimization and decision-support tools with advanced predictive analytics and machine intelligence. I am particularly interested in developing advanced models and methods for prognostics and health management of based on sensor information and design predictive and preventive strategies for manufacturing operations. These methods can assist to improve system’s resource utilization efficiency, to enhance system intelligence and resilience in responses to dynamic environment. Another area of my research interest is the design and control of remanufacturing systems which is within the scope of sustainable manufacturing. I have investigated new modeling and solution methods to advance the knowledge of remanufacturing strategies in faced with stochastic product returns. This research includes the design of methods for large scale sequential decision problems, such as Markov decision processes, approximate dynamic programming, etc.
Invisible performance degradation of machine and process impact both product quality and system productivity. Process variation and components (including sensors) degradation occurrence often are invisible until issues and faults are detected and diagnostic actions follow. To enable a resilient and near-zero-breakdown machine and processes, I have developed predictive tools and techniques to enhance intelligent prognostics and health management capabilities in components, machines, and production systems. More specifically, my research investigates modeling to deal with high dimensional and heterogeneous data environments thereby enhancing prognostics design with adaptive condition monitoring (i.e., Dual Extended Kalman Filters (DEKF), support vector regression based particle filters (SVR-PF), Haar-based power spectrum analysis) with applications in lithium-ion battery for electric vehicle, nickel-hydrogen battery, and progressive stamping machine [1-3]. These methods have enhanced the state-of-the-art on prognostics by taking the best advantage of using both information from physical knowledge and data either collected from practical usage or generated by virtual models. The results can be leveraged to the design of a unified Cyber Physical System platform.
My PhD dissertation has designed the optimal control strategies for remanufacturing systems of End-of-Life (EOL) products with stochastic return and demand, with a particular focus on the applications of electrical vehicle (EV) batteries. This research was funded by the Department of Energy (DOE) and was in close collaboration with General Motors R&D who is striving for new battery vehicle technology and clean vehicle development. EV is an energy efficient alternative to conventional vehicle powered by gasoline engines, but its market success depends on the reliability and life cycle performance of batteries. I have investigated remanufacturing as a sustainable solution that can significantly increase the total lifetime of expensive batteries by recovering their value in terms of material, energy and labor in the original products, thus reducing their costs to automotive OEMs and customers.
As an assistant research scientist at the University of Michigan, I have continued the line of study in these two areas. I have been leading several NSF I/UCRC projects on Intelligent Maintenance Systems and have developed and launched various models and simulation tools with industry collaborators to improve their performance and support automated decision-makings in complex manufacturing systems . My recent work with my colleague on Estimation of Maintenance Opportunity Windows (MOWs) has discovered hidden time windows for preventive maintenance [8-10] and has been successfully implemented in several assembly plants at GM and Ford. My study  was the first to bring the concept Real Option to the area of manufacturing systems. I developed a real-option based model to make joint decisions on maintenance schedule and production plan for systems subject to stochastic demand . This research provides a new perspective and way to mitigate the risks of machine tool degradation in manufacturing systems. The results have been reported in ASME conferences and CIRP Annual Meetings and published in leading peer-reviewed journals with international circulation.
My future research plans revolve around the development of predictive analytics for self-aware and resilient machines and manufacturing systems. I anticipate that the engineering applications of this research will be numerous and diverse, and the interdisciplinary research will provide researchers and practitioners in manufacturing enterprises with valuable insights through enhanced capability of prediction and decision-making. The remainder of this section briefly describes some specific potential research projects.