基于模糊C均值聚类算法的矿井煤炭自燃预测

Prediction of Coal Spontaneous Combustion in Mine Based on Fuzzy C Means Clustering Algorithm

为实现矿井煤炭自燃预测控制,对煤层自燃原因进行分析,得到煤炭含水量、发热量、含硫量、倾斜角和地质构造以及煤层瓦斯含量共6种自燃因素,首先将某煤矿历年来的数据作为原始数据,再通过模糊C均值聚类的方法对这些原始数据进行分类,建立标准模型库,再将新的数据通过模糊模式识别中贴近度的方法与模型库相匹配,预测矿井待开采煤层自燃的可能程度,针对不同的程度选择不同的开采方案以及控制煤炭自燃的方案,使煤矿安全生产得以保障。

To realize the prediction control of coal spontaneous combustion in mine,we analyze the reasons and get a total of 6 kinds of factors including coal moisture content,calorific value,sulphur content,angle of inclination,geological structure and gas content of coal seam. Through data selection over the years in a mine,we classify the data by fuzzy C means clustering algorithm and build standard model base. We match the new data with model base by close degree of fuzzy pattern recognition method,predict spontaneous combustion degree of the mining coal seam,and choose a different mining scheme to control coal spontaneous combustion and ensure safety production according to different degree.