Multiple model adaptive strategy for model predictive control

Date of Completion

January 2002


Engineering, Chemical




This research effort addresses the important issue of developing an adaptive strategy for Model Predictive Control (MPC). MPC differs from classical control by employing a model of the process internal to the control architecture. The controller uses this model to estimate how operational changes and process upsets will impact the process over a prediction horizon. The internal model of the process employed in MPC is linear, but most chemical processes are nonlinear. Hence, the performance of MPC will degrade as the operating level moves away from the original design level of operation. To maintain performance of the controller over a wide range of operating levels, an adaptive control strategy for MPC is investigated. The research addresses the important issue of creating an adaptive strategy for Dynamic Matrix Control (DMC), which is the process industry's standard for MPC. The novelty of the strategy is in the subtle but important details. The method of approach is to design multiple linear DMC controllers. The tuning parameters for the linear controllers are obtained using previously published correlations. The controller output of the adaptive DMC controller is a weighted average of the multiple linear DMC controllers based on the current value of the measured process variable. The capabilities of the multiple model adaptive DMC strategy are investigated through computer simulation and experimental demonstrations. This research provides an adaptive DMC strategy that is simple to implement and use, requires minimal time and effort for updating model parameters, requires minimal knowledge of first principle equations, relies on the linear control knowledge of plant personnel, does not require sophisticated analysis tools, and is reliable for a broad class of process applications. ^