Generalized predictive control algorithm overcoming model parametric mismatch

WU Xia-lai， LOU Gan-fei， CHEN Chao， FAN Sheng-wan
（Department of Automation， Lishui University， Lishui 323000， China）

Abstract： Aiming at the effect of model parametric mismatch to generalized predictive control，an output increment feedback generalized predictive control algorithm was presented. A speed function of output increment was introduced to design an output increment reference sequence and the system output was indirectly controlled by directly controlling output increment. The input increments were constrained by softness coefficient matrix to obtain smoothing input based on stepped control，and the computation of inverse matrix was avoided. Stability of system robustness was improved by using a factor to multiply the control increment. The simulation results indicate that the effect of model parametric mismatch is overcomed successfully by simplified predictive control algorithm. The output fluctuation during output adjusting process is offset，the adjusting time of dynamic response is reduced and the input is smoothed.
Key words： generalized predictive control algorithm； output increment feedback； model parametric mismatch