基于深度学习的地震岩相反演方法
|
摘要
复杂岩性预测是地震储层预测的难题,基于机器学习的非线性反演是识别岩性的有效手段。常规方法多以测井特征曲线(伽马曲线等)为学习目标,利用BP神经网络建立非线性映射预测岩性体,但这种方法存在两个问题,一是井震分辨率不匹配,二是BP神经网络在反演过程中存在局部收敛、效果不稳定以及非线性表征能力弱的问题。为解决这些问题,一是通过引入地震岩相概念解决井震分辨率不匹配问题,二是将深度学习引入到地震岩相反演中,经过优化样本采样、抽取相控伪井解决大样本集的构建问题,采用增量学习的策略进一步提高预测模型的精度和稳定性。以分频地震数据作为预测模型的输入,井岩相曲线为反演目标,实现了基于深度学习的地震岩相反演,有效解决了复杂岩性预测的难题。将该方法应用于海上某深水陆坡水道沉积研究区(该区发育灰岩、钙质砂岩、砂岩和泥岩4种岩相,岩石物理规律复杂,区分困难)岩性预测,结果表明,基于深度学习的地震岩相反演结果与井资料吻合,与地质认识相符。与叠前反演方法和BP神经网络学习岩相反演方法相比,基于深度学习的地震岩相反演方法准确度和分辨率更高,证明该方法是复杂岩性预测的有效手段。
Complex lithology prediction is a difficult problem in seismic reservoir prediction.Nonlinear inversion based on machine learning is an effective method for identifying lithology.Conventional method is to establish nonlinear mapping to predict lithologic bodies with logging characteristic curves(such as gamma)as the learning objectives.But there are two limitations,one is the mismatch of well seismic resolution,the other is that the BP neural network has local convergence,unstable effect,and weak nonlinear characterization in the inversion process.To overcome these limitations,the concept of seismic lithofacies is introduced to solve the mismatch of well seismic resolution,deep learning is introduced into seismic lithofacies inversion to construct large sample sets by optimizing sampling and extracting facies-controlled pseudo wells,and also an incremental learning strategy is adopted to further improve the accuracy and stability of prediction model.Using frequency division seismic data as the input to the prediction model and the well lithofacies curve as the inversion target,seismic lithofacies inversion based on deep learning was realized,and found to be effectively solving the problem with complex lithology prediction.The study area is a deep-water continental slope waterway in the sea,where four lithofacies develop,including limestone,calcareous sandstone,sandstone,and mudstone;the lithofacies are complicated and difficult to distinguish.Yet,the seismic lithofacies inversion results based on deep learning were consistent with both the well data and the geological knowledge.The proposed method is superior to prestack inversion and BP neural network learning method for its higher accuracy and resolution,proving its effectiveness for complex lithology prediction.
Complex lithology prediction is a difficult problem in seismic reservoir prediction.Nonlinear inversion based on machine learning is an effective method for identifying lithology.Conventional method is to establish nonlinear mapping to predict lithologic bodies with logging characteristic curves(such as gamma)as the learning objectives.But there are two limitations,one is the mismatch of well seismic resolution,the other is that the BP neural network has local convergence,unstable effect,and weak nonlinear characterization in the inversion process.To overcome these limitations,the concept of seismic lithofacies is introduced to solve the mismatch of well seismic resolution,deep learning is introduced into seismic lithofacies inversion to construct large sample sets by optimizing sampling and extracting facies-controlled pseudo wells,and also an incremental learning strategy is adopted to further improve the accuracy and stability of prediction model.Using frequency division seismic data as the input to the prediction model and the well lithofacies curve as the inversion target,seismic lithofacies inversion based on deep learning was realized,and found to be effectively solving the problem with complex lithology prediction.The study area is a deep-water continental slope waterway in the sea,where four lithofacies develop,including limestone,calcareous sandstone,sandstone,and mudstone;the lithofacies are complicated and difficult to distinguish.Yet,the seismic lithofacies inversion results based on deep learning were consistent with both the well data and the geological knowledge.The proposed method is superior to prestack inversion and BP neural network learning method for its higher accuracy and resolution,proving its effectiveness for complex lithology prediction.
引文
[1]李国发,岳英,熊金良.基于三维模型的薄互层振幅属性实验研究[J].石油地球物探勘探,2011,46(1):115-120LI G F,YUE Y,XIONG J L.Experimental study on amplitude properties of thin interbed based on 3Dmodel[J].Oil Geophysical Prospecting,2011,46(1):115-120
[2]林忠民.塔河油田奥陶系碳酸盐岩储层特征及成藏条件[J].石油学报,2002,23(3):23-26LIN Z M.Carbonate rock reservoir features and oil-gas accumulating condition in the Ordovician of Tahe oilfield in Northern Tamin Basin[J].Acta Petrolei Sinica,2002,23(3):23-26
[3]庚琪.辽河盆地东部凹陷火成岩储层特征及成藏模式[J].复杂油气藏,2016,9(1):6-11GENG Q.Igneous reservoir characteristics and hydrocarbon accumulation model in eastern sag of Liaohe Basin[J].Complex Hydrocarbon Reservoirs,2016,9(1):6-11
[4]彭俊,彭嫦姿,于少勇.复杂岩性有效储层地震预测技术——以B气田大安寨段一亚段气藏为例[J].断块油气田,2015,22(3):342-346PENG J,PENG C Z,YU S Y.Seismic prediction technology for effective reservoir in complex lithology:The1st interval of Da’anzhai Section in B Gas Field case[J].Fault-Block Oil and Gas Field,2015,22(3):342-346
[5]胡英,陈辉,贺振华,等.基于地震纹理属性和模糊聚类划分地震相[J].石油地球物理勘探,2013,48(1):114-120HU Y,CHEN H,HE Z H,et al.Seismic facies classification based on seismic texture attributes and fuzzy clustering[J].Oil Geophysical Prospecting,2013,48(1):114-120
[6]张繁昌,刘汉卿,钮学民,等.褶积神经网络高分辨率地震反演[J].石油地球物理勘探,2014,49(6):1165-1169ZHANG F C,LIU H Q,NIU X M,et al.High resolution seismic inversion by convolutional neural network[J].Oil Geophysical Prospecting,2014,49(6):1165-1169
[7]吴媚,符力耘,李维新.高分辨率非线性储层物性参数反演方法和应用[J].地球物理学报,2008,51(2):546-557WU M,FU L Y,LI W X.A high-resolution nonlinear inversion method of reservoir parameters and its application to oil/gas exploration[J].Chinese Journal of Geophysics,2008,51(2):546-557
[8] HINTON G E,OSINDERO S,TEH Y.A fast learning algorithm for deep belief nets[J].Neural Computation,2006,18(7):1527-1554
[9] BENGIO Y.Learning deep architectures for AI[M].Boston:Now Publishers Inc,2009:1-127
[10] RANZATO M,BOUREAU Y,CHOPRA S,et al.A unified energy-based framework for unsupervised learning[J].Proceedings of the Eleventh International Conference on Artificial Intelligence and Statistics,2007:371-379
[11]程国建,刘丽婷.深度学习算法应用于岩石图像处理的可行性研究[J].软件导刊,2016,15(9):163-166CHEN G J,LIU L T.Feasibility study of deep learning algorithm applied to rock image processing[J].Software Guide,2016,15(9):163-166
[12] CAO J X,WU S K.Deep learning:Chance and challenge for deep gas reservoir identification[C]∥Qingdao:International Geophysical Conference,2017:711-712
[13] LEWIS W,VIGH D.Deep learning prior models from seismic images for full-waveform inversion[J].Expanded Abstracts of 87th Annual Internat SEG Mtg,2017:1512-1517
[14] BESTAGINI P,LIPARI V,TUBARO S.A machine learning approach to facies classification using well logs[J].Expanded Abstracts of 87th Annual Internat SEG Mtg,2017:2137-2142
[15]刘力辉,李建海,刘玉霞.地震物相分析方法与“甜点”预测[J].石油物探,2013,52(4):432-437LIU L H,LI J H,LIU Y X.Seismic reservoir property facies analysis and sweet spot prediction[J].Geophysical Prospecting for Petroleum,2013,52(4):432-437
[16]刘帅师,程曦,郭文燕,等.深度学习方法研究新进展[J].智能系统学报,2016,11(5):567-577LIU S S,CHENG X,GUO W Y,et al.Progress report on new research in deep learning[J].CAAI Transactions on Intelligent Systems,2016,11(5):567-577
[2]林忠民.塔河油田奥陶系碳酸盐岩储层特征及成藏条件[J].石油学报,2002,23(3):23-26LIN Z M.Carbonate rock reservoir features and oil-gas accumulating condition in the Ordovician of Tahe oilfield in Northern Tamin Basin[J].Acta Petrolei Sinica,2002,23(3):23-26
[3]庚琪.辽河盆地东部凹陷火成岩储层特征及成藏模式[J].复杂油气藏,2016,9(1):6-11GENG Q.Igneous reservoir characteristics and hydrocarbon accumulation model in eastern sag of Liaohe Basin[J].Complex Hydrocarbon Reservoirs,2016,9(1):6-11
[4]彭俊,彭嫦姿,于少勇.复杂岩性有效储层地震预测技术——以B气田大安寨段一亚段气藏为例[J].断块油气田,2015,22(3):342-346PENG J,PENG C Z,YU S Y.Seismic prediction technology for effective reservoir in complex lithology:The1st interval of Da’anzhai Section in B Gas Field case[J].Fault-Block Oil and Gas Field,2015,22(3):342-346
[5]胡英,陈辉,贺振华,等.基于地震纹理属性和模糊聚类划分地震相[J].石油地球物理勘探,2013,48(1):114-120HU Y,CHEN H,HE Z H,et al.Seismic facies classification based on seismic texture attributes and fuzzy clustering[J].Oil Geophysical Prospecting,2013,48(1):114-120
[6]张繁昌,刘汉卿,钮学民,等.褶积神经网络高分辨率地震反演[J].石油地球物理勘探,2014,49(6):1165-1169ZHANG F C,LIU H Q,NIU X M,et al.High resolution seismic inversion by convolutional neural network[J].Oil Geophysical Prospecting,2014,49(6):1165-1169
[7]吴媚,符力耘,李维新.高分辨率非线性储层物性参数反演方法和应用[J].地球物理学报,2008,51(2):546-557WU M,FU L Y,LI W X.A high-resolution nonlinear inversion method of reservoir parameters and its application to oil/gas exploration[J].Chinese Journal of Geophysics,2008,51(2):546-557
[8] HINTON G E,OSINDERO S,TEH Y.A fast learning algorithm for deep belief nets[J].Neural Computation,2006,18(7):1527-1554
[9] BENGIO Y.Learning deep architectures for AI[M].Boston:Now Publishers Inc,2009:1-127
[10] RANZATO M,BOUREAU Y,CHOPRA S,et al.A unified energy-based framework for unsupervised learning[J].Proceedings of the Eleventh International Conference on Artificial Intelligence and Statistics,2007:371-379
[11]程国建,刘丽婷.深度学习算法应用于岩石图像处理的可行性研究[J].软件导刊,2016,15(9):163-166CHEN G J,LIU L T.Feasibility study of deep learning algorithm applied to rock image processing[J].Software Guide,2016,15(9):163-166
[12] CAO J X,WU S K.Deep learning:Chance and challenge for deep gas reservoir identification[C]∥Qingdao:International Geophysical Conference,2017:711-712
[13] LEWIS W,VIGH D.Deep learning prior models from seismic images for full-waveform inversion[J].Expanded Abstracts of 87th Annual Internat SEG Mtg,2017:1512-1517
[14] BESTAGINI P,LIPARI V,TUBARO S.A machine learning approach to facies classification using well logs[J].Expanded Abstracts of 87th Annual Internat SEG Mtg,2017:2137-2142
[15]刘力辉,李建海,刘玉霞.地震物相分析方法与“甜点”预测[J].石油物探,2013,52(4):432-437LIU L H,LI J H,LIU Y X.Seismic reservoir property facies analysis and sweet spot prediction[J].Geophysical Prospecting for Petroleum,2013,52(4):432-437
[16]刘帅师,程曦,郭文燕,等.深度学习方法研究新进展[J].智能系统学报,2016,11(5):567-577LIU S S,CHENG X,GUO W Y,et al.Progress report on new research in deep learning[J].CAAI Transactions on Intelligent Systems,2016,11(5):567-577