TITLE:
A Comparative Study of Two Tree-Based Models for Predicting Flyrock Velocity at Open Pit Bench Mining
AUTHORS:
Ezatullah Rawnaq, Bassir Esmatyar, Akihiro Hamanaka, Takashi Sasaoka, Hideki Shimada
KEYWORDS:
Flyrock, Machine Learning, Bench Blasting, Coefficient of Determination
JOURNAL NAME:
Open Journal of Applied Sciences,
Vol.14 No.2,
February
9,
2024
ABSTRACT: Blasting is a cost-effective technique to break
hard rock volumes by using explosives in the mining and civil engineering
realms. Moreover, although blasting is a designed process and plays an
indispensable role in these industries, it can also have multiple adverse
environmental impacts. One such effect is flyrock, which poses risks to nearby
machinery, and residential structures, and can even lead to injuries or
fatalities. To optimize blasting efficiency as well as restrict side effects,
prediction of the blast aftereffects is vital. Therefore, the present work
focuses on using two machine learning methods to predict the velocity of
flyrock in the open pit mine. To address this issue, a comprehensive dataset
was gathered from the open pit mine. Then, Decision Tree and Random Forest
algorithms were employed to predict flyrock velocity. The Random Forest model
demonstrated superior performance compared to the Decision Tree model.
Nonetheless, the performance of the Decision Tree model was deemed satisfactory,
as evidenced by its coefficient of determination value of 0.83, mean squared
error (MSE) of 4.2, and mean absolute percentage error (MAPE) of 5.6%.
Considering these metrics, it is reasonable to conclude that tree-based
algorithms can be effective in predicting flyrock velocity.