Application Research Of Fuzzy Neural Network In Distillation Process

Fuzzy neural network knowledge is used to do some studies on alcohol distillation column. For alcohol distillation column system with model complexity, time variation and variables coupling, fuzzy neural network is studied on the modeling and control of alcohol distillation column.In view of the complex model of distillation column system, a hybrid modeling method is adopted to establish the model of distillation column. Mechanism model is introduced and fuzzy neural network compensation model is established. The mixed model is established by combining the mechanism model and the fuzzy neural network compensation model. In MATLAB environment, the test results of the mixed model reveal that the temperature error between the mixed model and the actual system is obtained at ? 1℃, and the steady mixed model can follow the state of actual system.In view of the time variation of distillation column system, a fuzzy neural network selftuning PID controller is used to control the system. The top temperature system of alcohol distillation column is selected as the study object. In MATLAB environment, the results show that fuzzy neural network self-adjusting PID control method is better than traditional PID control method, and it can meet the time-varying characteristics of top temperature system of distillation column.In view of the coupling of distillation column system, the control scheme of the fuzzy neural network self-adjusting PID controller with decoupler is used. The top temperature and bottom temperature system of distillation column is selected as the research object. In the MATLAB environment, the compare between the scheme of fuzzy neural network self-adjusting PID with decoupler and the scheme of traditional PID controller with decoupler is done. The compare results show that the scheme fuzzy neural network self-adjusting PID with decoupler not only can solve the coupling problem but also can solve the time variation problem.