Control Of Coke Oven Combustion Chamber NO_x Based On Neural Network And The Optimization Of Heating Process

Nitrogen oxides( NO_x)is one of the main pollutants in coke oven heating combustion process.Therefore,it's of great significance to effectively control the emissions by grasping the NO_x emission characteristics in coking industry.Based on this reality,the author conducts a study of the control of coke combustion NOx emission.Coke oven is a complex multi-variable system with complex NO_x emission characteristics.It is difficult to calculate NO_x emissions by simple mathematical formula,and it's difficult to analyze data with the heavy testing workload in the coke oven production site,and limited measurement condition.What's more,the influence of the parameters on the stability coke oven combustion system is great.The advantage of neural network modeling is the black-box nature between the input and output.If the coke oven chamber is seen as a black box,a certain input will correspond to a certain output.Based on this,the prediction model used to test the characteristics of coke oven chamber NO_x emission with the application of artificial neural network is established.On the basis of analyzing the mechanism of production for coke oven NO_x emission,this paper elaborates the various influencing factors of NO_x emission in the heating and combustion process of coke oven,and find out two main factors :the temperature of heating combustion flame and the proportion of oxygen in the gas.According to the large amounts of data measured in experiment,this paper constructs the prediction model of combustion NO_x emission in coke oven by using the three layers neural network,and this model applies the flame temperature and the oxygen amount as the input,the NO_x emission as the output.And simulating the established model which is based on BP neural network in the matlab software.By controlling the emissions,we also optimize the synchronous heating process of coke oven,greatly reducing the fuel gas to coke oven heating system.This model can predict the value of NO_x emission through certain input directly when the model to predict the characteristics of coke oven combustion NO_x emission is built through the BP neural network.This paper presents a project application of this model in the actual project,the effect Application effects have been demonstrated.