Neural-fuzzy modelling of polymer quality in batch polymerization reactors

The estimation of parameters and obtaining an accurate and comprehensive mathematical model of the polymerization process is of strategic importance to the control engineering purposes in the polymerization industry. It is characteristic for these processes a grate non-linearity and many difficulties applying traditional estimation techniques. This paper describes an approach based upon neural-fuzzy representation of the model. A concrete model is constructed with the Sugeno fuzzy inference technique and a fuzzy-neural network is used to model the dynamic behavior of the polymer process. Such neural-fuzzy models of polymer quality could be used successfully for optimization and control of polymerization processes. Short example for such implementation is included with additional results for modeling of Mn and Mw.