Systems Identification and Control
The identification methods under study cover from traditional least squares (linear or nonlinear), relay feedback, PLS, to subspace. Methods using closed-loop data are emphasized. Controller designs for multi-loop and multivariable control systems have been focused on the inverse-based methods. Performance limits and on-line methods to monitoring dynamic performance for single loop systems and multi-loop systems are among the current studies. For time-varying plant, on-line adaptive control with a least interference to a closed loop system is also of interests.
Modeling, Data Mining, and Monitoring of Dynamic Systems
Steady state and dynamic simulations of plant-wide chemical processes, such as reaction and separation processes, are of interests. The purpose of these studies is aimed to build models for optimization, control system design and performance assessment. For safety of process operations, topics of input/output time series, model identification, and statistical multivariate analysis for static/dynamic systems are of interest. The purposes of this study is focused on the developments of on-line real time methods for monitoring and diagnosis of process faults.
Recent Research Topic
The Representative Publication