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

A broadly-applicable, control-relevant system identi®cation methodology for nonlinear restricted complexity models (RCMs) is presented. Control design based on RCMs often leads to controllers which are easy to interpret and implement in real-time. A control-relevant identi®cation method is developed to minimize the degradation in closed-loop performance as a result of RCM approximation error. A two-stage identi®cation procedure is presented. First, a nonlinear ARX model is estimated from plant data using an orthogonal least squares algorithm; a Volterra series model is then generated from the nonlinear ARX model. In the sec- ond stage, a RCM with the desired structure is estimated from the Volterra series model through a model reduction algorithm that takes into account closed-loop performance requirements. The e€ectiveness of the proposed method is illustrated using two che- mical reactor examples.