Research On Optimization Method Of Technological Parameters Of CO₂ Refining System

CO₂ is widely used as an important resource, and the discharge of CO₂ is also the key cause of the global warmth. So highly efficient recycling of CO₂ can save resources and reduce the air pollution and the greenhouse effect.In the CO₂ retrieving and refining system, which used adsorption distillation method, the product CO₂ purity quotient in the section of industrial CO₂ flash distillation is affected by many factors. By the analysis of balanced distillation principle and the flash distillation theory, we learned that the temperature of CO₂ at the liquid gas exit and the pressure of flash distillation are the two key parameters affecting product CO₂ purity quotient. The regular control plan adopts PID control method in controlling those two parameters, in which the technological parameters is determined by theoretical calculation and human experience, and this control plan's performance is not good enough and can cause instability of product CO₂ purity quotient.This paper adopts a technological parameter optimization method based on the acknowledging of the process of the industrial CO₂ flash distillation. Based on the historical data, the prediction model and technological parameter optimization model of the system are setup by the neural network, and then forecast the key parameters that affect the product CO₂ purity quotient the most. The key parameters are real time controlled by the PID controlling, which can overcome the defects of the instability of the relevant parameters and then product CO₂ purity quotient is stabled, and average purity is higher.This paper also includes a neural network parameters optimization method by Kalman filter, and the results show the optimization method can improve the performance of the neural network training.