AIChE Journal, Vol.52, No.2, 668-677, 2006
Step and pulse response methods for identification of Wiener processes
Lack of simple identification methods for nonlinear processes hinders field applications of nonlinear control systems. For identification methods that are as simple as those for the first order plus time delay models of linear dynamical processes, graphical and least squares methods to identify Wiener-type nonlinear processes from standard responses, such as step, pulse, and square-wave responses, are proposed. Static nonlinear functions are identified independently in Wiener-type nonlinear processes. Graphical methods extract discrete points of the nonlinear static function or a continuous non-parametric model of the nonlinear static function iteratively. The least squares method provides a parametric model of the nonlinear static function. The identified static nonlinear function can be used to design a simple linearizing control system. To illustrate the proposed identification methods, simulation and experimental results are given.