The health risks associated with exposure to respirable dust in coal mines have been attracted much attention by an increasing number of researchers. However, the accurate identification of potentially hazardous agents in respirable dusts and the evaluation of the potential health risks arising from it still remains controversial to varying degrees. A comprehensive understanding of the physicochemical properties of respirable dust is a prerequisite and an important basis for resolving this controversy. Therefore, in this study, the particle size distribution and morphology, pore structure, mineralogical and geochemical patterns, and oxidative potential (OP) of respirable coal mine dust were comprehensively investigated. Stepwise multiple linear regression was employed to identify dust components driving OP, such as anatase, tobelite, quartz, and ankerite, in respirable coal mine dust, along with Na, Ni, Se, W, and As. On this basis, we performed a single-factor risk prediction for different coal mines by considering factors that may impact miners' health, with the analyses yielding somewhat contradictory results. Therefore, a multifactor integrated prediction model is proposed using an entropy-based technique for order preference by similarity to the ideal solution to categorize coal mines in the study area into three risk categories, high-, medium-, and low-risk dust mines, which is important for the hierarchical classification and control of coal mines and for formulating appropriate dust prevention and control measures.

煤矿粉尘暴露带来的健康风险已引起越来越多研究人员的关注。然而，对于呼吸性粉尘中潜在有害物质的准确识别以及由此产生的潜在健康风险的评估仍存在不同程度的争议。全面了解呼吸性粉尘的理化特性是解决这一争议的前提和重要基础。因此，本研究对可吸入煤矿粉尘的粒径分布和形态、孔隙结构、矿物学和地球化学模式以及氧化电位（OP）进行了全面研究。采用逐步多元线性回归来识别驱动 OP 的粉尘成分，例如可吸入煤矿粉尘中的锐钛矿、钙硅钙石、石英和铁白云石，以及 Na、Ni、Se、W 和 As。在此基础上，我们考虑可能影响矿工健康的因素，对不同煤矿进行了单因素风险预测，分析结果有些矛盾。因此，提出了一种多因素综合预测模型，采用基于熵的技术，通过与理想解相似的顺序偏好，将研究区煤矿分为高、中、低风险尘矿三个风险类别，对煤矿分级分类治理、制定相应的扬尘防治措施具有重要意义。