Predicting the pyrite oxidation process within coal waste piles using multiple linear regression (MLR) and teaching-learning-based optimization (TLBO) algorithm

Predicting the pyrite oxidation process within coal waste piles using multiple linear regression (MLR) and teaching-learning-based optimization (TLBO) algorithm

Coal mining often leads to significant environmental hazards and health concerns when sulfide minerals, particularly pyrite, are associated with coal waste. The oxidation of pyrite typically generates acid mine drainage, a significant problem. This paper presents two mathematical relationships using a teaching-learning-based optimization (TLBO) algorithm for predicting pyrite oxidation and pH changes within a coal waste pile from Alborz-Markazi in northern Iran. A dataset was built based on historical data to achieve this goal. Some influential parameters comprising the depths of the various samples, oxygen fraction, and bicarbonate concentrations were considered as input data, while the outputs were pyrite content and pH. Then, the best statistical relationships were suggested between input and output parameters employing curve and surface fitting methods. Afterward, two multiple linear regression (MLR) models were presented for predicting pyrite content and pH. Also, two relationships have been suggested for predicting the same outputs by applying the TLBO algorithm. Comparison of the results of the latter method with the results obtained using the statistical technique, including correlation coefficient and root mean squared error (RMSE), revealed that the TLBO could predict the outcomes better than the MLR.

Files associated with this item cannot be displayed due to copyright restrictions.