Dynamic Prediction Of Nox Emissions Based On Factor Analysis And Narx Neural Network

NOx emissions of thermal power plants are closely related to the environment. It is very important to research on the prediction of NOx emissions. However, most of the current models are static and did not take into account the impact of the previous data. Boiler combustion is a process with large delay, and NOx generation also goes through a certain process, so we cannot just focus on the data at the current moment. In this paper, the dynamic prediction model of NOx emissions combining factor analysis and NARX dynamic neural network is proposed. The common factors that play the important roles in NOx formation are extracted based on the factor analysis method. It can eliminate the collinearity of the original data and reduce the complexity of modeling. The factor score matrix of common factors is used as the input of the NARX neural network, and a dynamic model is established, taking full account of the influence of various operating parameters and output parameters on the NOx emission at the previous moment. The simulation results show the effective and feasible of the model in NOx emissions prediction.