Intelligent control of underground mine ventilation processes based on deep learning and real-time sensor data analysis

Abstract

This article examines the intelligent control of ventilation processes in underground mines, based on deep learning models and the analysis of real-time sensor data. Parameters in underground mines, such as methane, carbon monoxide, oxygen levels, temperature, humidity, dust concentration, airflow velocity, and pressure, are crucial for ensuring miner safety and the energy efficiency of the ventilation system. In traditional ventilation systems, fans often operate in constant or predetermined modes, which do not adequately account for actual technological conditions, worker locations, or gas-dynamic changes. The study proposes an intelligent approach aimed at collecting and preprocessing multi-parameter sensor data, detecting hazardous situations, forecasting the state of the mine atmosphere, and adaptively controlling fan operating modes. Using deep learning models, incoming sensor signals are analyzed in real time, the current state of the ventilation network is assessed, and the probability of hazard zone formation is determined. Based on the resulting forecasts, the control system automatically adjusts fan speed, airflow distribution, and the ventilation mode. The proposed approach maintains safe microclimatic conditions in underground mines, predicts when gas concentrations will exceed regulatory limits, enables rapid emergency response, and reduces excessive energy consumption.