Machine Learning Approach to NOx Prediction for SCR Process of Thermal Power Plant

Abstract

This paper presents a machine learning application, to predict the formed nitrogen-oxides (NOx) in thermal power plants, and used in the control loops of the selective catalytic reduction (SCR) process. To deal with the big plant operation data, data reduction methods are also described. The predictions are applied for the feed-forward (FF) control and show improvement of control performance in the simulation study.