任何事故发生必有险兆。目前在安全生产管理中仍然存在重视事发后调查处理而 忽视事发前预防管控的倾向，致使大量险兆事件（Near-Miss）得不到有效控制而发展成为伤害事故。本文致力于煤矿险兆事件的辨识、上报及分级管控，在对国内外有关文献梳理研究的基础上，对险兆事件的特征要素、致因机理作了探究，对危险源、隐患、险兆事件、事故之间的区别和关系做了分析，重点围绕煤矿险兆事件的辨识上报、量化分级、管控体系构建等重点环节开展研究。 首先，利用扎根理论，获取了煤矿险兆事件主动上报意愿的影响因素，然后运用G1赋权法确定了上报影响因素各层次指标的权重值。三个维度中“险兆事件认知”对煤矿险兆事件主动上报的影响最大，而在三级指标体系中，“概念的明晰性”对煤矿险兆事件主动上报的影响最为明显。同时用结构方程模型对煤矿险兆事件主动上报意愿影响因素进行了验证性因子分析，讨论了各个变量对上报的影响。 其次，采用f值法、权重分析系统和集对分析法分别对煤矿险兆事件进行了定量分级。在作业条件危险性评价法三项评价指标的基础上，根据煤矿险兆事件的特点增加两项指标，提出了用于煤矿险兆事件分级的f值法分级模型，运用五项风险指标七级评分计算实现分级。f值法适用于险兆事件的五项指标清晰、明显，能够直接进行分值定量测评的简单典型险兆事件进行量化分级。对描述模糊、过程复杂、无法用f值法进行测评的险兆事件，用权重分析系统和集对分析法构建了煤矿险兆事件分级评价向量矩阵和分级模型进行量化分级。f值法、权重分析系统和集对分析法使煤矿险兆事件的分级由定性分级实现了定量分级，为在实际工作中根据险兆事件的不同危险级别而区分轻重缓急采取不同管控措施提供了依据。 第三，运用案例推理检索技术方法，构建了煤矿险兆事件分级管控体系。提出了案例检索的关键过程和学习维护要点，建立了应用于煤矿险兆事件分级管控的“143”管控体系，构造了煤矿险兆事件搜集辨识、上报、分级和决策的流程，设计出基于案例检索技术的决策支持系统基本结构和五大模块。研发了基于安卓（Andriod）手机操作系统的“险兆事件管理系统”和基于电脑（PC）端的“险兆事件管理服务平台”，实现了某煤矿险兆事件管理系统对煤矿险兆事件较为有效的管控。 本研究深化了煤矿险兆事件管控理论，对煤矿险兆事件辨识上报、分级管控等方面的工作起到了推动作用，有利于促进煤矿安全管理水平的进一步提升。

There must be a Near-miss of any accident. Nowadays, the proneness that investigations of accidents are given more attention than pre-control in safety management leads a large number of Near-misses to injury accidents. This paper dedicates to the identification, report and grade control of near-miss, based on the study of relevant literature at home and abroad , to analysis the discriminates causing mechanism and the concept elements of Near-miss, and the difference and relationship between hazard，potential dangers, Near-miss, accidents. Focus on the identification and report, quantitative classification, control system construction and other key areas to carry out research of coal-mine events. First of all, on the basis of in-depth interview survey, to obtain the influence factors of the coal-mine Near-miss events active report willing, and then used G1 weighting method to weight every hierarchies’ indexes of active response.The Near-miss perceiving of 3 dimensions influence active report most. However, the clearness of definition influence active report most in 3-class index system. Structural equation model is used to make confirmatory factor analysis of influencing factors of active report will. discusses the effect of each variable on reporting. Secondly, the f-value method, weighting analysis system and set pair analysis method were used to carry out quantitative classification of coal-mine Near-misses. Based on the three evaluation indexes of working conditions evaluation index method, according to the characteristics of coal-mine Near-misses increased two indicators, A classification model of f-value method for coal-mine Near-misses events’ classification was proposed.and using the five risk indicators and seven score calculation methods to implement classification, The applicable conditions of f-value method are clearly and obviously .It also used to quantitative classified the easily typical Near-misses which can achieve the obvious quantitative score.Weighting analysis system and set pair analysis method construct the classification evaluation vector matrix and classification model of coal-mine Near-misses to those fuzzy described, complex process and for those can not evaluate with f-value method Near-misses to quantitative classification. The f-value method, weighting analysis system and set pair analysis method can make the classification of coal-mine Near-misses by fuzzy qualitative classification to the accurate quantitative classification. And provide foundations of taking different control measures according to different risk level to distinguish the order of priority in the actual work. Thirdly, this paper using the case retrieval technology constructs coal-mine Near-misses classification control system, proposes the key process, the maintenance and learning points of case retrieval, then the system of "143" management and control system of Near-miss is established. Constructed a coal-mine Near-miss incident collect identification, reporting, classification and decision-making processes, designed the basic structure and the five modules of decision support system based on case retrieval technology, developed two “Near-miss Management Services platform” terminals respectively on the Android-based mobile network and the PC, to achieved more effective control for the Near-miss management system. This research has deepened the coal-mine Near-miss events’ management and control theory, promoted its identification report, hierarchical control and other aspects of the work, and to further improve the coal-mine safety management level.