我国煤矿开采环境复杂、地质条件恶劣，一线班组作为煤矿生产最核心的组织，其安全生产水平的高低直接影响着企业的经济效益。安全绩效评价作为衡量组织安全水平的有效手段，不仅能够保障一线员工的生命健康安全，还利于企业安全生产水平的提升。当前安全绩效相关的研究成果较为丰富，但鲜有研究立足于煤矿一线班组的角度。鉴于此，本文基于煤矿一线班组安全绩效影响因素的识别和特征分析建立了安全绩效的等级评估模型，提出了一线班组安全绩效的提升方法，从而实现对煤矿一线班组安全绩效的合理评价。具体开展的研究如下：

（1）基于扎根理论的质性研究方法，对煤矿一线班组的30起事故案例分析和深度访谈研究两种途径收集到的文本资料进行三级编码分析，得出了煤矿一线班组安全绩效影响因素的23项子范畴和6个主范畴，并进行理论饱和性检验，从而构建了煤矿一线班组安全绩效影响因素体系。

（2）以识别出的影响因素为基础，运用解释结构模型（ISM）方法构建了影响因素的8级递阶层次结构模型，明晰了因素间的层次结构；运用交叉影响矩阵相乘（MICMAC）进行因素的驱动力-依赖性分析，将影响因素划分为自治型、独立型和依赖型三大类，剖析了各类因素的属性特征。

（3）从安全人员、安全过程和安全结果三方面建立了影响一线班组安全绩效的评价指标体系，并结合定量与定性分析法制定了指标量化标准与评价结果分级标准；基于粒子群算法（PSO）优化BP神经网络建立煤矿一线班组安全绩效等级评估模型，收集75组样本数据对该模型进行了仿真实验。结果表明，PSO-BP模型的预测精度优于BP神经网络，说明使用粒子群算法优化BP神经网络可以提高模型预测精度。

（4）对山西某矿一线班组的安全绩效评价进行了实证研究，结果显示一线班组的安全绩效水平等级为2级，与该矿目前的实际情况基本吻合，进一步证明了该模型的合理可行性；结合指标量化数据和实际情况，识别了该矿安全管理工作中的薄弱环节，并提出相关的管理建议，助推煤矿企业安全发展。

The coal mining environment in China is complex and the geological conditions are harsh. As the core organization of coal mining production, the level of safety production of frontline teams directly affects the economic benefits of enterprises. Safety performance evaluation, as an effective means of measuring the level of organizational safety, not only ensures the life, health, and safety of frontline employees, but also benefits the improvement of the safety production level of enterprises. There are currently abundant research results related to safety performance, but few studies are based on the perspective of frontline coal mine teams. In view of this, this paper establishes a level evaluation model for safety performance based on the identification and characteristic analysis of factors affecting the safety performance of frontline coal mine teams, and proposes measures to improve the safety performance of frontline coal mine teams, in order to achieve a reasonable evaluation of the safety performance of frontline coal mine teams. The specific research conducted is as follows:

(1) Based on the qualitative research method of Grounded theory, a three-level coding analysis was carried out on the text data collected from the analysis of 30 accident cases and in-depth interviews of coal mine first line teams. From this, 23 sub categories and 6 main categories of the factors affecting the safety performance of coal mine first line teams were obtained, and the theoretical saturation test was carried out, so as to build the system of factors affecting the safety performance of coal mine first line teams.

(2) Based on the identified influencing factors, an 8-level hierarchical structure model of influencing factors was constructed using the Interpretative Structural Model (ISM) method, clarifying the hierarchical structure between factors. Using the Cross Influence Matrix Multiplication (MICMAC) to analyze the driving force dependency of factors, the influencing factors were divided into three categories: autonomous, independent, and dependent, and the attribute characteristics of each type of factor were analyzed.

(3) We have established an evaluation index system that affects the safety performance of frontline teams from three aspects: safety personnel, safety processes, and safety outcomes. By combining quantitative and qualitative analysis methods, we have established indicator quantification standards and evaluation result grading standards; Based on Particle Swarm Optimization (PSO), a BP neural network was optimized to establish a safety performance evaluation model for frontline coal mine teams. 75 sets of sample data were collected to conduct simulation experiments on the model. The results indicate that the prediction accuracy of the PSO-BP model is superior to that of the BP neural network, indicating that using particle swarm optimization to optimize the BP neural network can improve the model's prediction accuracy.

(4) An empirical study was conducted on the safety performance evaluation of a frontline team in a certain mine in Shanxi. The results showed that the safety performance level of the frontline team was level 2, which is basically consistent with the current actual situation of the mine, further proving the reasonable feasibility of the model; Based on the quantitative data of indicators and the actual situation, the weak links in the safety management work of the mine were identified, and relevant management suggestions were proposed to boost the safety development of coal mining enterprises.