Abstract:

By identifying  boiler operating parameters  closely related to  nitrogen oxides emissions, the improvement of PLS was used to predictthe nitrogen oxides emissions in the boiler coal combustion process. Based on calibration models uniformed by the random matrix and combined with several factors influenced by nitrogen oxides emissions, and variable single-output of the emissionsof nitrogen content, the PLS model was improved. The improved calculation modelwas used to test nitrogen oxides emissions. The result showed that this method was more accurate and faster than  conventional PLS when used to predict the emission of  nitrogen oxides. This method could  provide a means for the integration of  data calculation and process collection.

Key words: partial least squares(PLS); random matrix; model; nitrogenoxides