多元统计分析在矿井断层预测中的应用

摘要

矿井小构造预测是一个难度极大的研究课题,根据岩层厚度、岩层组合关系,运用判别分析、地质统计学等数学方法预测矿井小构造,是一种全新的思想方法.本次研究选取我国西部某一煤矿作为研究对象,研究主采煤层上下岩性变化特征,根据岩石的不同抗压强度,确定五个变量;将已采区的钻孔分为含有断层的钻孔和不含有断层的钻孔,统计这两类钻孔中五个变量的值,建立判别函数,对未采区的钻孔进行判别分类.对每个钻孔的岩层在可采煤层的上下划分成10个层段,判别出每个层段中未知钻孔的归属,从而得到钻孔属于有断层的概率,即断层分布的概率等值线图.经实例验证,精确度达70％以上,对生产矿井具有一定经济效益和实用价值.%The prediction of coalmine minor structure is an extremely difficult research subject, according to strata thickness, strata associative relations, using mathematical approaches of discriminating analysis and geostatistics to predict coalmine minor structures is a kind of bran-new method of thinking. This study has selected a coalmine in western China as the subject investigated, to study lithologic variation characteristies of strata above and below the main mineable coal seams. Based on different compressive strengths of rocks have determined 5 variables. Divided boreholes in mined district into fault intersected and non fault intersected, counted up values of the 5 variables of two kinds of boreholes and established discriminating functions, carried out discriminating classification for boreholes in non mined district. Divided strata above and below mineable coal seams into 10 sectors, distinguished unknown borehole attribution of each sector, and so obtained probability of borehole fault intersecting, drawn up fault distribution probability isogram. Examples have demonstrated that the accuracy rating can be above 70%, thus the method has certain economic effect and practical value.