基于SVM的石灰岩矿山碎石加工系统安全风险研究

发布时间：2018-08-23 07:54

【摘要】：随着机械自动化的迅猛发展，随之而来频发的附属设施系统事故给非煤矿山行业拉响了安全警报。碎石加工系统作为附属设施的重要组成部分。只有系统的分析风险、量化风险，找到行业存在问题，才能采取相应措施规避风险。本文通过作业分解树-风险分解树（WBS-RBS）方法进行危险源辨识，引用支持向量机（SVM）对碎石加工系统的安全风险因素进行分类，并提出SVM方法对碎石加工系统进行安全生产标准化等级进行评估。首先，在查阅碎石加工区域有关事故的基础上，根据安全法规和行业标准，结合现场调研及专家咨询，总结露天石灰岩矿山行业碎石加工系统的安全风险现状。其次，针对碎石加工系统相关机械设备及生产工艺流程，对碎石加工系统进行分类；结合作业分解树-风险分解树（WBS-RBS）方法进行危险源辨识，确定安全风险因素指标体系。再次，通过16个矿山企业危险源分类问卷调研并量化处理，在验证数据适用性后，任意选取10个矿山调研数据作为训练集，进行训练学习，其余的6个矿山数据作为测试集。经数据归一化处理，核函数选取，寻求最优参数c和g，得到基于SVM的碎石加工系统安全风险因素分类模型。经测试集测试验证后，得出碎石加工系统三种等级的危险源，，以此来判断风险因素对安全生产目标的影响程度。然后，结合危险源辨识结果和安全风险因素分类结果，建立碎石加工系统安全生产标准化等级评估指标体系。在此基础上建立调查要素，对目前现有的已取得不同标准化等级的露天石灰岩矿山碎石加工系统进行调研，根据不同标准化等级企业的安全生产现状，对各指标要素投入到位情况进行赋值，以得到训练样本。将训练集训练学习，构造SVM碎石加工系统安全生产标准化等级评估模型。选任一石灰岩矿山企业碎石加工系统进行测试，得到识别测试集输出的结果，即该石灰岩矿山企业碎石加工系统所处的安全生产标准化等级。最后，经测试验证，建立的SVM安全风险评估模型与矿山企业实际现状有较高的符合性。得出的高等级风险因素，有利于指导隐患排查与教育培训。SVM不失为一种好的安全风险评价方法，并带来一种新的安全生产标准化等级评定思路。
[Abstract]:With the rapid development of mechanical automation, the frequent accidents of auxiliary facilities alarm the non-coal mine industry. Gravel processing system is an important part of ancillary facilities. Only by systematically analyzing risks, quantifying risks and finding problems in the industry, can we take corresponding measures to avoid risks. In this paper, hazard source identification is carried out by job decomposition tree-risk tree (WBS-RBS) method, and support vector machine (SVM) is used to classify the safety risk factors of gravel processing system. The SVM method is put forward to evaluate the standardized grade of production safety of gravel processing system. First of all, on the basis of referring to the related accidents in the gravel processing area, according to the safety regulations and industry standards, combined with field investigation and expert consultation, the paper summarizes the current situation of safety risk of gravel processing system in open-pit limestone mining industry. Secondly, according to the related mechanical equipment and production process of gravel processing system, classification of gravel processing system is carried out; combined with job decomposition tree-risk decomposition tree (WBS-RBS) method, hazard source identification is carried out, and safety risk factor index system is determined. Thirdly, through the investigation and quantification of 16 mine enterprises' hazard source classification questionnaire, after verifying the applicability of the data, 10 mine survey data are chosen as the training set, and the remaining 6 mine data are used as the test set. After normalized data processing and kernel function selection, the optimal parameters c and g are obtained, and the classification model of safety risk factors of gravel processing system based on SVM is obtained. After testing and verification of the test set, three kinds of dangerous sources of gravel processing system are obtained to judge the influence of risk factors on the target of production safety. Then, combining the results of hazard source identification and the classification of safety risk factors, a standardized evaluation index system for production safety of gravel processing system is established. On this basis, the investigation elements are established, and the existing open-pit limestone mine gravel processing systems with different standardization grades are investigated, according to the safety production status of enterprises with different standardization grades. The input of each index element is assigned to get the training sample. The SVM gravel processing system safety standardization grade evaluation model is constructed by training and learning the training set. The lithotripsy processing system of a limestone mine enterprise is selected for testing, and the output result of the identification test set is obtained, that is, the standard grade of safety production of the lithotripsy processing system in the limestone mine enterprise. Finally, the SVM safety risk assessment model is proved to be in good agreement with the actual situation of mining enterprises. The high grade risk factors are helpful to guide hidden trouble detection and education training. SVM is a good method of safety risk evaluation and brings a new way of safety production standardization evaluation.
【学位授予单位】：重庆科技学院
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TD79;TD921.2

【参考文献】