A Fast Screening Method of Key Parameters from Coal for Carbon Emission Enterprises

Abstract

During the process of determining carbon emissions from coal using the emission factor method, third-party organizations in China are responsible for verifying the accuracy of the carbon emission data. However, these verifiers face challenges in efficiently handling large quantities of data. Therefore, this study proposed a fast screening method that utilizes multiple linear regression (MLR), in combination with the stepwise backward regression method, to identify problematic carbon emission data for the lower calorific value (LCV) and carbon content (C) of coal. The results demonstrated the effectiveness of the proposed method. The regression models for LCV and C exhibited high R-squared (R2) values of 0.9784 and 0.9762, respectively, and the root mean square error (RMSE) values of the validation set were 0.32 MJ/kg and 0.80% for LCV and C, respectively, indicating strong predictive capabilities. By analyzing the obtained results, the study established the optional error threshold interval for the LCV and C of coal as 2RMSE–3RMSE. This interval can be utilized as a reliable criterion for judging the quality and reliability of carbon emission data during the verification process. Overall, the proposed screening method can serve as a valuable tool for verifiers in assessing the quality and reliability of carbon emission data in various regions.

Keywords

• screening
• carbon dioxide emissions
• coal
• multiple regression
• lower calorific value
• carbon content