A Novel Simulator Using Minimal Resource Allocation Network and Its Application in Industrial Methanol Oxidation to Formaldehyde

Abstract

A simulator in order to calculate the rate of the reaction in the methanol oxidation to formaldehyde process is presented in this paper. Here the Radial Basis Function Network is used to model the process. To choose an optimum number of hidden neuron we use an algorithm called Minimal Resource allocation Network. It recruits hidden neuron based on the novelty of the input data. The training data were obtained from a model available in literature. The network is trained with the literature data and the resulted model gives a good prediction of rate of reaction of formaldehyde formation.