Deep belief network based NOx emissions prediction of coal-fired boiler

The emissions of nitrogen oxides (NOx) from boilers is related to various production parameters and has strong nonlinear characteristics, which is difficult to accurately predict. To address this problem, a comprehensive NOx emissions modeling method based on deep belief network (DBN) is proposed. In the algorithm, classification regression tree (CART) is used to select input variables by sorting the importance of input variables. In addition, a DBN is employed to establishe the NOx emissions prediction model. In this paper, the experimental data is obtained from the historical data of coal-fired power plants. The experimental results show that the average relative error of DBN prediction results is lower than extreme learning machine (ELM) and least squares support vector machine (LSSVM).