为了准确有效评估煤矿瓦斯爆炸风险，提出了基于熵权法改进的模糊贝叶斯网络瓦斯爆炸风险评估模型。首先，对瓦斯爆炸事故案例进行风险识别，提取出18个瓦斯爆炸主要风险因素；其次通过故障树模型，并根据映射规则建立出相应的贝叶斯网络模型；为减少专家判断的主观性，将熵权法结合模糊理论得到的组合权重作为贝叶斯网络的先验概率，然后通过此模型对贵州松林煤矿的瓦斯爆炸危险性进行了评估。结果表明：贵州松林煤矿瓦斯爆炸风险概率为25%，风险等级为一般风险；瓦斯积聚和煤炭自燃是导致瓦斯爆炸的主要风险因素；其中，通风阻力、瓦斯突出、防爆设备故障、违规爆破、煤自燃等因素是瓦斯爆炸的关键致因因素，评价结果与实际情况相符。

In order to accurately and effectively evaluate the risk of gas explosion in coal mines, an improved fuzzy Bayesian network gas explosion risk assessment model based on entropy weight method is proposed. Firstly, the risk identification of gas explosion accident cases is carried out, and 18 main risk factors of gas explosion are extracted. Secondly, the fault tree model is established, and the corresponding Bayesian network model is established according to the mapping rules. In order to reduce the subjectivity of expert judgment, the combination weight obtained by entropy weight method combined with fuzzy theory is taken as the prior probability of Bayesian network, and then the risk of gas explosion in Songlin Coal Mine of Guizhou Province is evaluated by this model. The results show that the risk probability of gas explosion in Songlin Coal Mine in Guizhou is 25 %, and the risk level is general risk. Gas accumulation and coal spontaneous combustion are the main risk factors of gas explosion. Among them, ventilation resistance, gas outburst, explosion-proof equipment failure, illegal blasting, coal spontaneous combustion and other factors are the key causes of gas explosion, and the evaluation results are consistent with the actual situation.