Hybrid Probabilistic–Machine Learning Model for Fire Risk Prediction in Underground Mines
Hybrid Probabilistic–Machine Learning Model for Fire Risk Prediction in Underground Mines
عطا اتحاد1 امین اتحاد2 سعید حکمت منش3
1) دانش آموخته کارشناسی ارشد زمینشناسی اقتصادی، دانشکده علوم، دانشگاه فردوسی مشهد، ایران
2) دانشآموخته کارشناسی ارشد مهندسی ایمنی، بهداشت و محیط زیست (HSE)، دانشگاه مهرالبرز، تهران، ایران
3) دانشآموخته کارشناسی ارشد مهندسی ایمنی، بهداشت و محیط زیست (HSE)، دانشگاه مهرالبرز، تهران، ایران
IOI Code
:
ACZA-ABHKDG
Publication :
(https://ceemconf.ir/2025/fa)
Abstract :
Fire incidents pose significant safety risks in deep underground copper mines, where elevated geothermal gradients, confined geometries, complex ventilation networks, and high equipment density increase both ignition probability and fire propagation intensity. Traditional deterministic fire risk assessment methods often fail to capture uncertainty and nonlinear interactions among geological, environmental, and operational factors, particularly under data-scarce conditions. This study proposes a hybrid probabilistic–machine learning framework for fire risk prediction in deep underground copper mining environments. A synthetic operational dataset comprising 10,000 scenarios was generated using Monte Carlo simulation to represent uncertainty across key risk variables. A probabilistic Fire Risk Index (FRI) was developed to quantify hazard intensity, and supervised learning models were employed for fire occurrence classification. Results indicate that the Random Forest model achieved an accuracy of 0.91 with an AUC exceeding 0.93, outperforming Logistic Regression. Carbon monoxide concentration, ventilation air velocity, and mining depth were identified as the most influential predictors. The proposed framework provides a scalable, data-driven approach for intelligent fire early-warning systems in deep underground mining operations.
Keywords : fire risk prediction deep underground mining probabilistic modeling machine learning monte carlo simulation ventilation airflow