1998 Volume 31 Issue 4 Pages 604-612
A method for identifying a wiener-type model for nonlinear processes is proposed. The linear subsystem of the model is represented by simple transfer functions of the form such as: FOPDT (i.e. First-Order-Plus-Dead-Time) or an SOPDT (i.e. Second-Order-Plus-Dead-Time). The static nonlinearity is represented by an invertible function defined on an operation domain. Data from a two-stage relay feedback experiment in one single run is required. Simple algorithms are formulated to identifying both structure and parameters of the model. Estimation of parameters in both linear and nonlinear parts is performed separately and in an one-by-one sequence.
