A Fuzzy Rule-Based Machine Learning Approach for Predicting Heat Stress Levels in Underground Metalliferous Mines

Abstract

Effective management of heat stress is vital to safeguarding miners’ health and maintaining productivity in underground mining environment. This study proposes a novel predictive approach for estimating Wet Bulb Globe Temperature (WBGT) using a Machine Learning (ML) based fuzzy logic approach. Key environmental parameters influencing heat stress, namely, working depth, air velocity, dry bulb temperature, wet bulb temperature, and barometric pressure were measured in an operational underground mine. Based on the collected data, a fuzzy logic model was developed, integrating expert-defined membership functions and inference rules tailored to the specific conditions of the underground mining setting. The proposed model offers reliable WBGT predictions, enabling proactive heat stress mitigation strategies. This, in turn, supports mine management in reducing heat-related risks and enhancing miners’ safety in underground workplaces.