Integrated proximate analysis and multiple linear regression for calorific value predictions of Tolay coal, South West Ethiopia

The rational approximation of the calorific value of coal is important for effective quality control and industrial consumption. However, in many new coal mining jurisdictions, the direct determination of calorific values remains a technically demanding and costly process. The proximate analysis, which measures moisture, ash, volatile matter content, as well as fixed carbon, offers a more hands-on and cost-effective alternative for analyzing coal characteristics compared to that of bomb calorimetric measurement. This study, therefore, demonstrates the relationship between the energy value of the coal and the proximate analysis parameters of coal gathered from the Tolay coal mine sites in Ethiopia. More than 54 representative samples were studied using ASTM standard procedures. Following the analysis, a statistical predictive model was established using the multiple linear regression (MLR) mechanism. The resulting model developed from the dataset is GCV (MJ/kg) = −12.73 + 0.05A +0.53VM + 0.23FC + 0.11 M, which showed a strong predictive performance with a coefficient of determination (R2) of 0.913, a root mean square error (RMSE) of 1.348, and a mean absolute error (MAE) of 0.886. The developed model indicated that volatile matter and fixed carbon have a positive correlation to the calorific value, while the ash content has a negative correlation. In conclusion, the findings have come to indicate that integration of proximate analysis with the regression method provides a reliable, consistent, and economically feasible approach for predicting the heat of combustion of coal.