A fuzzy logic tuned polynomial based predictor for processes having long dead-times

Abstract

Prediction using linear interpolation works well in systems where the destabilizing effects of long dead-times are not dominant. When the process contains a long dead-time, prediction using linear derivative control will not work because the measured signal will not contain sufficient information about any future changes that may occur in the loop. Under these conditions, the anticipatory nature of the derivative controller fails to ensure proper control of deadtime dominant processes. This paper proposes a fuzzy logic tuned, polynomial based nonlinear predictor to improve the control of plants experiencing instability due to long dead-times. The proposed predictor replaces the derivative controller in a PID controller, and contributes towards a general improvement of the control performance in systems having long dead-times.