【作者】 杨军；

【导师】 宋学锋；

【作者基本信息】 中国矿业大学 ， 管理科学与工程， 2013， 博士

【摘要】 煤矿的安全问题已经成为煤矿可持续发展的头等大事，而如何提前发现并判别潜在的安全风险是煤矿安全问题的重中之重。本文在理论分析与实际调研的基础上，主要解决两个问题。一是解决安全风险的评价问题，即判断煤矿目前是否存在安全风险，主要用于定期判断煤矿目前的安全风险状况。二是解决安全风险的预警问题，即如何通过一定的预警指标与预警区间进行日常性的安全风险判断与提示。首先，论文在对相关安全风险评价与预警理论综述的基础上，主要研究用于煤矿的安全风险评价原理与方法以及安全风险预警理论与方法，为后续研究提供理论与方法基础。论文对目前常用的风险评价方法与预警方法进了全面的总结，分析了这些方法的优点与确定，并总结了对煤炭企业安全风险评价与预警的适用性。对方法的总结力求全面与详细，为后文方法的选择提供帮助。另外，对风险评价与预警的理论也进行了一定的梳理，归纳出能够用于煤矿风险评价与预警的共性理论。然后，根据煤矿安全风险产生的特殊性，分析了选用安全风险评价、预警理论与方法所遵循的选择，并依据扎根理论，构建了煤矿安全风险评价体系与预警系统的建立步骤。第二，对煤矿安全生产风险评价指标体系与权重进行设计与确定。在分析煤矿生产的复杂性、动态变化性、随机性、灾害事故的关联性等风险特点的基础上，从科学性、系统性、全面性、易评价性、定性与定量相结合、独立性等六个方面提出了煤矿安全生产风险评价指标设计的原则。并在典型案例分析、文献研究的基础上，通过进行访谈和调研，在广泛征求专家意见的基础上，基于“人、机、物、环、法”的理论框架，建立了人力资源配置、规章制度及执行、作业环境、地测及防治水、一通三防、防止煤与瓦斯突出、采掘系统、机运系统、煤矿固有风险等九大类煤矿安全生产风险评价指标体系，并进一步进行了细化为105个具体指标。另外，在此基础上，进行了煤矿安全生产风险评价指标权重确定。主要对整个安全风险评价体系中权重进行计算，并归纳确定安全风险评价标准。依据调研问卷，使用网络层次分析法进行各个安全风险评价指标权重的确定。并在权重确定的基础上，建立了安全风险评价指标的评价标准。第三，构建了煤矿安全风险预警指标体系与预警机制。为了达到管理层日常安全管理的目的，论文依据安全风险评价指标与调研结果，使用扎根理论对安全评价指标进行遴选，确定安全风险预警指标体系，并依据聚类分析与判别分析确定安全预警单一指标预警区间与综合预警判别函数。首先从人力资源方面、安全投入效率与安全管理制度实施效率等三个方面建立衡量安全风险的预警指标体系；其次依据安全风险预警指标建立针对单一指标的预警区间，确定正常、轻警、中警与重警四个预警区间的临界值；再次使用FISHER判别模型构建安全风险综合预警判断函数，依据函数与煤矿现实指标状况判断煤矿目前所处的安全风险警度；最后依托煤矿安全风险预警体系，建立了煤矿安全风险预警判别与应对机制。第四，归纳煤矿安全风险预警的具体方法，并构建用于安全风险预警的配套机制与管理系统。主要归纳了适用于煤矿安全风险预警的具体方法，是以预警指标与预警区间作为基础的，并在此基础上，构建了与安全风险预警体系相配套的实施机制，包括在人力资源方面、安全投入方面与安全制度构建等方面的实施机制。同时，还构建了用于实施安全风险预警体系的信息管理系统。第五，选择袁店煤矿作为研究对象，具体对煤矿安全生产风险预警指标体系进行测试。主要将安全风险预警指标体系应用到典型煤矿应用，分析其适用性。具体根据该煤矿2011年的数据计算得到安全风险预警指标的具体数值，对应单一指标的预警区间，分别判断该矿在人力资源、安全投入与安全制度三个方面所处的警度。然后，对综合风险的测试从两个方面着手进行。一方面，使用FISHER判别模型，将具体预警指标代入，具体得到目前安全风险的总体警度。该方法能够比较直观地判断安全风险的警度状态。但该方法也存在一定的弊端，FISHER判别模型需要进行定期更新，使得模型能够非常接近于现状。另一方面，使用粗糙集进行经验衡量，具体依据对单项指标警度的判断，联合判断总体安全风险的警度。在实际应用中，一般将两种方法综合起来使用，以期达到更好的预警效果。更多还原

【Abstract】 The safety problem of coal mine has become the major event of coal minesustainable development, and how to discover and distinguish the potential safety riskis the most important of coal mine safety problem. This thesis mainly solves twoproblems on the basis of theoretical analysis and practical research. First, to solve theappraisal problem of security risk, namely judging whether coal mine has the securityrisk. This is mainly used in the regular judgment coal mine current security riskcondition. Second, to solve the early warning problem of security risk, namely howcarries on the daily security risk judgment and prompt through certain early warningtarget and early warning sector.First of all, the thesis mainly researches coal mine security risk appraisalprinciple and the method as well as the security risk early warning in the foundationwhich summarizes the correlation security risk appraisal and the early warning theory,and provides the theory and method foundation for the following research. The papercarries on the comprehensive summary to the commonly used risk assessment methodand the early warning method, analyses the advantages and disadvantages of thesemethods, and summarizes the applicability of the coal enterprise security riskassessment and early warning. Moreover, the thesis carries on certain combing to thetheory of risk assessment and early warning, and summarizes the common theory thatcan be used in the coal mine risk assessment and early warning. Then, according tothe particularity of the coal mine safety risk, analysis of the choice followed by theselection of security risk assessment as well as early warning theory and method, thethesis constructs coal mine safety risk assessment system and the early warningsystem based on grounded theory.Second, the thesis carries on the design and determination to the coal mine safetyproduction risk evaluating indexes system and weight. On the basis of analysis of coalmine production complexity, dynamics, randomness, relevance of disaster risk, thecoal mine safety production risk assessment indicators were proposed in combinationfrom scientific, systematic, comprehensive, easy evaluative, qualitative andquantitative. And on the basis of typical case studies, literature studies, carrying oninterviews and research. Based on "people, machines, materials, environmental, law"theoretical framework, the thesis establishes the human resources disposition, therules and regulations and the execution, the work environment, the surveying and guards against regulates waterways, NBC, prevents the coal and the gas prominent,the excavation system, the opportunity system, the coal mine inherent risk and so onnine big kind of coal mine safety production risk assessment target system on thebasis of extensive consultations with expert advice. And further refinement to105specific indicators. Moreover, in this foundation, the thesis carries on the coal minesafety risk evaluation index weight determination, the major heavy weights in theentire security risk evaluation system is calculated and summarized to determine thesafety risk evaluation criteria. According to the investigation and study questionnaire,uses the network analytic hierarchy process to carry on the determination of eachsecurity risk evaluation criterion weight. And based on weight determination,establishes the evaluation criteria of security risk evaluation standard.Third, the thesis constructs the coal mine safety risk early warning indicatorsystem and early warning mechanism. In order to achieve the purpose of managementdaily safety control, according to security risk evaluation indexes and investigationresults, papers use grounded theory on safety evaluation index for selection todetermine security risk warning index system, and to determine safe single indexwarning interval and comprehensive early warning discriminant functions based oncluster analysis and distinction analysis. First, from the efficiency of investment inhuman resources aspects, security and safety management system implementationefficiency to establish a measure of security risk warning index system; Secondly,based on security risk early warning indicators to establish early warning interval fora single indicator, to determine the critical value of the normal, mild early warning,moderate and severe warning four warning interval; Thirdly, using FISHER todistinguish the model construction security risk synthesis early warning estimatefunction once more, according to function and coal mine realistic target to estimatethe situation coal mine at present security risk police; Finally, depends on coal minesafety risk early warning system to establish the coal mine safety risk early warningdistinction and response mechanism.Fourth, the thesis summarizes the specific method of coal mine safety riskwarning, and builds the supporting mechanisms and management systems for earlywarning of security risks. The thesis summarizes the specific methods applicable tocoal mine safety risk warning, and builds a security risk early warning systemcoordinated implementation mechanisms, including in the area of human resources,investment in safety and security system and other aspects of the implementation mechanism on the basis of early warning indicators and early warning interval. At thesame time, and also to build the information management system for theimplementation of the security risk early warning system.Fifth, the thesis chooses Yuan shop coal mine as the object of study, conducts theconcrete test to the coal mine safety production risk early warning indicator system,mainly applies the security risk early warning target system the typical coal mine, andanalyzes its serviceability. According to the mine data in2011to calculate the safetyrisk early warning index of concrete numerical value, corresponding to a single indexearly-warning interval, respectively determine the mine in human resources, safetyinput and safety alert degree of the three aspects of the system. Then, the thesis teststhe integrated risk from two aspects. On the one hand, the use of FISHER’sdiscriminant model, will plug in specific warning index to get the overall alert degreeof security risks. This method can intuitively judge the alert degree of the security riskof the state, but also has certain drawbacks. FISHER discriminant model needs to beupdated regularly, so that the model can be very close to the status quo. On the otherhand, uses the random to carry on experience to weigh, judging the overall securityalert degree of the risk based on the judgment of the accuracy of individual indicators.In practical applications, the two methods are combined to use, in order to achievebetter warning effect.更多还原