Design Of Coke Oven Gas Collector Pressure Controller Based On Neural Network

The coke oven coking gas collection process is complex and the reactions occur inside the coke oven cannot be monitored.The collector pressure is used as a process indicator to determine the operating status of the coke oven through its changes.Due to the time-varying,strong coupling and non-linear characteristics of the collector pressure system,the collector pressure is complex and variable,and difficult to predict and control.In response to this situation,the collector pressure is studied from three aspects.Firstly,for the problem of no regularity of collector pressure,ARIMA method based on time series of collector pressure is used to model and predict the system,to explore the law of collector pressure change,to realize the early warning of abnormal collector pressure,to provide reference for blower power adjustment,and to avoid the situation that the butterfly valve opening falls into dead zone due to untimely blower adjustment.Second,to address the problem that there are many factors affecting the collector pressure and it is difficult to model,the production data are preprocessed and the mutual information is solved according to the kernel density estimation method.Based on the mutual information criterion,the conditional characteristics that have a large influence on the collector pressure are selected as control variables,and a multiple linear regression method is used to establish the system ARX model according to the selected variables.Finally,for the coupling problem between collectors and the bad effect of traditional control methods,the RBF neural net inverse controller is designed using inverse system to decouple and control the system,and the improved PSO algorithm is used to optimize the net,and the inverse controller combined with the original system to form a pseudolinear system to solve the decoupling of the system.Meanwhile,for the inverse system without feedback,the single-neuron PID controller and the PID controller are combined with the pseudo-linear system as auxiliary controllers to form a closed-loop feedback control,and the effects of controllers are compared.The results show the composite control system has better effect,stronger anti-disturbance capability.Figure 42;Table 10;Reference 54...