Modeling and parametric analysis of low-temperature oxidative self-heating in coal stockpiles driven by natural convection

To accurately ascertain the behavior of oxidation heating and emission of oxidation products in fully enclosed coal storage yards under the influence of natural convection, comprehensive adiabatic oxidation experiments were conducted, encompassing coal particles of multiple grain sizes. Linear regression analysis was applied to deduce the heating characteristics of coal particles within real coal stockpiles from the experimental data. Subsequently, numerical simulations were executed, leveraging the acquired oxidation parameters specific to the stored coal. These simulations unveiled that natural convection triggers the formation of spiral airflow on both sides of the stockpile, with the strength of this airflow increasing in tandem with rising temperatures. An inflection point becomes evident during the impact assessment of porosity on maximum temperature variation, whereby its augmentation improves the heating rate when below the critical threshold, but conversely diminishes it when surpassing that threshold. Furthermore, coal particles smaller than 3 mm effectively isolate oxygen. As the degree of metamorphism increases, the rate of temperature rise decreases, and the high -temperature region penetrates deeper into the coal stockpile. The genetic algorithm-optimized random forest algorithm emerges as a highly effective tool for the precise prediction of the efficient storage duration of the coal stockpile. Keeping porosity below 0.25 proves effective in oxygen isolation. Lower heights and steeper slope angles exhibit the capacity to decelerate oxidation during high-temperature phases, thereby minimizing the loss of coal calorific value and greenhouse gas emissions. This study is paramount for the achievement of secure, efficient, and environmentally friendly coal storage practices in the context of natural convection.