A risk assessment model of spontaneous combustion for sulfide ores using Bayesian network combined with grounded theory

Sulfide ores, collectively referred to as a class of minerals with extensive applications, are increasingly susceptible to spontaneous combustion incidents as the mining environment evolves. The spontaneous combustion of sulfide ores is a complex nonlinear process, and the mechanisms of which remain not fully understood. Consequently, it is challenging to delineate the accident trajectory through traditional accident causation models and subsequently construct a probabilistic model for spontaneous combustion risk. To address this issue, this paper proposed a risk assessment model based on grounded theory and Bayesian network. The advantage of this model is that it employs grounded theory to categorize and organize factors influencing the occurrence of accidents, thereby forming a clear accident trajectory and providing a framework and basis for constructing a probabilistic assessment model. Furthermore, the model presented in this paper introduces the D number theory and the Noisy-OR model to handle uncertainties in the risk assessment process and to establish more rational conditional probability tables. The findings suggest that environmental temperature, heat dissipation conditions, safety awareness, and safety skills are the most influential factors in spontaneous combustion incidents when considering a particular point in time. However, when considering the entire process of sulfide ore pile spontaneous combustion, the temperature of the ore pile is the most critical accident factor to monitor. Therefore, the key to suppressing the spontaneous combustion of sulfide ores lies in reducing environmental temperatures, enhancing heat dissipation, and elevating the safety awareness of relevant personnel.