Design and Application of Block-Oriented Fuzzy Models – Fuzzy Hammerstein Model

Recently, data driven fuzzy modelling has drawn a great deal of attention not only from academia but also from industry. However, a dynamic fuzzy model might be difficult to develop when training data do not contain sufficient information about the process nonlinearity. In order to avoid this problem and to reduce the model complexity, this paper presents a fuzzy Hammerstein (FH) modelling approach, where a static fuzzy model is integrated with a linear dynamic model in series. A constrained recursive least-squares algorithm is used for the simultaneous identification of the steady-state and the dynamic part of the FH model. To demonstrate the proposed modelling approach, an electrical water-heater process is used as an example.