论文部分内容阅读
松辽盆地陆相三角洲沉积的主力储层为分流河道砂体,其以“砂薄、砂地比低、横向分布不稳定”为特点.针对分流河道砂体平面展布及厚度预测难题,本文提出在对振幅、频率、相位、波形及相关等多种地震属性提取基础上进行优选组合,确定输入的最佳属性组合,应用支持向量机方法预测河道平面展布;同时应用测井曲线重构、随机反演方法,提高垂向分辨率,有效识别分流河道砂体厚度这一综合刻画方法.文中以G油田扶余油层F205沉积单元为例,识别出3条断续型分流河道沉积(河道1为基于井落实河道,河道2、3为地震预测河道),均为北北西走向.研究区中部的预测河道2在F205沉积单元内预测延伸长度为800 m,根据随机反演结果显示厚度在5 m左右.河道2相对于研究区东部预测河道3厚度大,延伸长,是下一步井位部署的潜力区.截止目前,在预测河道2中部设计了一口评价井G12-9,F205沉积单元钻遇砂岩厚度4.8 m,有效厚度为4.6 m,试油方式为压后抽汲,试油结论为高产工业油层.后经岩心、测井曲线等证实,该井为典型的河道砂沉积体.这充分证明了应用优选地震属性的支持向量机方法刻画河道平面展布,结合测井曲线重构参与的随机反演方法刻画河道厚度,在三角洲沉积体系分流河道的综合刻画上是有效的.
The main reservoir of the sedimentary facies in the Songliao Basin is distributary channel sandbody, which is characterized by “thin sand, low sand-land ratio and unstable lateral distribution.” Aiming at the problem of plane distribution and thickness prediction of distributary channel sand bodies, In this paper, we propose a combination of the best attributes based on the extraction of seismic attributes, such as amplitude, frequency, phase, waveform and correlation, and determine the best combination of attributes by using SVM to forecast the river plane distribution. At the same time, Structure, stochastic inversion method, vertical resolution enhancement and effective identification of distributary channel sand body thickness.This paper takes F205 sedimentary unit of Fuyu oil field of G Oilfield as an example to identify three intermittent distributary channel sediments 1 is based on well implementation of the river, river 2, 3 for the earthquake prediction of the river), are North-North-West trend in the study area of the predicted river 2 in the F205 sedimentary unit predicted extension of 800 m, according to the random inversion results show that the thickness of 5 m or so.Corresponding to the thickness and extension of the predicted channel 3 in the east of the study area, the channel 2 is the potential area for the next wellbore deployment.As of now, a well G1 2-9, F205 Sedimentary unit sandstone thickness of 4.8 m, effective thickness of 4.6 m, test oil pressure pumping after drawing, the conclusion of the test oil for the high-yield industrial reservoir after the core, well logging curve confirmed that the well was Which is a typical channel sand deposit.This fully proved that the application of the preferred seismic attribute support vector machine method to depict the river plane distribution, combined with the well logging curve reconstruction involved random inversion method to characterize the thickness of the river in the delta system, Characterization is effective.