Self-tuning PID controller using δ-model identification

This contribution presents an application of a self-tuning digital PID controller for process control modelled by delta-models. The process is identified by the regression (ARX) model using the recursive least-squares method (RLSM) with LD decomposition and applied directional forgetting. Controller synthesis is designed on the basis of a modified Ziegler-Nichols criterion for digital PID control loops. The ultimate (critical) proportional gain and period of oscillations have been derived for the second-order delta-model. Control results obtained using digital PID controller on the basic delta-models and z-models are compared. Copyright (C) 2002 John Wiley Sons, Ltd.