基于多属性变换的煤田波阻抗反演应用研究

英文篇名：Research on Coalfield Acoustic Impendence Inversion Based on Multi-attribute Transform
中文刊名：工程地球物理学报
英文刊名：Chinese Journal of Engineering Geophysics
作者：李娟娟 ; 崔若飞 ; 潘冬明 ; 胡明顺 ; 张应华
英文作者：Li Juanjuan 1 ; 2 ; 3 ; Cui Ruofei 1 ; 2 ; Pan Dongming 1 ; 3 ; Hu Mingshun 1 ; 3 ; Zhang Yinghua 4(1.School of Resources and Geosciences ; China University of Mining and Technology ; Xuzhou Jiangsu 221116 ; China ; 2.Coalbed Methane Resources and Reservoir Formation Process ; Key Laboratory of the Ministry of Education ; Xuzhou Jiangsu 221116 ; China ; 3.ChinaState Key Laboratory of Geomechanics and Deep Underground Engineering ; Xuzhou Jiangsu 221116 ; China ; 4.Yuannan DA RUI coal group Co ; LTD ; Xiaohaizi colliery ; Qujing Yunnan 655813 ; China)
中文关键词：多属性变换 ; 多属性线性回归 ; 径向基神经网络
英文关键词：multi-attributes transform ; multi-attributes linear regression ; radical basis neural network
出版日期：2012-11-30
机构：中国矿业大学资源与地球科学学院;煤层气资源与成藏过程教育部重点实验室;深部岩土力学与地下工程国家重点实验室;云南大瑞煤业集团有限公司小海子煤矿;
年：2012
期：06
出版单位：工程地球物理学报

摘要

随着煤炭资源勘探深度的日益增加,复杂的地质条件对岩性勘探提出了新的要求,常规的反演技术捉襟见肘。本文避开了以褶积模型为基础的常规反演,从基于地震属性技术的反演技术出发,选择利用多元线性回归技术和径向基神经网络技术进行反演研究。论文以山西阳泉某矿地震勘探资料为例进行反演计算,计算研究结果表明,基于多属性变换的反演技术效果优于常规的反演方法,在煤田反演中可以得到广泛的推广。这就为煤田岩性地震勘探提供了一项有效的技术,有很好的应用前景。
With the increasing exploration of coal resources,complex geological conditions set the new demands in lithological exploration,for the conventional inversion methods have many problems.The paper chooses multiple linear regression and radical basis neural network techniques to do some inversion research based on seismic attributes technique.And the results from the computation of coalfield Yangquan Shanxi show the inversion based on multi-attributes technique has better effects,so it can be widely applied to coalfield exploration.

引文

[1]姚逢昌,甘利灯.地震反演的应用与限制[J].石油勘探与开发,2000.27(2):53～58.
    [2]Russell B,D Hampson,J Schuelke,et al.Multiat-tribute seismic analysis[J].The Leading Edge,1997,16(10):1439～1443.
    [3]Hampson D,J Schuelke,J Quirein.Use of multiat-tribute transforms to predict log properties from seis-mic data[J].Geophysics,2001,66(1),220～236.
    [4]Specht,Donald.Probabilistic neural networks[J].Neural networks,1990,2:109～118.
    [5]Somanath Misra,Satinder Chopra,John Zhang.Neuralnetwork regression analysis and post-stack inversion-A comparison[C].SEG,2010,1473～1477.
    [6]Brian Henderson Russell.The application of multiva-riate statistics and neural networks to the predictionof reservoir parameters using seismic attributes[D].University of Calgary,2004.