NOx Emission Trend Prediction for the Waste Incineration Process Based on Partial Least Squares with the Time Series Reconstruction and Exponential Weighting

Abstract

Accurate prediction of nitrogen oxide (NOx) emission is crucial for effectively controlling pollution in municipal solid waste incineration processes. However, it is challenging to construct a NOx emission prediction model with high prediction accuracy and easy engineering application. To address this, this paper proposes a robust and easily applicable NOx emission trend prediction model oriented to engineering applications, utilizing the partial least squares (PLS) method with the time series reconstruction and exponential weighting (TS-EW-PLS). The model is verified using operational data from an actual waste incineration process, and comparative analysis with the PLS model showed that the TS-EW-PLS model achieved a remarkable improvement of 27–38 % in prediction performance.