【作者】 王爽英；

【导师】 吴超；

【作者基本信息】 中南大学 ， 安全管理工程， 2012， 博士

【摘要】 本文综合应用安全科学和系统科学的相关理论和方法,对中小型煤矿的生产区域安全度、生产安全的评价指标体系和综合评价方法、生产安全的投入产出效率和投入决策模型等方面进行了较系统的研究。(1)研究了中小型煤矿区域安全度和系统安全度。阐述了影响煤矿生产安全的危险因素,界定区域安全度和分析安全度和安全等级之间的关系,提出单一区域和组合区域的安全度计算方法,同时结合某煤矿的实际情况进行了实证分析。(2)构建了中小型煤矿生产安全水平评价指标体系。探讨了系统安全评价理论,针对目前我国煤矿生产安全评价体系存在的主要问题和不足,确定了一套较为全面且简单实用的指标体系,从生产区域安全、人员安全、安全管理、安全事故四个方面建立了一个三层22个评价指标的指标体系,并对各评价指标的具体含义进行说明。(3)综合评价了中小型煤矿的生产安全水平。运用主、客观的综合赋权法确定中小型煤矿生产安全水平评价指标的权重,运用灰色层次分析理论、模糊综合评价法、集对分析理论三种评价方法对中小型煤矿进行综合评价,并进行实证分析。针对三种评价方法的优劣进行分析比较。(4)从生产安全水平投入、产出的角度,利用数据包络(DEA)模型对中小型煤矿生产安全投入与产出的相对有效性进行研究。以第三章建立的中小型煤矿生产安全水平评价指标体系为基础,建立基于DEA的投入和产出评价指标体系,并运用CCR模型对15家中小型煤矿的相对有效性进行了实证研究。(5)分析了生产安全投入决策的常用方法,主要包括德尔菲法、头脑风暴法、教育交流法等定性决策,以及决策树法、贴现现金流法、实物期权法等定量方法,说明这些方法在生产安全投入决策中的具体应用,并比较分析方法的优劣。另外,说明了多种人工智能技术在煤矿生产安全决策领域的应用状况。(6)建立了基于BP网络的中小型煤矿生产安全投入决策模型,决策模型以区域安全度、事故发生率和年经济损失金额比例三项指标为模型输入数据,以生产安全投入的十项指标为输出指标,利用MATLAB对DEA有效的10个决策单元为训练样本,对网络进行训练。运用训练好的神经网络模型对5个无效的中小型煤矿进行生产安全投入的测算,通过对原数据和测算数的比较,得出需调整的主要投入指标：一是减少员工“三违”比例和降低矿工的流动性,二是完善安全管理机构组织和提高安全工程技术投入,三是提高矿工的持证率和安全教育的投入。通过上述的研究,建立了中小型煤矿生产安全水平评价指标体系,并进行综合评价,研究了中小型煤矿生产安全投入相对效率,并对其生产安全投入决策问题进行实证分析,形成了一套较为系统的中小型煤矿生产安全水平评价研究体系。更多还原

【Abstract】 Taken the relevant theories and methods from safety science and systems science as the foundation and on the middle&small scale coal mines, the production region safety degree, production safety evaluating indexes system, comprehensive evaluation methods, input&output efficiency and input decision-making models are systematically researched. The main contents are as follows:(1) By analysing region and system safety degree of the small&middle scale coal mines, the danger factors which cause the changes of the coal mine safety are clarified, the region safety degree is defined and the relationship between safety degree and safety level is explicited. Meanwhile safety degree’s algorithm for single region and combined region in production are proposed, and the empirical analysis is conducted.(2) For building production safety level evaluating indexes system of middle&small scale coal mine,and on the basis of studying the theories and methods of coal mines’ production safety evaluation and analysing the defects of present evaluating indexes system, a three-storys’ production safety level evaluating indexes system for middle&small scale coal mines is built, which contains22sub-indexes, all of which are based on respective4aspects, namely, production region safety degree, staff’s safety, safety management, safety accidents. Moreover, all evaluating indexes are illuminated by the specific definition.(3) For evaluating production safety level of middle&small scale coal mines,and on the basis of calculating the weight of production safety evaluation indexes of each index by applying integrate subjective with objective method, a comprehensive evaluation of middle&small scale coal mines with3evaluation methods, namely, grey analytic hierarchy process, fuzzy comprehensive evaluation method and set pair theory is conducted. Additionally, the advantages and disadvantages of three evaluation methods are analysed and compared.(4) From the perspective of input&output, the relative effectiveness of middle&small scale coal mines production safety level was studied by applying the data envelopment analysis(DEA) model. Based on the safety production evaluating index system of middle&small scale coals mine built in chapter3, an empirical research on the effectiveness of15coal mines was conducted by applying CCR model.(5) Common production safety input decision methods of mine coals are researched, which mainly include qualitative decisions, such as Delphi method, brainstorming methods, education and communication method, and quantitative methods, such as decision tree method, discounted cash flow method and real option method. Besides, application of these methods in coal mine production safety is indicated. Furthermore, the advantages and disadvantages of these methods are analysed and compared. In additions, the paper application status of a variety of artificial intelligence technology in coal mine production safety decision is clarified.(6) A safety input decision-making model of coal mine production is built, which is based on BP neural network. Input data of decision-making model include3indexes, namely, regional safety degree, accident incidence and proportion of annual economic loss. Besides, the model usess10indexes concerning each input of production safety as output indexes. By using the neural network toolbox based on MATLAB, the model applies the10decision making units that are effective to DEA to train the network, and the input in production safety of5ineffective coal mines is measured and calculated by applying the well-trained neural network. By comparing the original data with calculated data, a few input indexes which should be adjusted are as follows:firstly, reducing the ratio of "3violation" of staffs and mobility of miners; secondly, improving the organization of production safety and increasing safety projects and technology input; thirdly, enhancing the percentage of licensed miners and safety education input and so on. Through the above research, the evaluating indexes system of middle&small scale coal mine production safety level is built, and safety level is evaluated. Besides, the input relative efficiency of middle&small scale coal mine production safety is investigated by applying DEA model. Morever, an empirical analysis on the issue of decision-making of production safety is conducted. Therefore, a set of coherent production safety level evaluating research system of middle&small scale coal mine is formed.更多还原