Predicting Mining Excavator Fuel Consumption using Machine Learning Techniques

Abstract

Fuel consumption represents about 30% of the total energy used in
surface mines. Modelling and prediction of fuel consumption of
mining equipment, including excavators is a valuable tool in assessing
both energy cost and greenhouse gas emissions. However, only a few
studies have reported on fuel predictions in mining operations. This
study presents the implementation of four machine learning
techniques (Random Forest, Gradient Boosting, k-Nearest Neighbor
and Multi-Layer Perceptron Neural Networks) in mining excavator
fuel consumption modelling and prediction based on collected dataset.
Multiple regression analysis was used as a baseline study. Coefficient
of correlation (R), root mean squared error (RMSE) and mean
absolute error (MAE) were the statistical metrices used to assess the
performance of the various models. The results indicate that all the
implemented machine learning algorithms performed better than the
multiple linear regression model. Although all the machine learning
models gave good performance in predicting fuel consumption, the
Gradient Boosting algorithm showed superior performance with high
R value (0.7330) and lowest errors (RMSE = 762.58, MAE = 582.15).
The Random Forest model showed poor performance in analyzing and
explaining the datasets. This study has shown the possible application
of machine learning models in predicting surface mine excavation
energy using operating parameters. These models can therefore be
used to analyze and improve mining excavator energy consumption
through the control of crucial factors which significantly impact on
energy consumption.