A methodology for control-relevant nonlinear system identification using restricted complexity models

A broadly-applicable, control-relevant system identification methodology fo r nonlinear restricted complexity models (RCMs) is presented. Control desig n based on RCMs often leads to controllers which are easy to interpret and implement in rear-time. A control-relevant identification method is develop ed to minimize the degradation in closed-loop performance as a result of RC M approximation error. A two-stage identification procedure is presented. F irst, a nonlinear ARX model is estimated from plant data using an orthogona l least squares algorithm; a Volterra series model is then generated from t he nonlinear ARX model. In the second stage, a RCM with the desired structu re is estimated from the Volterra series model through a model reduction al gorithm that takes into account closed-loop performance requirements. The e ffectiveness of the proposed method is illustrated using two chemical react or examples. (C) 2001 Elsevier Science Ltd. All rights reserved.