Application Study Of Neural Network Optimization Algorithm In Hydrocracking

With the rapid development of modern industry and the continuous innovation of chemical production technology,the production process is becoming larger and more complicated.At the same time,modern production often has higher and higher requirements on product quality,the original single simple control system can not meet its requirements,so the advanced control system came into being.Predictive control is one of the most representative and effective control methods.It is a new type of computer control algorithm developed directly from industrial process control.Because it uses multi-step prediction,rolling optimization,feedback correction and many kinds of control strategy,thus the control effect is good,more suitable for those who is not easy to build accurate mathematical model and those with the more complex industrial process,so brought to the attention of the engineer and professor of control at home and abroad,is a kind of very promising new type of computer control algorithm.In order to study the above problems,the hydrocracking reaction,which is a common and important chemical link in modern refinery enterprises,was modeled in this paper,and the conversion rate of hydrocracking products was estimated by using the advanced control algorithm of neural network.In the process of research,it is found that the pure neural network algorithm does not perform well in the prediction model and has many limitations.In view of the limitation of neural network,an optimization method for neural network model based on adaptive genetic algorithm is proposed after a lot of references.By comparing the prediction accuracy of neural network model with that of neural network model based on adaptive genetic algorithm,it is found that the problem of the limitation of neural network such as random initial weight is well solved.Finally,the hydrocracking conversion prediction model presented in this paper was constructed through MATLAB programming,and real chemical process parameter data were used for conversion prediction,and the accuracy of the prediction model was analyzed through simulation.In modern chemical production,many production processes are complex industrial production processes,and the reaction mechanism itself is complex,and the reaction process is non-linear,uncertain,time-varying and strong coupling,so it is of great practical significance to study the application of advanced control.Hydrocracking reaction is a common and important chemical process in refinery enterprises.The study on hydrocracking product conversion rate is helpful to control it in actual production,so as to optimize the reaction process and improve the production efficiency of chemical enterprises.