Aprendizaje

Blast Movement Simulation Through a Hybrid Approach of Continuum, Discontinuum, and Machine Learning Modeling

Wei Fu, Jason Furtney, Jorge Valencia

Abstract

Estimating and optimizing rock movement during blasting is important to prevent unnecessary material handling, reduce ore loss and dilution, and minimize environmental footprint. It has been challenging and computationally burdensome to model the whole dynamic process because rock blasting consists of a complex process that generally involves explosive detonation, gas expansion, stress wave propagation, rock fragmentation and throw, and muckpile formation. In this regard, we propose a hybrid approach that captures the first-order impacts on the rock movement due to blasting while achieving accelerated simulation. Specifically, a small-scale continuum model is established to represent an annulus of rock, with explosive in the center and gas pressure resulting from the detonation applied to the borehole surface, which reduces as the borehole deforms. The continuum model simulates the early-stage, near-field rock blasting process and forms a synthetic dataset based on realistic explosive data to train a machine learning model. Key parameters, such as expanded hole diameter, burden velocity, and time-dependent gas pressure, are readily obtained from the constructed machine learning model. Informed by the machine learning model, the subsequent discontinuum model simulates the dynamic rock movement and predicts the muckpile formation in the far field using the rolling resistance contact model. Our results demonstrate the efficacy of the proposed approach to capture the key physics of blast-induced rock movement and realize accelerated blast design optimization aided by machine learning.

Fu, W., Furtney, J., & Valencia, J. (2023). Blast Movement Simulation through a Hybrid Approach of Continuum, Discontinuum, and Machine Learning Modeling. In Proceedings, 57th U.S. Rock Mechanics/Geomechanics Symposium (ARMA, Atlanta, Georgia, June 2023), ARMA 23-831. https://doi.org/10.56952/ARMA-2023-0831