Monitoring NOx Emissions from Coal Fired Boilers Using Generalized Regression Neural Network

The formation of nitrogen oxides (NOx) associated with coal combustion systems is a significant pollutant source in the environment as the utilization of fossil fuels continues to increase, and the monitoring of NOx emissions is an indispensable process in coal-fired power plant so as to control NOx emissions. A novel "one-pass" neural network, generalized regression neural network (GRNN) was proposed to establish a non-linear model between the parameters of the boiler and the NOx emissions. The selection of the GRNN model's parameter is discussed. The method presented in this paper is applied to a case boiler of 300 MW steam capacity. The results show that the GRNN model predicted NOx emissions much more accurate than the widely-used "iterative" BPNN model and the multiple linear regression model. The main advantage of the GRNN model, by comparing with the traditional BPNN model, consists of the certainty of the predictive result, simplicity in network structure, quick convergence rate and much better predictive accuracy, especially for the case with a very large number of training samples. This approach will be a good alternative to the BPNN model which is commonly used to implement the predictive emission monitoring system (PEMS).