浊积岩储层物性预测技术研究
|
摘要
随着油气田勘探开发技术的不断发展,各大油田的开发程度不断提高,岩性油气藏越来越引起勘探开发人员的高度关注。浊积岩油藏作为一种复杂的岩性油气藏,在油气勘探开发中具有良好的前景。本文以史南地区沙三段浊积岩沉积层段为研究对象,开展浊积岩储层物性参数—厚度、孔隙度的地质、地球物理研究。在认真细致分析研究区内的地震、地质、测井、钻井等资料的基础上,对浊积岩目的层段的砂岩累计厚度、孔隙度空间物性分布、砂泥岩薄互层的测井曲线、地震反射特征,进行了详细分析研究,为沙三段浊积岩储层物性预测研究奠定了基础。
在研究掌握了史南地区浊积岩沉积特征、构造特征及成藏规律的基础上,针对目的层段砂泥岩薄互层情况,研究设计了砂岩薄层厚度预测与孔隙度预测技术方法。本文在研究了砂泥岩薄互层地震响应在时域与频域的特征基础上,针对薄互层预测的这一难点问题,研究了建立时、频域目标函数,在厚层稀疏大反射系数的约束下,进行薄互层反射系数联合反演技术方法。针对非线性算法的的随机性缺点,讨论了L1- L2范数联合约束稀疏脉冲反褶积计算方法,在薄层反射系数反演方法研究过程中,设计了分段模拟退火反演的方法,极大提高了算法的结果的可靠性和计算效率。经理论模型与实际资料的验证,本方法取得了良好的效果。在反演结果的基础上,对研究区内的砂岩累计厚度进行了计算,经与井资料对比分析,预测结果具有较高的精度。
针对储层孔隙度物性参数的特点,在大量统计研究区内井资料与地震属性数据的基础上,分析了本地区物性预测孔隙度硬数据与地震属性软数据的相关关系,对地震属性软数据进行了优化分析,选取了相关度较高的软数据参与计算。在此基础上,针对贝叶斯网络推理预测中的不足,研究了贝叶斯-马尔科夫毯网络孔隙度预测方法,并设计了相应算法,最后利用该算法对史南地区目的层段孔隙度进行了预测。对预测结果通过多口井的抽稀验证分析,该方法的物性预测结果可靠,并且具有较高的精度,是一种孔隙度预测切实有效的方法。
As the developing of oilfield exploration technology ,the development of major oil fields keeps Increasing. Exploring scientists increasingly show great attention to the lithologic hydrocarbon reservoir. As a complex lithologic,the turbidite reservoir have well prospect in the oil and gas exploration. In this paper,we use turbidite deposition of shinan area as the research object. Carrying out the research of physical properties parameters of the Turbidite reservoir, such as geologic and geophysics characteristics of thickness and porosity. Based on the careful analysis of the seismic, geological, logging and drilling data, the sandstone thickness and porosity of target zone of turbidite and the spatial distribution of physical properties were predicted. It’s provided a good basis of seismic prediction for the reservoir physical properties prediction of the turbidite.
Based on the researching and mastering of the sedimentary characteristics, structural characteristics and forming law of the turbidite in the Shinan area, in terms of the thin interbedded situation of the sandstone and mudstone in the target layers, thickness and porosity forecast technology method of sandstone the thin interbedded are designed. On the basis of characteristics of seismic response of sand and mud thin interbeds in time and frequency domains, aiming at the difficulties of thin interbeds forecasting ,this article established time-frequency domain target function, carry on research in joint inversion method of thin interbedded reflection coefficient, in constraint of larger constrained sparse reflection coefficient. Aiming at the weakness of randomness of non-linear algorithm, discussed the inverse method of the constraint of the result of L1-L2 norm joint constrained sparse spike deconvolution. In the research of the inversion method of thin interbedded reflection coefficient, puting forward the Simulated Annealing Inversion, greatly improve the computational reliability and efficiency. Based on the results of conversion, we calculated the thickness of sandstone in this area. Compared with the well data, show that the prediction have high precision, and get the good results.
According to the characteristics of the physical parameters of the reservoir, we analysis the relation between hard data on physical properties prediction and earthquake soft data of the area, optimized the soft data on the seismic, and select the relevant data in a higher degree. All these operations are based on a large number of statistics of logging data and seismic data in the study area. For this reason ,we study Bayesian-Markov blanket network prediction method for porosity prediction, designed the relevant algorithm and use the algorthm to predict porosity in target stratum of shinan area, aiming at the weakness of bayesian network prediction. By means of vacuating verification for multi-well,the prediction results show reliablely and have higher precision, and it is an effective reservoir parameter prediction method.
在研究掌握了史南地区浊积岩沉积特征、构造特征及成藏规律的基础上,针对目的层段砂泥岩薄互层情况,研究设计了砂岩薄层厚度预测与孔隙度预测技术方法。本文在研究了砂泥岩薄互层地震响应在时域与频域的特征基础上,针对薄互层预测的这一难点问题,研究了建立时、频域目标函数,在厚层稀疏大反射系数的约束下,进行薄互层反射系数联合反演技术方法。针对非线性算法的的随机性缺点,讨论了L1- L2范数联合约束稀疏脉冲反褶积计算方法,在薄层反射系数反演方法研究过程中,设计了分段模拟退火反演的方法,极大提高了算法的结果的可靠性和计算效率。经理论模型与实际资料的验证,本方法取得了良好的效果。在反演结果的基础上,对研究区内的砂岩累计厚度进行了计算,经与井资料对比分析,预测结果具有较高的精度。
针对储层孔隙度物性参数的特点,在大量统计研究区内井资料与地震属性数据的基础上,分析了本地区物性预测孔隙度硬数据与地震属性软数据的相关关系,对地震属性软数据进行了优化分析,选取了相关度较高的软数据参与计算。在此基础上,针对贝叶斯网络推理预测中的不足,研究了贝叶斯-马尔科夫毯网络孔隙度预测方法,并设计了相应算法,最后利用该算法对史南地区目的层段孔隙度进行了预测。对预测结果通过多口井的抽稀验证分析,该方法的物性预测结果可靠,并且具有较高的精度,是一种孔隙度预测切实有效的方法。
As the developing of oilfield exploration technology ,the development of major oil fields keeps Increasing. Exploring scientists increasingly show great attention to the lithologic hydrocarbon reservoir. As a complex lithologic,the turbidite reservoir have well prospect in the oil and gas exploration. In this paper,we use turbidite deposition of shinan area as the research object. Carrying out the research of physical properties parameters of the Turbidite reservoir, such as geologic and geophysics characteristics of thickness and porosity. Based on the careful analysis of the seismic, geological, logging and drilling data, the sandstone thickness and porosity of target zone of turbidite and the spatial distribution of physical properties were predicted. It’s provided a good basis of seismic prediction for the reservoir physical properties prediction of the turbidite.
Based on the researching and mastering of the sedimentary characteristics, structural characteristics and forming law of the turbidite in the Shinan area, in terms of the thin interbedded situation of the sandstone and mudstone in the target layers, thickness and porosity forecast technology method of sandstone the thin interbedded are designed. On the basis of characteristics of seismic response of sand and mud thin interbeds in time and frequency domains, aiming at the difficulties of thin interbeds forecasting ,this article established time-frequency domain target function, carry on research in joint inversion method of thin interbedded reflection coefficient, in constraint of larger constrained sparse reflection coefficient. Aiming at the weakness of randomness of non-linear algorithm, discussed the inverse method of the constraint of the result of L1-L2 norm joint constrained sparse spike deconvolution. In the research of the inversion method of thin interbedded reflection coefficient, puting forward the Simulated Annealing Inversion, greatly improve the computational reliability and efficiency. Based on the results of conversion, we calculated the thickness of sandstone in this area. Compared with the well data, show that the prediction have high precision, and get the good results.
According to the characteristics of the physical parameters of the reservoir, we analysis the relation between hard data on physical properties prediction and earthquake soft data of the area, optimized the soft data on the seismic, and select the relevant data in a higher degree. All these operations are based on a large number of statistics of logging data and seismic data in the study area. For this reason ,we study Bayesian-Markov blanket network prediction method for porosity prediction, designed the relevant algorithm and use the algorthm to predict porosity in target stratum of shinan area, aiming at the weakness of bayesian network prediction. By means of vacuating verification for multi-well,the prediction results show reliablely and have higher precision, and it is an effective reservoir parameter prediction method.
引文
[1]李祥辉,王成善,金玮等.深海沉积理论发展及其在油气勘探中的意义[J].沉积学报, 2009,27(1):77-86
[2]杨凤丽,王清,沈财余.渤南一区浊积岩沉积学研究[J].沉积与特提斯地质,2006,26(1):37-40
[3]罗红梅.地震属性技术在现河地区滑塌浊积岩油藏勘探的应用[J].工程地球物理学报,2005,2(3):224-227
[4]杜文凤,彭苏萍.利用地质统计学预测煤层厚度[J].岩石力学与工程学报, 2010, 29(z1):22-27
[5]唐小彪.基于对应分析的支持向量机回归在地震储层厚度预测中的应用[J].物探与化探, 2009, 33(4): 67-71
[6]邹华胜,宁书年.支持向量机在储层厚度预测和计算中的应用[J].地球物理学进展, 2007 22(5):11-19
[7]周开明.薄层的二阶功率谱特征研究及厚度预测[J].石油物探, 2008, 47(1) :45-51
[8]李瑞. BP神经网络在储层物性参数预测中的应用[J].西北大学学报,2002, 32(3) : 271-272
[9]印兴耀,吴国忱,张洪宙.神经网络在储层横向预测中的应用[J].石油大学学报(自然科学版),1994,18(5):20-25
[10]金业权,王越之,李自俊.地震资料预测地层孔隙压力应基于欠压实成因[J].石油钻探技术,2000,28(3):7-11
[11]钟淑敏,刘传平,杨青山.一种有效确定特低渗透砂岩储层有效厚度的方法[J].大庆石油质与开发,1999,18(5):46-47
[12]陆文凯,李衍达,牟永光.误差反传播神经网络法地震反演.地球物理学报[J],1996, 39(增):292-30
[13]崔炯成,吴媚,李维新等.高分辨率非线性地震波阻抗反演方法和应用[J].石油物探, 2007,46(1):1-12
[14]张建宁.牛庄洼陷浊积岩砂体储层物性地震预测[J].石油地球物理勘探,2005,40(6):716-720
[15]杨剑萍,陈顺霞,查明等.惠民凹陷中央隆起带古近系地震浊积岩特征研究[J].煤田地质与勘探,2006, 34(5): 1-4
[16]李凤杰,王多云,苑克增等.人工神经网络技术在油田储层物性预测中的应用-以西峰油田为例[J] .天然气地球科学,2004,15(3):244—246
[17]杨文采.非线性反演方法的补充及比较[J].石油物探,1995,34(4):39
[18]孙洪泉.地质统计学及其应用[M].徐州:中国矿业大学出版社, 1990, 25-75
[19]陈科.基于模拟退火的谱反演方法研究[D].东营:中国石油大学, 2010.
[20]秦德文,基于谱反演的薄层预测与反演方法研究[D].东营:中国石油大学, 2009
[21]王宇,韩立国,周家雄等. L1_L2范数联合约束稀疏脉冲反演的应用[J].地球科学-中国地质大学学报, 2009,34 (5):835-840
[22] Charles I. Puryear and John P. Castagna,Layer-thickness determination and stratigraphic interpretation using spectral inversion: Theory and application [J],GEOPHYSICS,2011,73(2),37-48
[23]李振春,张军华.地震数据处理方法[M].东营:中国石油大学出版社,2006, 99-145
[24]李庆忠.走向精确勘探的道路[M].北京:石油工业出版社,1993, 25-90
[25]姜建国等.信号与系统分析[M].清华大学出版社,2006:50-169, 17-31
[26]张智江李岳生等.介质反射系数的反演方法及其计算机实现[J].高等学校计算数学学报,1996,44(4):69-72
[27] Kelyn Paola Casta?o, Optimizing thin-layer mapping through Spectral Inversion: Performance of Genetic Algorithms and Simulated Annealing[C], SEG Denver 2010 Annual Meeting, SEG, 2010:1610-1614
[28]王文涛.地震储层评价与预测的贝叶斯反演方法研究[D].武汉:中国地质大学,2008
[29]王文涛,朱培民等.地震储层预测中贝叶斯反演方法的研究[J].石油天然气学报(江汉石油学院学报),2009,31(5):263-266
[30] Horn, R. A. and Johnson, C. R. "Norms for Vectors and Matrices." Ch. 5 in Matrix Analysis[M]. Cambridge, England: Cambridge University Press, 1990, 2-25
[31]雷桂媛,关于蒙特卡罗及拟蒙特卡罗方法的若干研究[D].杭州:浙江大学, 2003
[32]詹海刚,施平,陈楚群.基于贝叶斯反演理论的海水固有光学特性准分析算法[J].科学通报,2006, 51(2): 204-207
[33]韩世勤,张玉芬.小波变换及其在信号分析中的应用[J].中南民族学院学报, 1996, 15(4), 86-91
[34]皇甫堪,陈建文,楼生强.现代数字信号处理[M].北京:电子工业出版社, 2003, 26-35,340-345
[35]赵启蒙.储层随机建模方法研究[D].大庆:大庆石油学院,2004
[36]邓继新,基于正演模型的储层谱分解响应特征分析[J].石油学报,2008,29(4):539-544
[37]张璐,印兴耀,李红梅.基于地震奇异性属性划分砂砾岩扇体沉积界面[J].物探化探计算技术,2009,31(4):361-373
[38] Stockwell.R.G, L.Mansinha,Lowe.R.P. Localization of the complex spectrum: The S transform[J]. IEEE Transacations on Signal Processing .1996, 44: 998-1001
[39] Bracewell R N. The Fourier transform and its application[M].New York: McGraw-Hill, 2000, 256-260
[40] Chakraborty A,Okaya D.Frequenct-time decomposition of sersmic data using wavelet-based methods[J]. Geophysics, 1995, 60(6): 1906-1916
[41] Bracewell R N. The Fourier transform and its application[M]. New York: McGraw-Hill, 2000, 256-260
[42]贺锡雷,黄德济,贺振华.薄互层反射系数序列时~频特征研究[J].物探化探计算技术,2009,31(3):227-232
[43]温祥泉,薄互层砂岩储层的人机交互预测[J].石油实验地质,1995,17(2):193-199
[44]汪恩华,贺振华,李庆忠,薄储层厚度计算新方法探索[J].物探化探计算技术, 2001,23(1):21-25
[45]姜镭.基于时频谱特征的薄互层分析[D].成都:成都理工大学,2009
[46] Stephen J. Hill, Golden ,Inversion-based thickness determination[J].The Leading Edge,2005,35(6):477-487
[47]付震,范振忠.大庆油田葡萄花油层有效厚度的确定[J].科学技术与工程, 2010, 10(22): 5503-5510
[48]李玮玮.贝叶斯网络结构学习方法的研究[D].南京:南京航空航天大学, 2008
[49]张剑飞.贝叶斯网络学习方法和算法研究[D].长春:东北师范大学, 2005
[50]王双成,程新章,王振海.贝叶斯网络中变量的最优预测[J].计算机应用与软件, 2007, 24(5): 9-12
[51]王双成苑森森,基于贝叶斯网络的马尔科夫毯预测学习[J].模式识别与人工智能, 2004, 17(1): 17-23
[52]张隆.基于信息论的特征选择和分类算法研究[D].重庆:西南农业大学, 2005
[53]王辉,张剑飞,王双成等.基于预测能力的贝叶斯网络结构学习[J].东北师大学报自然科学版, 2005, 37(1): 53-68
[54]高美娟,朱庚忠,张赦华.利用贝叶斯～克里金估计技术进行储层参数预测[J].石油地球物理勘探, 1999, 34(4): 390-397
[55] Constantin F. Aliferis, Alexander Statnikov, Ioannis Tsamardinos. Local Causal and Markov Blanket Induction for Causal Discovery and Feature Selection for Classification [J]. Journal of Machine Learning Research 11. 2010, 15(6): 171-234
[56] Jean-Philippe Pellet, Andr′e Elisseeff. Using Markov Blankets for Causal Structure Learning[J]. Journal of Machine Learning Research , 2008, 19(6): 1295-1342
[57] Jacobs E W,Fisher Y,Boss R D. Image compression: a study of iterated transform method[J]. Singal P Processing,1992,29:251-263
[58] Jacquin A E.Fractal image coding based on a theory of iterated contractive image transformation[J]. SPIE Visual Comm Image Process,1990,1360:227-239
[59]李来林等.用马尔科夫-贝叶斯方法预测孔隙度的分布[J].石油物探, 1995, 34(4): 33-41
[60]蒋国萍.软件项目风险管理的贝叶斯网络模型研究[D].长沙:国防科学技术大学,2005
[61]石玉梅,姚逢昌,甘利灯.多渠道数据融合及其应用[J].石油物探, 2003, 42(1): 22-24
[62] Wei Xiaodong, Zhang Yanqing, Cao Lili, Wang Yanan and Zhang Yazhing, The gradient of the amplitude spectrum of sersmic data and its application in reservoir prediction[C]. SEG Houston 2009 International Exposition and Annual Meeting
[63] Theophanis. S, Queen.J. Color display of the localized spectrum[J]. Geophysics, 2000, 65(4): 1330-1340
[64] Tirado. S. Sand thickness estimation using spectral decomposition[M]. M.S.thesis. University of Oklahoma, 2004
[65]黄飞. PNN及其在储层物性参数预测中的应用[J].物化探计算技术. 2002, 25(1):78-82
[2]杨凤丽,王清,沈财余.渤南一区浊积岩沉积学研究[J].沉积与特提斯地质,2006,26(1):37-40
[3]罗红梅.地震属性技术在现河地区滑塌浊积岩油藏勘探的应用[J].工程地球物理学报,2005,2(3):224-227
[4]杜文凤,彭苏萍.利用地质统计学预测煤层厚度[J].岩石力学与工程学报, 2010, 29(z1):22-27
[5]唐小彪.基于对应分析的支持向量机回归在地震储层厚度预测中的应用[J].物探与化探, 2009, 33(4): 67-71
[6]邹华胜,宁书年.支持向量机在储层厚度预测和计算中的应用[J].地球物理学进展, 2007 22(5):11-19
[7]周开明.薄层的二阶功率谱特征研究及厚度预测[J].石油物探, 2008, 47(1) :45-51
[8]李瑞. BP神经网络在储层物性参数预测中的应用[J].西北大学学报,2002, 32(3) : 271-272
[9]印兴耀,吴国忱,张洪宙.神经网络在储层横向预测中的应用[J].石油大学学报(自然科学版),1994,18(5):20-25
[10]金业权,王越之,李自俊.地震资料预测地层孔隙压力应基于欠压实成因[J].石油钻探技术,2000,28(3):7-11
[11]钟淑敏,刘传平,杨青山.一种有效确定特低渗透砂岩储层有效厚度的方法[J].大庆石油质与开发,1999,18(5):46-47
[12]陆文凯,李衍达,牟永光.误差反传播神经网络法地震反演.地球物理学报[J],1996, 39(增):292-30
[13]崔炯成,吴媚,李维新等.高分辨率非线性地震波阻抗反演方法和应用[J].石油物探, 2007,46(1):1-12
[14]张建宁.牛庄洼陷浊积岩砂体储层物性地震预测[J].石油地球物理勘探,2005,40(6):716-720
[15]杨剑萍,陈顺霞,查明等.惠民凹陷中央隆起带古近系地震浊积岩特征研究[J].煤田地质与勘探,2006, 34(5): 1-4
[16]李凤杰,王多云,苑克增等.人工神经网络技术在油田储层物性预测中的应用-以西峰油田为例[J] .天然气地球科学,2004,15(3):244—246
[17]杨文采.非线性反演方法的补充及比较[J].石油物探,1995,34(4):39
[18]孙洪泉.地质统计学及其应用[M].徐州:中国矿业大学出版社, 1990, 25-75
[19]陈科.基于模拟退火的谱反演方法研究[D].东营:中国石油大学, 2010.
[20]秦德文,基于谱反演的薄层预测与反演方法研究[D].东营:中国石油大学, 2009
[21]王宇,韩立国,周家雄等. L1_L2范数联合约束稀疏脉冲反演的应用[J].地球科学-中国地质大学学报, 2009,34 (5):835-840
[22] Charles I. Puryear and John P. Castagna,Layer-thickness determination and stratigraphic interpretation using spectral inversion: Theory and application [J],GEOPHYSICS,2011,73(2),37-48
[23]李振春,张军华.地震数据处理方法[M].东营:中国石油大学出版社,2006, 99-145
[24]李庆忠.走向精确勘探的道路[M].北京:石油工业出版社,1993, 25-90
[25]姜建国等.信号与系统分析[M].清华大学出版社,2006:50-169, 17-31
[26]张智江李岳生等.介质反射系数的反演方法及其计算机实现[J].高等学校计算数学学报,1996,44(4):69-72
[27] Kelyn Paola Casta?o, Optimizing thin-layer mapping through Spectral Inversion: Performance of Genetic Algorithms and Simulated Annealing[C], SEG Denver 2010 Annual Meeting, SEG, 2010:1610-1614
[28]王文涛.地震储层评价与预测的贝叶斯反演方法研究[D].武汉:中国地质大学,2008
[29]王文涛,朱培民等.地震储层预测中贝叶斯反演方法的研究[J].石油天然气学报(江汉石油学院学报),2009,31(5):263-266
[30] Horn, R. A. and Johnson, C. R. "Norms for Vectors and Matrices." Ch. 5 in Matrix Analysis[M]. Cambridge, England: Cambridge University Press, 1990, 2-25
[31]雷桂媛,关于蒙特卡罗及拟蒙特卡罗方法的若干研究[D].杭州:浙江大学, 2003
[32]詹海刚,施平,陈楚群.基于贝叶斯反演理论的海水固有光学特性准分析算法[J].科学通报,2006, 51(2): 204-207
[33]韩世勤,张玉芬.小波变换及其在信号分析中的应用[J].中南民族学院学报, 1996, 15(4), 86-91
[34]皇甫堪,陈建文,楼生强.现代数字信号处理[M].北京:电子工业出版社, 2003, 26-35,340-345
[35]赵启蒙.储层随机建模方法研究[D].大庆:大庆石油学院,2004
[36]邓继新,基于正演模型的储层谱分解响应特征分析[J].石油学报,2008,29(4):539-544
[37]张璐,印兴耀,李红梅.基于地震奇异性属性划分砂砾岩扇体沉积界面[J].物探化探计算技术,2009,31(4):361-373
[38] Stockwell.R.G, L.Mansinha,Lowe.R.P. Localization of the complex spectrum: The S transform[J]. IEEE Transacations on Signal Processing .1996, 44: 998-1001
[39] Bracewell R N. The Fourier transform and its application[M].New York: McGraw-Hill, 2000, 256-260
[40] Chakraborty A,Okaya D.Frequenct-time decomposition of sersmic data using wavelet-based methods[J]. Geophysics, 1995, 60(6): 1906-1916
[41] Bracewell R N. The Fourier transform and its application[M]. New York: McGraw-Hill, 2000, 256-260
[42]贺锡雷,黄德济,贺振华.薄互层反射系数序列时~频特征研究[J].物探化探计算技术,2009,31(3):227-232
[43]温祥泉,薄互层砂岩储层的人机交互预测[J].石油实验地质,1995,17(2):193-199
[44]汪恩华,贺振华,李庆忠,薄储层厚度计算新方法探索[J].物探化探计算技术, 2001,23(1):21-25
[45]姜镭.基于时频谱特征的薄互层分析[D].成都:成都理工大学,2009
[46] Stephen J. Hill, Golden ,Inversion-based thickness determination[J].The Leading Edge,2005,35(6):477-487
[47]付震,范振忠.大庆油田葡萄花油层有效厚度的确定[J].科学技术与工程, 2010, 10(22): 5503-5510
[48]李玮玮.贝叶斯网络结构学习方法的研究[D].南京:南京航空航天大学, 2008
[49]张剑飞.贝叶斯网络学习方法和算法研究[D].长春:东北师范大学, 2005
[50]王双成,程新章,王振海.贝叶斯网络中变量的最优预测[J].计算机应用与软件, 2007, 24(5): 9-12
[51]王双成苑森森,基于贝叶斯网络的马尔科夫毯预测学习[J].模式识别与人工智能, 2004, 17(1): 17-23
[52]张隆.基于信息论的特征选择和分类算法研究[D].重庆:西南农业大学, 2005
[53]王辉,张剑飞,王双成等.基于预测能力的贝叶斯网络结构学习[J].东北师大学报自然科学版, 2005, 37(1): 53-68
[54]高美娟,朱庚忠,张赦华.利用贝叶斯～克里金估计技术进行储层参数预测[J].石油地球物理勘探, 1999, 34(4): 390-397
[55] Constantin F. Aliferis, Alexander Statnikov, Ioannis Tsamardinos. Local Causal and Markov Blanket Induction for Causal Discovery and Feature Selection for Classification [J]. Journal of Machine Learning Research 11. 2010, 15(6): 171-234
[56] Jean-Philippe Pellet, Andr′e Elisseeff. Using Markov Blankets for Causal Structure Learning[J]. Journal of Machine Learning Research , 2008, 19(6): 1295-1342
[57] Jacobs E W,Fisher Y,Boss R D. Image compression: a study of iterated transform method[J]. Singal P Processing,1992,29:251-263
[58] Jacquin A E.Fractal image coding based on a theory of iterated contractive image transformation[J]. SPIE Visual Comm Image Process,1990,1360:227-239
[59]李来林等.用马尔科夫-贝叶斯方法预测孔隙度的分布[J].石油物探, 1995, 34(4): 33-41
[60]蒋国萍.软件项目风险管理的贝叶斯网络模型研究[D].长沙:国防科学技术大学,2005
[61]石玉梅,姚逢昌,甘利灯.多渠道数据融合及其应用[J].石油物探, 2003, 42(1): 22-24
[62] Wei Xiaodong, Zhang Yanqing, Cao Lili, Wang Yanan and Zhang Yazhing, The gradient of the amplitude spectrum of sersmic data and its application in reservoir prediction[C]. SEG Houston 2009 International Exposition and Annual Meeting
[63] Theophanis. S, Queen.J. Color display of the localized spectrum[J]. Geophysics, 2000, 65(4): 1330-1340
[64] Tirado. S. Sand thickness estimation using spectral decomposition[M]. M.S.thesis. University of Oklahoma, 2004
[65]黄飞. PNN及其在储层物性参数预测中的应用[J].物化探计算技术. 2002, 25(1):78-82