Hybrid estimation technique for predicting butene concentration in polyethylene reactor

Abstract

A component of artificial intelligence (AI), which is fuzzy logic, is combined with the so-called conventional sliding mode observer (SMO) to establish a hybrid type estimator to predict the butene concentration in the polyethylene production reactor. Butene or co-monomer concentration is another significant parameter in the polymerization process since it will affect the molecular weight distribution of the polymer produced. The hybrid estimator offers straightforward formulation of SMO and its combination with the fuzzy logic rules. The error resulted from the SMO estimation will be manipulated using the fuzzy rules to enhance the performance, thus improved on the convergence rate. This hybrid estimation is able to estimate the butene concentration satisfactorily despite the present of noise in the process.