Prediction of strength of coal briquettes from Karakichi coal deposit using regression models

Abstract

The purpose of this study was to predict the strength of fuel briquettes based on Karakichi coal using regression models. This study used experimental data obtained by varying the technological para meters of briquetting, including compaction pressure, compaction time, temperature, humidity, bentonite clay content, and the ash content of the initial mixture. A detailed pairwise correlation analysis was carried out to identify the factors having the greatest influence on briquette strength. This analysis examined the resulting index (strength) and nine technological variables, including pressing pressure, pressing time, temperature, mixture moisture, bentonite clay content, briquette thickness, fractional composition, and duration of natural drying. The analysis quantified the degree of correlation between each factor and briquette strength, identifying the most significant ones. The greatest strength of correlation was demonstrated by the variable ash content of coal raw material (X₉), with which briquette strength is positively and almost linearly related (correlation coefficient r = 0.9787). Based on these data, a one-factor linear regression model was constructed according to the least squares method, which helped to simplify the forecast calculations without significant loss of accuracy. According to the model, at 32% ash content in the mix, the predicted strength value is 5.56 MPa. The high coefficient of determination (R²=0.9575) confirmed the reliability of the obtained model and its suitability for practical application. Thus, the proposed approach provided an objective and reproducible means of assessing the quality of briquettes, optimizing the production process and selecting technological parameters aimed at increasing the mechanical strength of fuel briquettes.

https://doi.org/10.70974/mat09225169