Authors: Yu.N. Linnik, V.Yu. Linnik

Title of the article: Analysis of machine learning methods used in the prediction of rock bursts

Year: 2025, Issue: 5, Pages: 39-47

Branch of knowledge: 2.8.8. Geotechnology, Mining machines (engineering)

Index UDK: 622-1/-9

DOI: 10.26730/1816-4528-2025-5-39-47

Abstract: Rock bursts in underground coal mining operations remain a significant threat today, leading to fatalities among personnel and damage to underground excavations and mining equipment. This circumstance has prompted researchers in Russia and abroad to explore alternative methods for predicting the likelihood of rock burst occurrences. However, due to the complex interplay between geological, mechanical, and geometrical parameters of the mine workings, traditional forecasting methods based on mechanics do not always yield accurate results. With the advent of machine learning techniques in recent years, a breakthrough in predicting rock bursts has become possible. This article provides an overview of selected machine learning methods applicable to predicting the probability of rock bursts. The first part presents a general examination of the issue of rock bursts and reviews conventional prediction methods. Statistical data from Russian mines regarding gas hazards is provided, revealing that 57% of Russian mines fall into Category III in terms of gas hazard, as well as being categorized as super-category or dangerous in terms of rock bursts and sudden emissions. Subsequently, the article reviews the application of machine learning models in predicting rock bursts, detailing corresponding mechanisms, technical details, and efficiency analyses.

Key words: minerals mining coal drilling mining transportation coal processing machine learning deep learning signal processing computer vision

Receiving date: 14.04.2025

Approval date: 30.06.2025

Publication date: 09.10.2025

CITE DOWNLOAD