Application of Fuzzy Artificial Neural Networks Model for Modeling NOx Emissions in a Power Station Boiler on Generation Mechanism

Up to now, the research on artificial neural network modeling NOx emissions is based on simple neural networks, not taking full advantage of the characteristics between NOx generation mechanism and neural network. In this paper, an improved fuzzy artificial neural network based on NOx generation mechanism is presented. The formation mechanism layer of this model is based on generation mechanism of NOx in order to decompose the input parameters. An improved RBF sequential algorithm and a particle swarm algorithm based on optimal stopping method are used in structure identification and parameter optimization. The results show that: in comparison to the traditional model, this model reduces the number of fuzzy rules and the complexity of the network, thus to improve the stability and the generalization ability of the model. Consequently, the model presented can provide reference for huge power station boiler to reduce NOx emissions by auto-adjustment or reconstruction.