Modeling of Drilling Rate and Costs in Open-Pit Mining: a Comparative Study of Machine Learning Algorithms

This study aimed to employ machine learning techniques, including Multilayer Perceptron Artificial Neural Networks (ANN-MLP), Decision Trees (DT), and Random Forests (RF), to predict drilling speed and operational drilling costs. The models were applied to a drilling database consisting of 288 samples containing operational information and drilling costs. The data were allocated into 15% for testing, 15% for validation, and 70% for training. The ANN-MLP model successfully predicted the drilling rate with an R2 of 0.78 and an MAE of 3.55, and it achieved an R2 of 0.76 and an MAE of 0.44 for predicting drilling costs. The DT model predicted the drilling rate with an R2 of 0.81 and an MAE of 3.07, and for drilling cost prediction, it achieved an R2 of 0.85 and an MAE of 0.32. The RF model predicted the drilling rate with an R2 of 0.93 and an MAE of 1.63, and for drilling cost prediction, it obtained an R2 of 0.96 and an MAE of 0.14. The RF model exhibited the highest effectiveness in predicting both the drilling rate and drilling costs in an open-pit mine, achieving a rate of 41.3 m/h and a cost of 5.25 USD/m.