矿山瓦斯灾害的空间数据挖掘.doc

矿山瓦斯灾害的空间数据挖掘 【中文摘要】矿山瓦斯灾害时刻迫害着矿工性命财产安全,国度矿山安全监察机构加大了安全检讨力度,煤矿出产单位也越来越重视对瓦斯的安全监测,树立了矿山环境监测体系。灾害的预防优于灾害的管理。因此,本文主要是应用空间数据挖掘的方式,从矿山瓦斯空间数据库中提取未知的、有效的和可操纵的知识,预测瓦斯灾害未来发生的可能性。本文通过在已有的矿山环境监测体系数据库2006年上半年数据中,提取瓦斯监测的数据,应用已有的CAD矿图,在ArcGIS中入行数字化,使用ArcGIS的联系关系功能把数字化的传感器位置与瓦斯数据库入行联系关系,运行专门编制的数据库查询软件,可以查询任意传感器任意时光段的数据,并且绘制相关的瓦斯数据折线图。在BP神经网络的基本上,使用MATLAB树立矿山瓦斯灾害预测模型,应用BP神经网络的自我学习和调整神经元权值的特色,可以不斟酌庞杂的影响因子而直接树立的相关的模拟模型,通过对比得出模型模拟本月的精度可以到达99.3,通过学习前3个月的数据,可以较好的预测未来的瓦斯凸起情形。;【Abstract】 The safety and property of miners was often endangered by the mine gas disaster. Thus, the State Administration of Coal Mine Safety checks the security increasingly, and the coal mine pays more and more attention to monitor the gas emission. They have established the mine environment monition system. Preventing the disaster would be better than controlling the disaster. So the main purpose in this *** is that we can use the knowledge to forecast the possibility of gas disaster with of spatial data mining. The knowledge which is thought unknown, effective and operative, which can be extracted from the mine gas spatial databases.In this ***, we extract the gas monitor database from the environment monition system which is from first half year of 2006, and then digitalize the CAD map with the software of ArcGIS. At last, we use the connection function to link the digitized sensing position and the gas database so that we can get the spatial database. After that, we can get the anytime of the sensor data and make chart with the inquire software.Finally, the *** uses the BP neural network to establish the mine gas disaster forecast model in MATLAB. Because the BP neural network has the character of self-study and the adjustment neuron weight, we may not consider the complex influence factor, but directly establishing the model. The precision between the model simulation and the true data could be possible to achieve 99.3. If we want to forecast future gas situation better, we need to research the first 3 month-long data.