<-- Back to Group Members

Email: hjiang@ualberta.ca

Webpage: http://www.ualberta.ca/~hjiang/

Location of work: Chemical and Materials Engineering Building (Room 266B) - University of Alberta

Personal Details: Bachelor of Engineering in Automation (2003), Zhejiang University, China, currently a PhD student in Computer Process Control, at the University of Alberta. I am big fan of sports. My favourite sport is soccer and I started to play since I was 9. I really enjoy playing soccer. I also love to watch soccer games in Serie A. English Primer league, LFP, Euro Champions league, Euro Cup, World Cup, etc. Besides soccer, I also have great interest in basketball, tennis, squash and hockey (Go Oilers Go!).

Research area: Detection and Diagnosis of Poor Control Performance

Summary: In most chemical plants, there are several hundred or even thousands of regulatory control loops. For eficiency, all critical loops must operate at optimum levels. With such a wide span-of-control, process and performance monitoring systems are increasingly becoming necessary for early detection of faults, safety violations and performance degradation before they lead to unexpected disruptions or even catastrophic failures. Thus the monitoring of the performance of chemical processes has received much attention in the engineering research literature over the past few decades. However, the diagnosis of poor control performance has received little attention and remains an open research area. Performance diagnosis requires identification of the cause(s) of poor performance;and among the many possible reasons for poor control performance, the presence of oscillations and model-plant mismatch (MPM) are two common situations.

Oscillations are a common type of plant-wide disturbance and the root causes can be poorly turned controllers, sticky valves, oscillatory disturbance etc. The oscillation effects can propagate to many units and thus may impact the overall process performance. The presence of oscillations in a plant increases the variability of the process variables and thus naturally results in cause poor control performance, inferior quality products and larger rejection rates. Increasing emphasis on plant safety and plant profitability strongly motivates the search for techniques to detect and diagnose plant-wide oscillations.

Model-plant mismatch (MPM) is another a common factor that causes poor control performance, espe- cially for model based control system, e.g. model predictive control (MPC). The precondition of good performance of model based control is usually that the MPM is negligible, or at least the model dy- namics are fairly close to the plant dynamics. However, change in plant dynamics and therefore MPM, is inevitable as operating conditions change. How to detect and diagnose the model-plant mismatch (MPM) is an open and challenging problem. A satisfactory solution to this problem would be helpful in research in both MPC design and MPC monitoring.