Research On Control Of The Acrylonitrile Polymerization Reactor Temperature Based On LMI

Acrylon is indispensable man-made fiber material in the textile industry. In the production processes of acrylon, polymerization reaction occupies a very important position. There are some parameters such as temperature, pressure, flow, liquid level and so on, which affect the polymerization reaction, but the foremost one is the temperature in the polymerization reactor. The control effect of temperature determinates the production's quality and quantity. The polymerization reaction with characteristics such as time-varying, nonlinear as well as large lag means that it is difficult to establish accurate mathematic model, so the classical PID control method based on the accurate mathematic model can hardly reach excellent control effect.According to the problems above, the author introduced mechanism of Acrylonitrile polymerization process and the structure of polymerization reactor. By analyzing the temperature model of polymerization process, the heat transfer status in the polymerization reactor can be known. After the model was simplified, the MATLAB software was used to simulate the model. The simulation results indicated that the polymerization reactor temperature model had nonlinear characteristic and lag characteristic.At second, the definition of linear matrix inequalities and three solving methods which conclude linear matrix inequality problem, eigenvalue problem, generalized eigenvalue problem were introduced. The principle and the characteristic of model predictive control also was explain. Characteristic was Predictive Model, rolling optimization, feedback control. From advantage and disadvantage of LMI and MPC, the method of model predictive control based on linear matrix inequality and the algorithmic steps was given. In order to prove this method the author give a two input and two output numerical examples that has uncertainties. the simulation result indicated that the model predictive control based on linear matrix inequality can control the output successfully.At last the uncertainty set was acquired by dealing with the temperature model, then model predictive control based linear matrix inequality controller was designed. The MATLAB software was used to simulate the effect of the controller. A traditional PID controller was designed to compare the effects between the different control methods. The simulation results indicated that the temperature control system of polymerization reactor which was designed linear matrix inequality based model predictive control had less overshoot and less steady state error, so it solved the problem of temperature control in polymerization reactor better.