Prediction of NOx Emission Concentration Based on Weighted Optimal Combination Model

In order to achieve accurate prediction of NOx emissions concentration from coal-fired boiler outlets, this paper proposes a nitrogen oxide emission concentration prediction model based on a weighted optimization combination model. Firstly, in order to extract the depth time-frequency information in the data, the RLMD decomposition is used to decompose the input data. Secondly, considering the existence of redundant information in the input data, the ReliefF feature selection is used to select the strongly correlated feature variables as the modeling input. Finally, in order to improve the accuracy of the prediction model, Extreme Learning Machine (ELM), Echo State Network (ESN), Deep Neural Networks (DNN) and Long Short-Term Memory (LSTM) are combined, and the weights of the four models are optimized using the ant colony algorithm to construct the NOx emission concentration prediction model. Experimental results demonstrate that the proposed algorithm has a prediction error of less than 2% for NOx emission concentration under different operating conditions, and can achieve accurate prediction of NOx emission concentration from coal-fired boiler outlets.