GRNN-FA hybrid model for predicting the useful life of dump truck tyres

Festus Kunkyin-Saadaari, University of Mines and Technology, Department of Mining EngineeringFollow
Alfred Kesseh, University of Mines and Technology, Department of Mining Engineering, Ghana
Victor Kwaku Agadzie, University of Mines and Technology, Department of Mining Engineering, Ghana

Author ORCID Identifier

Festus Kunkyin-Saadaari: 0000-0002-8202-2021

Alfred Kesseh: 0009-0004-4427-7666

Victor Kwaku Agadzie: 0000-0001-9616-3725

Abstract

The useful life of dump truck tyres is crucial for optimising economic efficiency and safety in mining operations. This study employs advanced hybrid models, including a generalised regression neural network (GRNN) combined with an artificial bee colony (ABC), ant colony optimisation (ACO), and firefly algorithm (FA), to predict tyre life at Perseus Mining Ghana Limited to enhance prediction accuracy. Sensitivity analysis using simple linear regression (SLR) identified tread depth as the most influential factor, with a correlation coefficient of 0.9638. Among the models, GRNN-FA demonstrated superior performance, achieving the highest correlation coefficient (r = 0.9586) and lowest mean squared error (MSE = 0.1019), root mean squared error (RMSE = 0.3191) and, mean absolute error (MAE = 0.2761). The study highlights that improving operational factors such as floor and road conditions, payload weight, and cycle time can significantly extend the tyre life. Implementing the GRNN-FA model can optimise tyre management, reduce downtimes, and enhance operational efficiency, thereby setting a new standard for tyre life prediction in the mining industry.

Recommended Citation

Kunkyin-Saadaari, Festus; Kesseh, Alfred; and Agadzie, Victor Kwaku (2025) "GRNN-FA hybrid model for predicting the useful life of dump truck tyres," Journal of Sustainable Mining: Vol. 24 : Iss. 2 , Article 4.
Available at: https://doi.org/10.46873/2300-3960.1451