Intelligent big data-driven system for predicting gas-dynamic phenomena in underground mines

Abstract

This paper develops a data-integration model for early assessment of gas-dynamic hazards in underground coal workings. Gas measurements are considered together with ventilation, microseismic, geophysical, and production parameters rather than as independent alarm signals. The framework assigns four operating states: low, moderate, high, and critical risk. Four computational scenarios were used to examine responses to changes in methane release, airflow, gas pressure, drainage performance, and production load; an additional virtual case reproduced the deterioration sequence used in the M-25-inspired example. In the reference case, methane remained at 0.5–0.7% with a fan capacity of 195 m 3 /s. Raising methane emissions from 28 to 45 m 3 /min increased the concentration to 1.1–1.4% and resulted in a moderate-risk state. When fan capacity was instead reduced to 145 m 3 /s, methane reached 1.6–1.9% even though emission remained near 28–30 mVmin, and the system assigned a high-risk state. The critical case combined 55 m 3 /min methane emission, 130 m 3 /s fan capacity, and 75% drainage efficiency; methane reached 2.2–2.5%. The scenario comparison indicates that loss of ventilation and drainage performance can create a greater local hazard than an increase in methane emission alone.