Auto-tuning Predictive Pi Based On Relay Feedback And Its Application In Carbon Fiber Production

As one of the heating furnaces,the typical features of high temperature furnaces（HT furnaces）are time delay,non-linearity,time-varying parameters,system uncertainty and harsh environment of the furnace,which significantly deteriorate the temperature control process of the furnaces system.In order to achieve accurate,fast and robust temperature tracking performance,the advanced temperature control system of high temperature furnaces in carbon fiber production line is studied.The main research contents and innovations are as follows:1.In this paper,a temperature dynamic model is established by using the law of heat conservation,which is simplified into a first-order plus time delay（FOPDT）model,and the parameters of the biased relay feedback system with hysteresis are identified based on the model.Because the identification process does not use Nyquist Stability Criterion to approximate the non-linear link and the relay has hysteresis and bias characteristics,this method has higher accuracy,simpler identification steps and stronger anti-disturbance ability than traditional methods.In addition,it can also reduce the order of higher-order objects.2.HT furnaces have the characteristics of time delay and non-linearity.The conventional PID cannot meet the control standard.In this paper,a predictive PI（PPI）controller is proposed,which is composed of PI module and prediction module.PPI controller has the advantages of good dynamic quality,small overshoot,fast response speed,good robustness and simple parameter setting.In addition,a new PID is proposed for FOPDT model and integral plus time delay（IPDT）model,which is a simplified form of PPI controller and can approximate the control effect of PPI.3.In this paper,the anti-windup idea of traditional controller is summarized,and a series of anti integral saturation strategies for PPI controller are proposed for the first time.The feedback suppression method suppresses the integral term according to the saturation.Compared with the traditional anti integral saturation method,it can reduce the overshoot,has the advantages of less parameters,simple parameter adjustment and high adaptive ability.4.The coordination between system response speed and robustness has an important impact on the control system.In this paper,a control parameter optimization method based on particle swarm optimization（PSO）is proposed to improve the response speed and robustness of the system.Compared with manual parameter adjustment,PSO has higher efficiency,higher adaptive ability and better robustness.In addition,the PSO method can adjust parameters according to different quantitative indexes to meet the requirements of different systems,which is not the ability of manual parameter adjustment.