Prediction of NOx Generation Process Based on A Nonlinear MA model

Abstract

In the process of a thermal power generation, the prediction of the amount of NOx in contaminant has a positive effect on the elimination of NOx. In this paper, the nonlinear moving average (MA) model with time-delays is used as the prediction model to predict the process of NOx generation. Firstly, the correlation coefficient method is used to divide all variables affecting the NOx generation into several categories, and the variable whose correlation coefficient with NOx is the biggest in each category is selected as a main variable. Then BP neural network method is used to select the three variables with the greatest influence among the main variables as the input variables in the prediction model. Next, the correlation coefficient method is used to determine the time-delay parameters of the three input variables in the prediction model. What’s more, the least square method is used to estimate other parameters of the MA model to obtain a prediction model of a NOx generation process. Finally, the practical data from the generation process of a power plant are used to verify the effectiveness of the proposed prediction method.