基于SVR和地震属性的构造煤厚度定量预测
详细信息 本馆镜像全文    |  推荐本文 | | 获取馆网全文
摘要
为了提高谱分解和甜面属性组合预测构造煤厚度的精度和可靠性,利用回归型支持向量机(SVR)的非线性处理能力,将SVR和地震属性相结合,研究采区构造煤厚度定量预测方法。在SVR预测模型建立时,以正演模型数据为基础,通过训练和测试获得SVR预测模型的主要参数;结合井旁道数据,建立了采区构造煤厚度SVR预测模型。通过输入实际谱分解属性和甜面属性,定量预测了构造煤厚度。相对于传统地震属性预测来说,本次所预测的构造煤厚度精度较高、误差较小。当核函数类型为径向基核函数、输入为谱分解属性和甜面属性时,预测模型的预测效果最好。由于模型建立时未考虑地震资料信噪比的影响,预测模型不能克服其造成的不确定性。
In order to improve the precision and reliability of thickness prediction,the authors combined the Support Vector Regression( SVR),which is good at dealing with nonlinear issues,and seismic attributes together to build a model for the prediction of tectonic coal seam thickness. During model building,the authors,first of all,optimized the key parameters of SVR model through training and testing forward model's data,then,combined those optimized parameters with in situ near-well traces to build a SVR predicting model. Through inputting the real attributes of spectral decomposition and sweetness into this model,the authors achieved a prediction of tectonic coal seam thickness in the study area. By comparison with true thickness at wells,the thickness prediction of the model has a higher precision and lower absolute error than that with the prediction using traditional seismic attributes. Setting RBF( Radial Basis Function) as kernel and spectral decomposition attributes and sweetness attribute as inputs,the model generates its best results. Since the influence of signal-to-noise ratio of seismic data was not considered during the model development,the model could not overcome its corresponding uncertainty.
引文
[1]张子敏.瓦斯地质学[M].徐州:中国矿业大学出版社,2009.
    [2]张玉贵,张子敏,曹运兴.构造煤结构与瓦斯突出[J].煤炭学报,2007,32(3):281-284.Zhang Yugui,Zhang Zimin,Cao Yunxing.Deformed-coal structure and control to coal-gas outburst[J].Journal of China Coal Society,2007,32(3):281-284.
    [3]Shepherd J,Rixon L K,Griffiths L.Outbursts and geological structures in coal mines:A review[J].International Journal of Rock Mechanics and Mining Sciences&Geomechanics Abstracts,1981,18(4):267-283.
    [4]琚宜文,王桂梁.煤层流变及其与煤矿瓦斯突出的关系——以淮北海孜矿为例[J].地质评论,2002,48(1):96-105.Ju Yiwen,Wang Guilang.Rheology of coal seams and their relation with gas outbursts:A case study of the Haizi coal mine,Huaibei coalfield[J].Geology Review,2002,48(1):96-105.
    [5]孙学凯,崔若飞,毛欣荣,等.联合弹性波阻抗反演与同步反演确定构造煤的分布[J].煤炭学报,2011,36(5):778-783.Sun Xuekai,Cui Ruofei,Mao Xinrong,et al.Elastic impedance inversion associated with simultaneous inversion in determining the distribution of tectonic coal[J].Journal of China Coal Society,2011,36(5):778-783.
    [6]Partyka G,Gridley J,Lopez J.Interpretational applications of spectral decomposition in reservoir characterization[J].The Leading Edge,1999,18(3):353-360.
    [7]Marfurt K J,Kirlin R L.Narrow-band spectral analysis and thin-bed tuning[J].Geophysics,2001,66(4):1274-1283.
    [8]Castagna J P,Sun S,Siegfried R W.Instantaneous spectral analysis:detection of low frequency shadows associated with hydrocarbon[J].The Leading Edge,2003,22(2):127-129.
    [9]Odebeatu E,Zhang J,Chapman M.Application of spectral decomposition to detection of dispersion anomalies associated with gas saturation[J].The Leading Edge,2006,25(2):205-210.
    [10]Charles I P,Castagna J P.Layer-thickness determination and stratigraphic interpretation using spectral inversion:Theory and application[J].Geophysics,2008,73(2):R37-R48.
    [11]Loizou N,Liu E,Chapman M.AVO analyses and spectral decomposition of seismic data from four wells west of Shetland,UK[J].Petroleum Geoscience,2008,14(4):355-368.
    [12]Williams G,Chadwick A.Quantitative seismic analysis of a thin layer of CO2in the Sleipner injection plume[J].Geophysics,2012,77(6):245-256.
    [13]Hart B S.Channel detection in 3-D seismic data using sweetness[J].AAPG Bulletin,2008,92(6):733-742.
    [14]Ahmad M N,Rowell A P.Application of spectral decomposition and seismic attributes to understand the structure and distribution of sand reservoirs within Tertiary rift basins of the Gulf of Thailand[J].The Leading Edge,2012,31(6):630-634.
    [15]Jiao S,Zeng G M,He L.Prediction of dust fall concentrations in urban atmospheric environment through support vector regression[J].Journal of Central South University,2010,17(2):307-315.
    [16]Akpona O,Sebastian V L,Laurent T.Support vector regression and synthetically mixed training data for quantifying urban land cover[J].Remote Sensing of Environment,2013,137(10):184-197.
    [17]Zhang H K,Huang B.Support vector regression-based downscaling for intercalibration of multiresolution satellite images[J].Geoscience and Remote Sensing,2013,51(3):1114-1123.
    [18]Cortes C,Vapnik V.Support-vector network[J].Machine Learning,1995,20(3):273-297.
    [19]Chang C C,Lin C J.LIBSVM:A library for support vector machines[J].ACM Transactions on Intelligent Systems and Technology,2011,2(3):1-27.
    [20]王新.煤与瓦斯突出高危煤层的地震属性预测方法研究[D].徐州:中国矿业大学,2012.
    [21]Stockwell R G,Mansinha L,Lowe R P.Localization of the complex spectrum:The S transform[J].IEEE Transaction on Signal Processing,1996,44(4):998-1001.
    [22]Arthur E B.A tutorial on complex seismic trace analysis[J].Geophysics,2007,72(6):W33-W43.
    [23]Liu G,Wang L Y,Chen G M.Parameters optimization of plasma hardening process using genetic algorithm and neural network[J].Journal of Iron and Steel Research International,2011,18(12):57
    [24]Rodriguez J D,Perez A,Lozano J A.Sensitivity analysis of k-fold cross validation in prediction error estimation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2010,32(3):569-575.

版权所有:© 2021 中国地质图书馆 中国地质调查局地学文献中心