【作者】 魏野；

【导师】 李正文；

【作者基本信息】 成都理工大学 ， 地球探测与信息技术， 2004， 硕士

【摘要】 地震信息是储层各种特征的综合反映体,它包括了储层岩性、物性、流体、裂缝等诸多特征;另一方面,地震信息又是几十种乃至上百种参数的综合,每一种地震参数从不同侧面不同程度反映了储层或气藏的特征,任何单一参数储层预测都有局限性和多解性,并给勘探带来较大风险。尤其是碳酸盐岩薄储层,由于薄层干涉以及分辨率低,用地震速度反演或单一参数作储层预测往往效果不理想。 对同一构造单元来说,储层岩性和物性横向相对稳定,可看作是系统常量,而流体、裂缝等就是地震信息函数的变量,各种参数与气藏之间必然存在着联系,参数综合预测就是要分析这种相关性。以往我们是通过人工方法进行,即在M个特征参数中选取K个特征组成新的特征向量。这种做法需要掌握详尽的地质、测井等资料,并了解各种参数的适用范围,直接删除那些在该地区效果不好的参数,然而这样选择特征参数工作量太大,做法也比较粗糙,人为性较强。一般来说,M个特征均来自反映被处理对象的某种特征,简单删除不能全面刻划原对象。 通过数学及计算机手段,将地震综合信息分解成多种单一参数,对这些参数作归一化、压缩降维,采用广义线性分类判别及加权综合处理,可获得能反映储层特征的综合参数,提高地震参数法预测的精度,在实际应用中也见到了良好的效果。 碳酸盐岩薄储层气藏地震参数综合预测法是由地震数据输入、储层地震参数提取、参数分析和储层综合预测等组成。该预测系统中信息分解、分析、压缩、综合均由计算机自动实现,减少了人为性,可客观地突出与油气存在有关的参数,预测出油气储层的最佳部位和局部油气富集带。 本文所用的参数综合预测不是参数简单筛选、人工综合分析或叠加,而是结合钻探实际,采用程序化、科学的方法加权综合为一种能反映气藏的特征参数。发展了碳酸盐岩薄储层预测的新思维和新方法技术,在储层预测的深度上获得较好的新进展和重要成果。更多还原

【Abstract】 Seismic information is compositive reflection of all kinds of characteristic in reservoir bed, which includes lithologic character, physical property, fluid, fracture etc. On the other hand, seismic information integrates tens of seismic parameters anyone of which reflected character of reservoir bed or gas reservoir in different side and different degree. Prediction with one parameter has limitation and multi-interpretation and bring risk to petroleum exploration. Especially, prediction with single parameter or seismic velocity reversion is poor in thin carbonate bed.To the same tectonic unit, lithologic character and physical property of reservoir bed are stable relatively, which can be thought of as constant of system, while fluid, fracture etc are variable to seismic information function. All kinds of seismic parameter must have relations to gas reservoir , parameters comprehensive prediction just about analyze these relativity. Before, we do it by manual work, that’s to say, new characteristic vector is composed of K kind characteristic parameters selected from M kind characteristic parameters. This method needs hold exhaustive geologic, logging datum, know applicability of and delete those parameters not adapt to the area, while this method will bring on mass workload , rough result and strong jamming. In a general way, M kind characteristic parameters all come from certain character which reflected object and deleting simply can not depict original object.Seismic compositive parameter can be devided into multiple single parameters by mathematics and computer means. By deposing these parameters, such as normalization, reducing dimensions, generalized linear classify differentiating and statistical weight comprehensive disposal etc, a comprehensive parameter can be acquired reflecting character of reservoir bed and prediction precision of seismic parameter method can be improved. This method also get good effect.Seismic parameter comprehensive prediction method is composed of inputing seismic data, extracting seismic data of reservoir bed, analyzing parameter and predicting gas reservoir synthetically etc. This prediction system can finish disassociation, analyzing, compressing, integrating of information by computer. This will reduce subjectivity of man, stand out parameters relative to gas/oil, predict best part of reservoir bed rich of oil/gas and local area rich of oil/gas.Seismic parameter comprehensive prediction method used in this paper is not filtering parameters and analyzing artificially simply, but integrating a new parameter reflecting character of gas reservoir, by combining with drillhole exploration fact, adopting routinization and scientific method. This method develops new method and technology about thin reservoir bed in carbonate rock, acquired better new progression and important outcome in prediton of reservoir bed.更多还原