Statistical Models for Predicting Cement Factory Emissions

Abstract

This research studied the relationship among important factors of cement production – namely, raw materials, fuels, and manufacturing processes – and gaseous and particulate emissions. Two types of statistical prediction models, multiple regression (MR) and artificial neural network (ANN), were developed and compared. The recorded daily average data of raw materials, coal fuels, alternative (hazardous waste) fuels, production processes, and gaseous and particulate emissions in 2007 were used in the analysis. Results show that the MR and ANN models for predicting NO₂, SO₂, CO₂, HCl and TSP, have the Adjusted R2 values in the range of 0.25-0.57 and 0.44-0.66, respectively. It is also found that the independent variables that have significant effect on the of models are quantity of clay, quantity of limestone, raw mill running time, alternative fuels used, kiln running time, and quantity of clinker. Overall, the ANN models perform slightly better than the MR models.