基于混合模糊神经网络储层裂缝地震反演研究
|
摘要
基于储层裂缝系统具有非线性特征,储层裂缝地震反演是由遗传算法(GA)、模糊神经网络(ANF IS)和禁忌搜索算法(TS)有机地结合而构成的自适应混合模糊神经网络技术。该技术在成像测井约束下,形成的自适应混合算法分别训练ANF IS网络的前提参数和结论参数,从而获得满足精度要求的储层裂缝密度的最佳估计值。针对目标储层段,应用储层裂缝地震反演方法对过井地震剖面和联井地震剖面进行了储层裂缝密度反演处理,获得了可用于地质解释和油气预测的视裂缝密度剖面。这种裂缝密度剖面含有裂缝定量信息,其裂缝密度相对误差为:0.8%~24%,满足勘探开发的要求。经与研究区的地质对比分析表明,视裂缝密度剖面上的裂缝展布特征符合研究区的沉积相分布和岩石力学性质的变化特征,对研究区的勘探开发具有重要意义。
Reservoir fracture system is of non-linear character.The seismic inversion of reservoir fracture is the adaptive mixed fuzzy neural network technique which is composed of combination of genetic algorithms(GA),adaptive neural fuzzy inference system(ANFIS) and tabu search algorithms(TS).The parameters of ANFIS are exercised by adaptive mixed algorithms in seismic inversion of reservoir fracture under restriction of formation micro-resistivity imaging log(FMI),and so get optimal estimated value of reservoir fracture density.This method is applied to real target reservoir and the apparent fracture density section is obtained.The section has quantitative fracture information which can be used for geological interpretation and oil-gas prediction.The relative error of fracture density is 0.8% to 24%.Compared with the geology of target area,It is showed that fracture distributing character on apparent fracture density section is in accordance with sedimentary facies and rock geomechanic property.It is of important significance to exploration and development in target area.
Reservoir fracture system is of non-linear character.The seismic inversion of reservoir fracture is the adaptive mixed fuzzy neural network technique which is composed of combination of genetic algorithms(GA),adaptive neural fuzzy inference system(ANFIS) and tabu search algorithms(TS).The parameters of ANFIS are exercised by adaptive mixed algorithms in seismic inversion of reservoir fracture under restriction of formation micro-resistivity imaging log(FMI),and so get optimal estimated value of reservoir fracture density.This method is applied to real target reservoir and the apparent fracture density section is obtained.The section has quantitative fracture information which can be used for geological interpretation and oil-gas prediction.The relative error of fracture density is 0.8% to 24%.Compared with the geology of target area,It is showed that fracture distributing character on apparent fracture density section is in accordance with sedimentary facies and rock geomechanic property.It is of important significance to exploration and development in target area.
引文
[1]李勇,邵泽辉,李琼.地震优化非线性反演方法及应用研究[J].矿物岩石,2004,24(1):101-104
[2]李正文,李琼,吴朝荣.沉积盆地有效储集层综合识别技术[M].成都:四川科学技术出版社,2002.
[3]李正文,李琼.油气储集层裂缝非线性预测技术及应用研究[J].石油地球物理勘探,2003,38(1):48-52.
[4]周文.裂缝性油气储层评价方法[M].成都:四川科学技术出版社,1998;38(1):48-52.
[5]贺振华,黄德济.缝洞储层的地震检测和预测勘探[J].地球物理进展,2003;26(2):79-83.
[6]A ndreas R iiger等.马玉春译.用AVO方法检测裂缝:解析基础和实用解[J].国外油气勘探,1998;10(4):41-44.
[7]R on it S trah ilev ityect.伊墁坦译.用P波AVO进行裂缝探测.SEG第65届年会论文集[M].北京:石油工业出版社,1996:162-165.
[8]王小平,曹立明.遗传算法-理论、应用与软件实现[M].西安:西安交通大学出版社,2002.
[9]李琼,李正文,魏野.同铁构造嘉陵江组储层裂缝非线性预测与分析研究[J].矿物岩石,2004,24(2):78-81.
[10]陈章良.遗传算法及其应用[M].北京:人民邮电出版社,2001,2.
[11]Jang J S R,ANF IS:A daptive-N etw ork-Based Fuzzy In ference System[J].IEEE Transaction on System,M an and Cybernetics,1993,23(3):665-685.
[12]G o ldberg D E.G enetic A lgor ithm s in Search,O ptim ization,and m ach ine L earn ing[M].R ead ing,M A,A dd ison-W esley,1989.
[13]S ie ism a J,D ow R J F.N eura l net prun ing-W hy and how[J].P roceed ings of IEEE In ternationa l C on ference on N eura l N etworks,1998,(1):325-333.
[14]G lover F.Fu ture paths for In teger P rogramm ing and links to IEEE N cura lN etw orks C ouncil[J].C om puters and O perations R esearch,1986,5:533-549.
[2]李正文,李琼,吴朝荣.沉积盆地有效储集层综合识别技术[M].成都:四川科学技术出版社,2002.
[3]李正文,李琼.油气储集层裂缝非线性预测技术及应用研究[J].石油地球物理勘探,2003,38(1):48-52.
[4]周文.裂缝性油气储层评价方法[M].成都:四川科学技术出版社,1998;38(1):48-52.
[5]贺振华,黄德济.缝洞储层的地震检测和预测勘探[J].地球物理进展,2003;26(2):79-83.
[6]A ndreas R iiger等.马玉春译.用AVO方法检测裂缝:解析基础和实用解[J].国外油气勘探,1998;10(4):41-44.
[7]R on it S trah ilev ityect.伊墁坦译.用P波AVO进行裂缝探测.SEG第65届年会论文集[M].北京:石油工业出版社,1996:162-165.
[8]王小平,曹立明.遗传算法-理论、应用与软件实现[M].西安:西安交通大学出版社,2002.
[9]李琼,李正文,魏野.同铁构造嘉陵江组储层裂缝非线性预测与分析研究[J].矿物岩石,2004,24(2):78-81.
[10]陈章良.遗传算法及其应用[M].北京:人民邮电出版社,2001,2.
[11]Jang J S R,ANF IS:A daptive-N etw ork-Based Fuzzy In ference System[J].IEEE Transaction on System,M an and Cybernetics,1993,23(3):665-685.
[12]G o ldberg D E.G enetic A lgor ithm s in Search,O ptim ization,and m ach ine L earn ing[M].R ead ing,M A,A dd ison-W esley,1989.
[13]S ie ism a J,D ow R J F.N eura l net prun ing-W hy and how[J].P roceed ings of IEEE In ternationa l C on ference on N eura l N etworks,1998,(1):325-333.
[14]G lover F.Fu ture paths for In teger P rogramm ing and links to IEEE N cura lN etw orks C ouncil[J].C om puters and O perations R esearch,1986,5:533-549.
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心