Modeling of Methanol to olefin converting Reactor: A neural network Approach

In this paper an isothermal fixed bed reactor for methanol to olefins converting was modeled using artificial neural network. The space time, weight fraction of water in the feed and temperature of reactor bed were input variable to network. The output variables of network were methanol, Ethylene, propylene and C4 fractions. Levenberg-Marquardt algorithm was used for training of network. The optimum topology of network was a three layers network with light neurons in hidden layer. The RMS(root mean square error) for network output were calculated and it is approximately agreed with the experimental values.

کلیدواژه ها:

Methanol ، olefin ، levenberg-Marquardt ، neural network