基于静态电成像的砂砾岩储层岩性识别研究
摘要
深层砂砾岩是重要的油气勘探目标,但这类储层岩性复杂,岩相变化快,埋藏深,岩石骨架对电阻率影响大,因此常规测井识别岩性非常困难。
地层微电阻率扫描测井(FMI)的分辨率非常高,可以区分0.5cm的砾石,被称为“地层显微镜”。成像数据生成图像后,可以与取心照片进行对比,而且更加直观和具有连续性。因此广泛用于识别岩性、裂缝、孔隙等,对于研究沉积环境、地层构造具有很大的优势。
人工神经网络(Artificial Neural Networks)是近年来人们广泛关注的一门学科,结合生物学、数学模型和计算机算法,模拟大脑处理问题的思维,解决实际应用的数据处理问题,取得了巨大的成功。人工神经网络具有很强的容错性、自适应性及联想记忆等特点,用于岩性的识别和预测是可行的。
本文的主要内容是利用成像测井识别砂砾岩岩性,首先研究成像测井的静态图像,掌握研究区内的各种砂砾岩的岩性特征,总结成像模式;然后利用数字图像处理技术,对图像进行预处理和特征提取;最后研究人工神经网络常用的模型BP网、Hopfield网及SOFM网的网络结构和算法流程,论述了人工神经网络识别岩性的过程。通过MATLAB编程,主要采用BP算法进行砂砾岩的岩性识别,取得了很好的效果。
Deep glutenite reservoirs are an important target for petroleum exploration in China. Due to glutenite reservoir which has complex lithology, lithology facies changed rapidly, and the resistivity of rock matrix has a great impact on the electrical logs , it is very difficult to lithology identification in Deep glutenite reservoirs with conventional well logs.
Fullbore Formation MicroImage is promoted by Schlumberger, it has very high resolution is up to 5mm, named as“formation microscope”. The static images generated by FMI data can compared with cores, and are more intuitive and continuous.,and have been widely used to identify lithology, fractures, and pores,and have great advantages in the study of sedimentary environment and formation structure.
ANN (Artificial Neural Networks) is drawing widespread attention in recent years, combined biology, mathematical models and computer algorithms to simulate the brain so as to deal with the problem. It is used to solve data processing , and has achieved great success. Artificial neural network has a strong fault-tolerant, adaptive and associative memory, so it is feasible for lithology identification and prediction.
This paper focuses on identifying glutinite lithology. First, we study the static Image logging data, calibrate borehole image logs to cores, and summarize the features of glutenite rock; Second, we utilize image analysis algorithms to perform quantitative analysis and feature selection to borehole images, and transfer image features to characteristic value; Last, we investigate the network structure and study three models of ANN—BP Network, Hopfield Network and SOM network, especially study BP network and mainly utilize the BP neural network to identify lithology. This paper realizes the BP network to identify well lithology, and achieves satisfactory results using MATLAB code.
地层微电阻率扫描测井(FMI)的分辨率非常高,可以区分0.5cm的砾石,被称为“地层显微镜”。成像数据生成图像后,可以与取心照片进行对比,而且更加直观和具有连续性。因此广泛用于识别岩性、裂缝、孔隙等,对于研究沉积环境、地层构造具有很大的优势。
人工神经网络(Artificial Neural Networks)是近年来人们广泛关注的一门学科,结合生物学、数学模型和计算机算法,模拟大脑处理问题的思维,解决实际应用的数据处理问题,取得了巨大的成功。人工神经网络具有很强的容错性、自适应性及联想记忆等特点,用于岩性的识别和预测是可行的。
本文的主要内容是利用成像测井识别砂砾岩岩性,首先研究成像测井的静态图像,掌握研究区内的各种砂砾岩的岩性特征,总结成像模式;然后利用数字图像处理技术,对图像进行预处理和特征提取;最后研究人工神经网络常用的模型BP网、Hopfield网及SOFM网的网络结构和算法流程,论述了人工神经网络识别岩性的过程。通过MATLAB编程,主要采用BP算法进行砂砾岩的岩性识别,取得了很好的效果。
Deep glutenite reservoirs are an important target for petroleum exploration in China. Due to glutenite reservoir which has complex lithology, lithology facies changed rapidly, and the resistivity of rock matrix has a great impact on the electrical logs , it is very difficult to lithology identification in Deep glutenite reservoirs with conventional well logs.
Fullbore Formation MicroImage is promoted by Schlumberger, it has very high resolution is up to 5mm, named as“formation microscope”. The static images generated by FMI data can compared with cores, and are more intuitive and continuous.,and have been widely used to identify lithology, fractures, and pores,and have great advantages in the study of sedimentary environment and formation structure.
ANN (Artificial Neural Networks) is drawing widespread attention in recent years, combined biology, mathematical models and computer algorithms to simulate the brain so as to deal with the problem. It is used to solve data processing , and has achieved great success. Artificial neural network has a strong fault-tolerant, adaptive and associative memory, so it is feasible for lithology identification and prediction.
This paper focuses on identifying glutinite lithology. First, we study the static Image logging data, calibrate borehole image logs to cores, and summarize the features of glutenite rock; Second, we utilize image analysis algorithms to perform quantitative analysis and feature selection to borehole images, and transfer image features to characteristic value; Last, we investigate the network structure and study three models of ANN—BP Network, Hopfield Network and SOM network, especially study BP network and mainly utilize the BP neural network to identify lithology. This paper realizes the BP network to identify well lithology, and achieves satisfactory results using MATLAB code.
引文
[1]董经利,张晋言.东营凹陷北带砂粒岩体天然气测井评价方法[J].测井技术,2007,3:24-27
[2]顾玉君,申晓娟,吴爱红等.泌阳凹陷南部陡坡带砂砾岩储层岩性识别研究[J].石油地质与工程,2009,23(2):40-42
[3]马海,王延江,胡睿等.测井岩性识别新方法研究[J].地球物理学进展,2009,2,(1):263-269
[4]张守谦,顾纯学,曹广华.成像测井技术及应用[M].北京:石油工业出版社,1997,52-53
[5]贾文玉,田素月,孙耀庭.成像测井技术与应用[M].北京:石油工业出版社,2000,15-17
[6]王小军,陈钢花,余杰等.测井相分析技术在岩性油气藏勘探中的应用-以乌夏地区三叠系为例[J].新疆石油地质,2007,28(4):441-443
[7]寻知锋,余继峰.聚类和判别分析在测井岩性识别中的应用[J].山东科技大学学报:自然科学版,2008,27(5):10-13
[8]孙健,周魁,冉小丰等.Bayes判别分析方法在岩性识别中的应用[J].石油天然气学报,2009,31(2):74-77
[9]乔德新.成像测井资料定量计算方法研究及软件开发[D].中国地质大学(北京),北京:2005
[10]章毓晋.图像处理和分析[M].北京:清华大学出版社.1999,31-36
[11]吴元敏.基于Freeman链码的图像中多个区域面积的计算方法[J].计算机工程与应用,2008,44(15):199-201
[12]何登春,鲍金平.地层倾角测井资料速度校正处理技术[J].测井与射孔,1999;23(2):37-41
[13]吴燕.成像测井的计算机实现[D].西安:西安科技大学,2006
[14]宋毅,崔平远,居鹤华.一种图像匹配中SSDA和NCC算法的改进[J].计算机工程与应用,2006:42-44
[15]吴培景,陈光梦.一种改进的SSDA图像匹配算法[J].计算机工程与应用,2005,41(33):76-78
[16]姜凯,陈海霞,汤建华.一种快速图像匹配算法的设计与实现计算机工程与应用[J]. 2004,40(11):64-102
[17]李强,张钹.一种基于图像灰度的快速匹配算法[J].软件学报,2006,17(2):216-222
[18]孙家广,杨长贵.计算机图形学(新版)[M].北京:清华大学出版社,1994:125-130
[19]张丽莉,刘瑞林.两种图像分割算法在FMI成像资料中的应用[J].江汉石油学院学报,1999;4(21):88-90
[20]周云才,刘瑞林.图像识别方法在FMI图像处理中的应用[J].江汉石油学院学报, 2003;6:50-51
[21]陈钢花,吴文圣,毛克宇,等.利用地层微电阻率成像测井识别裂缝[J].测井技术,1999;23(4):279-281
[22]耿会聚,王贵文,李军.成像测井图像解释模式及典型解释图板研究[J].江汉石油学院学报,2002;24(1):26-29
[23]李灿元.成像测井的图像自适应识别方法[D].哈尔滨:哈尔滨工业大学,2003
[24]曹莉.成像测井的图像识别方法[D].哈尔滨:哈尔滨工业大学,2001
[25]王满.基于FMI的火成岩岩组构分析[D].吉林:吉林大学,2007
[26]吴艳.成像测井的计算机实现[D].西安:西安科技大学,2006
[27]谭琨.测井岩性自适应识别[D].吉林:吉林大学,2005
[28]张涛.BP神经网络在测井解释中的应用研究[D].西安:西北大学,2010
[29]李安宗.成像测井地面系统软件的设计与实现[D].西安:西北工业大学,2002
[30]张治国.人工神经网络及其在地学中的应用研究[D].吉林:吉林大学,2006
[31]李茂兵.电成像测井自动识别和定量评价研究[D].东营:中国石油大学(华东),2010
[32]张孝珍.砂砾岩储层测井评价方法研究[D].东营:中国石油大学(华东),2009
[33]石宁.盐家—永安地区沙四上亚段砂砾岩体期次精细划分[D].青岛:中国海洋大学,2009
[34]张挺.基于多点地质统计的多孔介质重构方法及实现[D].合肥:中国科学技术大学,2009
[35]李河.基于构件复用的测井解释系统及成像测井图像处理与自动识别技术研究[D].吉林:吉林大学,2005
[36]朱剑兵.Matlab软件在测井资料处理中的应用[J].油气地球物理,2006,4(1):42-44
[37]尤征,杜旭东,侯会军,等.成像测井解释模式探讨[J].测井技术,2000,24(5):393-398
[38]熊伟,运华云,赵铭海,等.成像测井在砂砾岩体勘探中的应用[A].2009,31(1):48-51
[39]薛林福,潘保芝.用自组织神经网络自动识别岩相[A].1999,29(2):144-147
[40] Dezayes.C,Villemin.T.Inherited and induced fractures characterized from acoustic and electric borehole images,1996:197-212
[41] Mikhail H Atallah.Faster Image Template Matching in the Sum of Absolute Value of Diferences Measure[J].IEEE Trans on Image Processing,2001,10(4):659-662
[42] LuthiS.M,Souhaite.P.Fracture apertures from electrical borehole scans[J].Geophysics,1990;55 (7):821-833
[43] Luigi Di Stefano,Stefano Matocia, federico Tombari.ZNCC-based template matching
[44] Schlumberger.Serra O.Formation MicroScanner Image Interpretation.Schlumberger,1991
[45] Delhomme J.P,Colley N.M.Permeability and porosity up scaling in the near-wellbore domain-the contribution of borehole electrical images.Society of Petroleum Engineers,1996;6(1):89-101
[46] Chai Hua,Li Ning,Xiao Chengwen,et al.Automatic discrimination of sedimentary facies and lithologies in reef-bank reservoirs using borehole image logs[J].APPLIED GEOPHYSICS,2009,1(6):17-29
[47] Atlas.STARFRAC[Z].Express Reference Manual,1999;(1):1-56
[48] Massimo Fomasier,Domenico Toniolo.Fast ,robust and efficient 2D pattern recognition for re-assembling fragmented images[J].PATTERN RECOGNITION ,2005 ,38 :2074-2087
[49] Zhou Shuang-quan,Zhao Da-zun.Neural Network Method for Colorimetry Calibration Of Video Cameras.Journal of Beijing Institute of Technology,2000;No.1:256-258
[50] Hough P.C.Method and means for recognizing complex patterns [P].US Patent,1962;(1):235-238
[2]顾玉君,申晓娟,吴爱红等.泌阳凹陷南部陡坡带砂砾岩储层岩性识别研究[J].石油地质与工程,2009,23(2):40-42
[3]马海,王延江,胡睿等.测井岩性识别新方法研究[J].地球物理学进展,2009,2,(1):263-269
[4]张守谦,顾纯学,曹广华.成像测井技术及应用[M].北京:石油工业出版社,1997,52-53
[5]贾文玉,田素月,孙耀庭.成像测井技术与应用[M].北京:石油工业出版社,2000,15-17
[6]王小军,陈钢花,余杰等.测井相分析技术在岩性油气藏勘探中的应用-以乌夏地区三叠系为例[J].新疆石油地质,2007,28(4):441-443
[7]寻知锋,余继峰.聚类和判别分析在测井岩性识别中的应用[J].山东科技大学学报:自然科学版,2008,27(5):10-13
[8]孙健,周魁,冉小丰等.Bayes判别分析方法在岩性识别中的应用[J].石油天然气学报,2009,31(2):74-77
[9]乔德新.成像测井资料定量计算方法研究及软件开发[D].中国地质大学(北京),北京:2005
[10]章毓晋.图像处理和分析[M].北京:清华大学出版社.1999,31-36
[11]吴元敏.基于Freeman链码的图像中多个区域面积的计算方法[J].计算机工程与应用,2008,44(15):199-201
[12]何登春,鲍金平.地层倾角测井资料速度校正处理技术[J].测井与射孔,1999;23(2):37-41
[13]吴燕.成像测井的计算机实现[D].西安:西安科技大学,2006
[14]宋毅,崔平远,居鹤华.一种图像匹配中SSDA和NCC算法的改进[J].计算机工程与应用,2006:42-44
[15]吴培景,陈光梦.一种改进的SSDA图像匹配算法[J].计算机工程与应用,2005,41(33):76-78
[16]姜凯,陈海霞,汤建华.一种快速图像匹配算法的设计与实现计算机工程与应用[J]. 2004,40(11):64-102
[17]李强,张钹.一种基于图像灰度的快速匹配算法[J].软件学报,2006,17(2):216-222
[18]孙家广,杨长贵.计算机图形学(新版)[M].北京:清华大学出版社,1994:125-130
[19]张丽莉,刘瑞林.两种图像分割算法在FMI成像资料中的应用[J].江汉石油学院学报,1999;4(21):88-90
[20]周云才,刘瑞林.图像识别方法在FMI图像处理中的应用[J].江汉石油学院学报, 2003;6:50-51
[21]陈钢花,吴文圣,毛克宇,等.利用地层微电阻率成像测井识别裂缝[J].测井技术,1999;23(4):279-281
[22]耿会聚,王贵文,李军.成像测井图像解释模式及典型解释图板研究[J].江汉石油学院学报,2002;24(1):26-29
[23]李灿元.成像测井的图像自适应识别方法[D].哈尔滨:哈尔滨工业大学,2003
[24]曹莉.成像测井的图像识别方法[D].哈尔滨:哈尔滨工业大学,2001
[25]王满.基于FMI的火成岩岩组构分析[D].吉林:吉林大学,2007
[26]吴艳.成像测井的计算机实现[D].西安:西安科技大学,2006
[27]谭琨.测井岩性自适应识别[D].吉林:吉林大学,2005
[28]张涛.BP神经网络在测井解释中的应用研究[D].西安:西北大学,2010
[29]李安宗.成像测井地面系统软件的设计与实现[D].西安:西北工业大学,2002
[30]张治国.人工神经网络及其在地学中的应用研究[D].吉林:吉林大学,2006
[31]李茂兵.电成像测井自动识别和定量评价研究[D].东营:中国石油大学(华东),2010
[32]张孝珍.砂砾岩储层测井评价方法研究[D].东营:中国石油大学(华东),2009
[33]石宁.盐家—永安地区沙四上亚段砂砾岩体期次精细划分[D].青岛:中国海洋大学,2009
[34]张挺.基于多点地质统计的多孔介质重构方法及实现[D].合肥:中国科学技术大学,2009
[35]李河.基于构件复用的测井解释系统及成像测井图像处理与自动识别技术研究[D].吉林:吉林大学,2005
[36]朱剑兵.Matlab软件在测井资料处理中的应用[J].油气地球物理,2006,4(1):42-44
[37]尤征,杜旭东,侯会军,等.成像测井解释模式探讨[J].测井技术,2000,24(5):393-398
[38]熊伟,运华云,赵铭海,等.成像测井在砂砾岩体勘探中的应用[A].2009,31(1):48-51
[39]薛林福,潘保芝.用自组织神经网络自动识别岩相[A].1999,29(2):144-147
[40] Dezayes.C,Villemin.T.Inherited and induced fractures characterized from acoustic and electric borehole images,1996:197-212
[41] Mikhail H Atallah.Faster Image Template Matching in the Sum of Absolute Value of Diferences Measure[J].IEEE Trans on Image Processing,2001,10(4):659-662
[42] LuthiS.M,Souhaite.P.Fracture apertures from electrical borehole scans[J].Geophysics,1990;55 (7):821-833
[43] Luigi Di Stefano,Stefano Matocia, federico Tombari.ZNCC-based template matching
[44] Schlumberger.Serra O.Formation MicroScanner Image Interpretation.Schlumberger,1991
[45] Delhomme J.P,Colley N.M.Permeability and porosity up scaling in the near-wellbore domain-the contribution of borehole electrical images.Society of Petroleum Engineers,1996;6(1):89-101
[46] Chai Hua,Li Ning,Xiao Chengwen,et al.Automatic discrimination of sedimentary facies and lithologies in reef-bank reservoirs using borehole image logs[J].APPLIED GEOPHYSICS,2009,1(6):17-29
[47] Atlas.STARFRAC[Z].Express Reference Manual,1999;(1):1-56
[48] Massimo Fomasier,Domenico Toniolo.Fast ,robust and efficient 2D pattern recognition for re-assembling fragmented images[J].PATTERN RECOGNITION ,2005 ,38 :2074-2087
[49] Zhou Shuang-quan,Zhao Da-zun.Neural Network Method for Colorimetry Calibration Of Video Cameras.Journal of Beijing Institute of Technology,2000;No.1:256-258
[50] Hough P.C.Method and means for recognizing complex patterns [P].US Patent,1962;(1):235-238