Adaptive Inferential Control for Chemical Processes with Limited Measurement Problems

An adaptive procedure is incorporated in the inferential control system which has been developed to address the limited measurement problem in the presence of unmeasured disturbances. A stable approximation to the inverse of a primary process model forms the basis of the inferential control law. The use of the adaptive inferential controller based on a discrete convolution model is examined and compared with one based on a discrete transfer function model. Simulation results show that the proposed adaptive inferential control system performs well and provides improvements over conventional PID and inferential control in the face of unmeasured load changes and process variations.