Application Of Neural Network And Internal Model Control In Atmospheric And Vacuum Distillation Unit

The internal system of the atmospheric and vacuum distillation unit is complex and has many variables,and it is a complex system with nonlinear,large time delay,and strong coupling characteristics.Therefore,the application of PID control has great limitations in atmospheric and vacuum distillation unit.In order to improve the characteristics of the system,this artical proposes to apply the internal model control with simple design and strong robustness to the atmospheric and vacuum devices,and to use the artificial neural network to assist modeling and improve the control effect.First of all,according to the single-input single-output control circuit in the atmospheric and vacuum devices,the atmospheric column top temperature and the atmospheric furnace outlet temperature are selected as controlled objects,and a feedforward-feedback control loop and a cascade control loop are designed respectively.Apply the internal model control to the control loop,and the BP artificial neural network is used for identification.The module is simulated in simulink and compared with the PID control to prove that the artificial neural network internal model control has better control effect.Then,in order to solve the problem of high coupling in the distillation column,the product concentration at the top of the column and the bottom of the column is selected as the control object,and a multivariable internal model decoupling controller is designed in combination with the internal model control algorithm.The design requirements of time delay and RHP zero in the transfer matrix function are analyzed,and the BP neural network is used to identify the process system model.Simulate in simulink and compare it with PID feedforward compensation decoupling control.The results show that the decoupling control of the NN-IMC has rapid response and strong anti-interference,and it has a strong robustness.Finally,an advanced control experiment was conducted in the JT3000 atmospheric and vacuum experimental apparatus designed by SUPCON and the liquid level of the atmospheric column bottom was selected as the controlled object.In the distributed control system,PID control and internal model control experiments were conducted respectively,which verified the feasibility of internal model control in this device.The other major task of this project is to modify the equipment of a micro flow rectification experimental device that has not been used due to various reasons such as large measurement error,and the control loops and schemes in the DCS are designed and modified.Unsteady flow equivalent constant flow method was used to detect the small variable parameters in the device,so that the experimental device completed the detection and control of the micro flow,which can be used as an update topic in experimental teaching.