Lab Research Projects

Process Control Analysis, Tuning and Training
We develop methods and procedures used by thousands of practitioners, researchers and students around the world for process control system analysis, tuning and training. Current research efforts in our lab seek to extend these methods to such important areas as:
    - whole-plant modeling, simulation and control
    - automated controller design and tuning
    - controller performance monitoring
    - multivariable process identification and control

Adaptive Process Control
Adaptive controllers maintain performance by automatically adjusting their design as process conditions change. For adaptive controllers to make a major impact on industrial practice, however, they must be made easy for nonexperts to use and they must be made stable and robust over a broad range of applications. Lab publications of our research focus on methods for automating the design, validation and implementation of adaptive controllers. Also under study are methods for automatically supervising controller performance after implementation to ensure the controller remains stable and robust while it is online.

A Pattern-Based Control Methodology
The patterns exhibited in the recent history of certain process variables indicate the effectiveness of a controller in maintaining desired performance. For example, rapid, slowly damping oscillations in a controlled variable indicate that a controller is tuned too aggressively. Conversely, a slow response or a persistent offset indicates that the controller tuning is too sluggish. Lab publications on this research explore how pattern analysis tools can be used for the online evaluation of current controller performance. The pattern analysis result is then used to adjust the controller design or tuning to restore desired performance. Current efforts are devoted to experimental demonstrations of pattern-based control methods.