Prediction of NOx in SCR Inlet of Thermal Power Plant Based on Mutual Information and Long Short-Term Memory Network

NOx produced by thermal power plant incineration is one of the main sources of atmospheric pollutants. Most thermal power plants use selective catalytic reduction (SCR) technology to denitrate. The current NOx concentration sensor has a large delay and cannot accurately reflect the real-time change of the NOx concentration at the SCR inlet, which will affect the economy and safety of the denitration control. This paper uses balanced mutual information feature selection (BMIFS) for feature extraction. Driven by historical data, a long short-term memory network (LSTM) model of inlet NOx concentration was established. By comparing the error with the field measured data, the accuracy of the prediction model and the effectiveness of the prediction method are verified. The conclusion shows that the balanced mutual information feature selection method can reduce the coupling degree and data dimension between the input variables, and achieve the purpose of improving modeling accuracy and reducing computational complexity.