Propylene Oxide Production Process Optimization, Control

Propylene oxide is important material of organic chemistry industry. At present, in our country, the production plant of propylene oxide is of low level of automation. And its DCS system has not been taken full advantage of. Using advanced technology of automation and information to reconstruct the existing plant and optimize its production process is an effective method to improve the production capacity of propylene oxide of our country.The production plant of propylene oxide of Jinzhou Chemistry Industry Ltd. is the study object of the paper. Based on the reconstruction of its DCS system, the technology of advanced control, neural network and on-line optimization is introduced into its control system to solve its existing problem. At first its DCS system is redesigned and an on-line optimization scheme is proposed. Then, aiming at the existing problem, the algorithm of dynamic recurrent neural network, RBF neural network and adaptive inverse control is studied in the paper. A very good result of disturbance-resistant performance of adaptive inverse control is reached. They are all successfully applied to the plant.The purpose, improving production quality, reducing consume and improving production efficiency, can be attained by the scheme proposed in the paper. The results of simulation show that the existing control problem of the plant can be successfully solved by neural network, adaptive inverse control and the result is good.