【摘要】 根据山西宁武煤田某煤矿2#煤层采掘及钻孔揭露情况,发现该井田中部和南部存在古河流冲刷带。通过对比基于优选的4个地震属性的四元一次、四元二次多项式回归模型与BP神经网络预测模型,决定将BP人工神经网络模型预测煤层厚度数据应用于整个测区古河流冲刷带的预判工作。首先利用GeoFrame系统和Landmark公司Poststack模块,提取2#煤层反射波的各类沿层切片,分析并圈定出2#煤层古河流冲刷带的大致范围,在此基础上,利用垂直时间剖面中2#煤层反射波的各种波形特征,进一步判别2#煤层古河流冲刷带解释的可靠性,然后结合BP神经网络预测模型获得的2#煤层厚度变化趋势图,最终解释出2#煤层古河流冲刷带范围:勘探区内2#煤层厚度变化范围0～5.3m,根据其煤层厚度变化趋势,将全区划分出一大二小3个古河流冲刷带。更多还原

【Abstract】 Based on the winning and opening of No.2 coal seam and drilling data of a coalmine in the Ningwu coalfield,Shanxi have found paleochannel scouring belts existed in central and southern parts of the minefield.Through comparison and based on quaternion linear,quaternion quadratic polynomial regression model and BP neural network prediction model of optimized 4 seismic attributes have determined to apply BP artificial neural network model predicting coal seam thickness data on whole prospecting area paleochannel scouring belts prediction.Firstly to use GeoFrame system and Landmark Company’s Poststack module,picked up various bedding slices of No.2 coal seam reflections,analyzed and delineated No.2 coal seam paleochannel scouring belts in the main,on this basis,to use various reflection waveform features on vertical time sections,to further distinguish interpretation reliability,then combined with No.2 coal seam thickness variation trend chart obtained from BP neural network prediction model and interpreted No.2 coal seam paleochannel scouring belt extent finally:No.2 coal seam thickness range of variation in the prospecting area is 0～5.3m,based on thickness variation trend,delineated one large and two small total three paleochannel scouring belts in the whole area.更多还原