Dynamic fuzzy modelling of cooling coil system

Modelling of complex non-linear systems using the process data is a challenging issue. This paper presents the dynamic fuzzy modelling of a cooling coil system using the input-output process data. The structure of the model is kept fixed as zero-order Takagi-Sugeno (TS) fuzzy nonlinear output error (NOE) model. The parameter identification is done using the recursive least square (RLS) technique. There are three inputs to the system and a single output Le. a MISO system. The modelling is carried out in three steps i.e. offline parameter identification, the online parameter identification and then dynamic modelling. Simulation results have been presented which demonstrate the efficiency of dynamic modelling with online parameter identification as compared to the techniques. The online models are extremely useful in model based control techniques.