基于K-means算法的采空区“三带”划分

"Three Zones"Division of Goaf Based on K-means Algorithm

采空区“三带”范围的划分可以有效防治煤自燃。为进一步解决不同区域的煤自燃“三带”划分存在差异性的问题,本文构建了基于K-means算法的采空区“三带”划分模型。基于东欢坨煤矿3015工作面的束管数据构建数据集,引入三种聚类算法和六个聚类评价指标,建立三个采空区“三带”划分模型,通过分析模型效果选出最优聚类模型。结果表明,K-means模型聚类效果最好,模型评价指标轮廓系数、CH指数和DH指数分别为0.573、121.291和0.47,基于氧气的采空区“三带”划分结果为4.435-20.294 m。本文所构建的多参数、多算法、多组合的煤自燃采空区“三带”划分模型精确度高,适用性强,可以针对不同环境的煤自燃三带进行划分,进而保障煤矿开采的安全性。

The division of the"three zones"in goaf can effectively prevent coal spontaneous combustion.To further address the issue of differences in the division of coal spontaneous combustion"three zones"in different regions,this paper constructs a"three zone"division model for goaf based on the K-means algorithm.Based on the bundle data of the 3015 working face in Donghuantuo Coal Mine,a dataset was constructed.Three clustering algorithms and six clustering evaluation indicators were introduced to establish three"three belt"division models for goaf.The optimal clustering model was selected by analyzing the model's effectiveness.The results showed that the K-means model had the best clustering performance,with evaluation indica⁃tors such as contour coefficient,CH index,and DH index of 0.573,121.291,and 0.47,respectively.The results of the"three zones"division of goaf based on oxygen were 4.435-20.294 meters.The multi parameter,multi algorithm,and multi combina⁃tion coal spontaneous combustion goaf"three zone"division model constructed in this article has high accuracy and strong ap⁃plicability,and can be divided for different environments of coal spontaneous combustion three zones to ensure the safety of coal mining.