Efficient Constrained Predictive Control Algorithm for 2-Input Systems

摘要

A multivariable constrained predictive controller based on the generalizedpredictive control approach is formulated. The quadratic cost function used includes a term involving the full value of the control signal. This is in anticipation of applications where the control objective is to minimize the control energy instead of the control activity. Incremental model is used in the prediction equation to reduce the effect of steady state offset. A special case of 2 input/1 input system is treated in detail and a simple solution which avoids the quadratic programming solution is obtained. The algorithm also gives an interpretation to the sequencing and range splitting control strategies often found in industrial controllers.