Statistical Models for Predicting Air Temperature in Vietnamese Underground Coal Mines

Abstract

Vietnamese underground coal mining plays a critical role in the country’s energy sector. However, increasing energy demand, deeper coal extraction, and the growing use of mechanized longwalls have led to significant deterioration of thermal conditions in mine workings. Ensuring a safe and compliant microclimate now requires reliable tools for predicting air temperature in underground excavations. The study proposes a statistical modeling approach for forecasting air temperature at the outlet of longwall workings based on operational data collected from ten Vietnamese coal mines between 2017 and 2020. Multiple linear and nonlinear regression models were developed using seven independent variables, including inlet air temperature, relative humidity, airflow volume, equipment power, excavation depth, daily production, and excavation length. The models were validated against independent datasets and demonstrated high predictive accuracy (R² = 0.82 – 0.86), with maximum deviations below 3.2%. Simplified versions of the regression equations were also derived to facilitate practical application in ventilation engineering. The proposed models can serve as effective tools for optimizing airflow distribution and cooling strategies, supporting improved microclimate control, energy efficiency, and occupational safety in Vietnamese underground coal mines.