Application Of Fuzzy Neural Network Based On DEBP Algorithm In Light Gasoline Etherification System

With the continuous development of the national economy,the use of cars in our country is increasing,and the automobile has caused environmental pollution problems as well as the convenience of our travel.For this reason,the production of higher quality gasoline has become an urgent problem to be solved in our refineries.Light gasoline etherification technology can effectively reduce olefin content of gasoline,which is also an important way to improve gasoline quality in refinery.In the etherification process of light gasoline,improving the conversion rate of olefin is the core target of the whole process,while the essential factor affecting the conversion of olefin is the ratio of alkene in the reactor.The control process of alcohol alkene ratio in reactor is characterized by nonlinearity and large delay.The control effect of traditional PID control is not stable enough.In this paper,fuzzy neural network based on DEBP algorithm is applied to control the ratio of alcohol to alkene in light gasoline etherification system.First,a mathematical model is established for the control process of the ratio of alcohol and alkene in the reactor.Second,the fuzzy neural network controller is constructed,and the parameters are optimized by the BP algorithm.The control effect of the fuzzy neural network on the ratio of alcohol and alkene is better than that of the PID control through simulation and comparison.Because of the disadvantages of BP algorithm: the convergence speed of fuzzy neural network parameters is slow and the local extremum is easily trapped,this paper proposes a combination of differential evolution algorithm and BP algorithm to form a DEBP algorithm.After verifying the validity of the DEBP algorithm by using the standard function,the algorithm is used for fuzzy neural network reference.After learning and training,the parameters are optimized online byBP algorithm.The simulation results show that the control effect of fuzzy neural network based on DEBP algorithm is superior to the fuzzy neural network optimized only by the simple BP algorithm.In this paper,the DeltaV DCS system is used to control the etherification system of light gasoline.After studying the configuration control mechanism of this system,the driving operation of light gasoline etherification system has been successfully realized.The fuzzy neural network based on DEBP algorithm is used to control the ratio of alkene.The data processing and analysis system of oil etherification system shows that the control effect of alkene ratio controlled by advanced algorithm is superior to that of traditional PID control,and the conversion rate of olefin is greatly improved,which have great significance for improving the quality of refined oil in refinery.