基于数据挖掘与信息融合的瓦斯灾害预测方法研究

发布时间：2018-03-23 20:43

本文选题：信息融合　切入点：数据挖掘　出处：《中国矿业大学（北京）》2013年博士论文

【摘要】：煤炭资源一直以来作为重要的能源来源，在生活和工业中的需求日益增加，而煤矿的开采因为其特殊的环境因素而有一定的危险性，尤其是煤矿瓦斯灾害作为煤矿首要的恶性事故，其发生频率高，破坏性广，社会影响大。长期以来，人们一直是以瓦斯的检测为主，无法提前预测瓦斯灾害的发生。本文综合阐述了煤矿瓦斯监测的主要问题所在，讨论了预测灾害发生的重要性和可能性。通过对监测数据资源进行数据挖掘、分析，采用支持向量机、模糊集等理论建立多源、多平台、多传感器煤矿瓦斯灾害在线辨识、隐患判别和决策模型；攻克了瓦斯监测系统不能及时发现重大瓦斯灾害隐患的关键技术。为了解决瓦斯灾害在线辨识中的不确定性和不精确性的问题，分别建立了基于Bayes network和D S证据理论的煤矿瓦斯灾害特征级融合模型及算法。并对这两种模型进行实验研究，检验这两种模型的有效性。提出利用多种监测信息进行“数据挖掘”、分析、处理、融合和综合判断瓦斯灾害危险性的先进理论和方法，为解决预测问题提供了理论依据和科学方法，具有重要的实际意义。
[Abstract]:As an important source of energy, coal resources have been increasing demand in daily life and industry, and coal mining is dangerous because of its special environmental factors. Especially, the gas disaster in coal mine is the most serious accident in coal mine, its occurrence frequency is high, destructive is wide, social influence is big. For a long time, people always take gas detection as the main factor. It is impossible to predict the occurrence of gas disaster in advance. This paper comprehensively expounds the main problems of gas monitoring in coal mine, and discusses the importance and possibility of predicting disaster. Based on support vector machine and fuzzy set theory, the online identification, hidden trouble discrimination and decision model of gas disaster in multi-source, multi-platform and multi-sensor coal mine are established. In order to solve the problem of uncertainty and inaccuracy in online gas disaster identification, the key technology that gas monitoring system can not find the hidden danger of major gas disaster in time is overcome. The characteristic level fusion model and algorithm of coal mine gas disaster based on Bayes network and DS evidence theory are established, and the experimental research on these two models is carried out. To test the validity of these two models, the advanced theory and method of "data mining", analyzing, processing, merging and synthetically judging the hazard of gas disaster using various monitoring information are put forward. It provides theoretical basis and scientific method for solving prediction problem and has important practical significance.
【学位授予单位】：中国矿业大学（北京）
【学位级别】：博士
【学位授予年份】：2013
【分类号】：TD712;TP202

【相似文献】