A Hybrid Model XBORE for Predicting the Performance of Truck-Haulage Systems in the Ore Transportation Process at Underground Mine

Abstract

Most mining companies worldwide endeavor to promote the continued flow of ore at every mining stage to ensure the mines’ productivity and profitability. However, issues such as ore transportation are among the major challenges contributing to the failure to meet planned production targets. In this paper, a thorough study of the challenges faced during the ore transportation process at the Mopani underground mine located in Kitwe town in the Copperbelt province of Zambia is undertaken. Then XBORE hybrid prediction model using XGBOOST, RF, and Voting Regression algorithms is developed that will help to predict and forecast the production of copper in tonnage by analyzing the links between the ore extraction and transportation from the underground sites to production outputs at the metallurgical plants. Historical ore transportation data records from the Mopani underground mine and real-time data at the mine site were used to train the proposed XBORE hybrid model. In evaluating the XBORE hybrid model’s performance, Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Root Mean Squared Logarithmic Error (RMSLE), Mean Absolute Percentage Error (MAPE), and R-Squared (R2) were the metrics used. Results revealed the XBORE model exhibited superior performance and high accuracy. In the test phase, R2 of 0.981 was obtained compared to 0.93 and 0.42 for the AdaBoost regressor and K-neighbors regressor respectively. Evaluation and analysis showed that XBORE performed better than other state-of-the-art algorithms (namely; GBR, ADA, DT, LR, BR, KNN, SVR, and CNN) concerning its robustness, accuracy, and computational efficiency.

Keywords

• Algorithms
• hybrid model
• machine learning
• production
• transportation