Predicting Models of Coal Thickness Based on Seismic Attributions and Their Applications

Abstract

Abstract As a technology widely used in lithology and structure interpretation, seismic attribute technology has been playing an important role in coal and oil exploration. Based on 3‐D seismic exploration data of coal seam 13‐1 in the Xieqiao colliery, Huainan coal field of China, 28 seismic attributes are extracted. Through analysis of seismic attributes, four usable seismic attributes, including average peak amplitude, kurtosis in amplitude, maximum absolute amplitude and slope rate of instantaneous frequency, are selected as the basic parameters of predicting models of coal thickness. Combined with the real drill data, the predicting models between coal thickness and multi attributes are established by using analytical methods of multivariant polynomial regression and BP neural networks (BPNN), and error analysis of predicting coal thickness is carried out. From the comparison of predicting results of coal thickness of coal seam 13‐1 in the Xieqiao colliery, Huainan coal field of China by using these two models, it is shown that BPNN model has higher accuracy in predicting coal thickness.