Abstract:
In the process of methanol conversion purification, a portion of the carbon monoxide in the feed gas will be converted to carbon dioxide and hydrogen so as to enlarge the content of hydrogen. In order to control the conversion rate of carbon monoxide quickly, this paper integrates the advantages of the PLS (partial least squares) in extracting information, removing noise and streaming data with mixed Pi Sigma fuzzy neural network to establish a model of carbon monoxide conversion rate. The simulation results show that the proposed model has a quicker simulation speed and higher accuracy, and can guide and adjust the hydrocarbon ratio of purified gas in the methanol synthesis.