摘要
在淮北杨柳矿区,层状或脉状侵入的岩浆岩导致主采煤层呈现较大范围的吞蚀、烧变、焦化、变薄现象,而常规地震时间剖面及顺层层拉平切片上难于确定其分布范围。为了精细确定主采煤层中岩浆侵入的影响范围,基于煤层反射波振幅、波形、能量的同一性及差异性特征,利用优化后的地震多属性无监督神经网络聚类分析技术对其进行分析。解释结果表明,3煤层未受到岩浆侵入,而10煤层岩浆侵入比较明显,其在成果图上呈现出大范围的黄色团块状及脉状特征。与实际钻孔揭露资料对比,利用地震多属性聚类分析技术圈定的岩浆岩分布范围符合率达78%,具有较高准确度。
In the Yangliu mine area, Huaibei, layered or veined intrusive magmatic rocks causing main mineable coal seams present phenomena of eroding, burning, carbonizing and thinning to a relatively great extent, thus their distribution ranges are hard to define on conventional seismic time sections and bedding layer-flattened slices. To define magmatic intrusion impacting area accurately, based on coal seam reflection wave amplitude, waveform, energy identity and discrepancy characteristics, using seismic muhiple attributes non-supervising neural network clustering analysis technology carry out analysis investigation. Interpreted results have demonstrated that the No.3 coal seam has no magmatic intrusion, while the No.10 coal seam obviously intruded, on its interpretation map has presented large extent yellow lumpy and veined features. Comparison with drilling revealed data, the coincidence rate can be 78% when using seismic multiple attributes clustering analysis technology to identify and determine magmatic intrusion, thus rather high accuracy.
出处
《中国煤炭地质》
2012年第2期64-66,共3页
Coal Geology of China
关键词
地震多属性
岩浆岩
无监督神经网络聚类分析
seismic multiple attributes
magmatic rock
non-supervising neural network clustering analysis
