【作者】 杨世夺；

【导师】 郑浚茂；

【作者基本信息】 中国地质大学（北京） ， 矿产普查与勘探， 2013， 博士

【摘要】 非均质性是碳酸盐岩储层的主要特点和挑战。非均质性主要来源于裂缝、溶蚀、岩性分布及发育的多样性和不确定性。岩性的分布受沉积物的来源、沉积环境和后期成因作用多种因素的控制。裂缝发育受到岩性、埋藏深度、局部构造和区域水平地应力的综合影响。岩性、古地形、古气候和海平面的升降变化则控制了溶蚀和喀斯特的发育区域。在沉积微相识别的基础上，应用高分布率层序地层学是预测岩性变化的方法之一。它以岩心、测井、露头和高分辨率地震反射剖面为基础，通过精细层序划分和对比技术，建立各种级别的成因地层格架，对各种沉积体进行评价和预测。然而，高分辨率层序地层学的分析结果受到地震资料、常规测井低分辨率和有限钻井取心资料的制约。在本文中提出了一套新的基于微电阻率成像测井的进行高分辨率层序地层学研究及有利储层预测的分析方法。首先应用钻井取心资料对比分析微电阻率成像典型沉积构造特征，并在此基础上应用有监督神经网络的方法进行多井沉积微相的识别。然后，综合考虑地震和微电阻率成像资料，划分不同级别的基准面旋回；并根据沉积纹层厚度和纵向上沉积相的变化分析地层的叠置关系。在层序地层学分析的基础上，恢复碳酸盐岩古地貌，进而总结出溶蚀和喀斯特在层序框架内的发育特点。最后根据地质构造来研究裂缝的发育特点，并预测有利的储层区域。在塔中82区块实施和应用了上述分析方法：根据3口井的取心资料识别出7种不同的沉积微相，应用有监督神经网络方法建立单井沉积微相，并推广到其他4口无钻井取心井。结合地震资料和微电阻率成像划分出3种不同级别的基准面旋回，在分析确定地层叠加模式的前提下，建立本区块的层序地层框架。应用顶部长期基准面旋回的厚度进行残留古地貌恢复，结合单井次生孔隙计算结果，分析溶蚀作用的发育规律。通过单井裂缝参数和灰岩顶部构造曲率的计算，研究裂缝发育的控制因素。在分析溶蚀和裂缝发育规律的基础上，预测有利储层的分布范围。最近新钻井试油资料验证了相关分析结果，证明了该分析流程的可靠性和可行性，为陆棚边缘厚层碳酸盐有利储层预测提供了解决方案，具有一定的参考价值和应用前景。更多还原

【Abstract】 The complexity in carbonate reservoirs is formation heterogeneity due to fractures,vugs, and mixed facies. The lithology distribution is controlled by multiple effects, such assediment source, depositional environment, and diagenesis. Fracture development isinfluenced by lithology, burial depth, local structure and far-field stress. Lithology, ancienttopography, ancient weather, and sea level change control the vugs and Karst development.High-resolution sequence stratigraphy is one of the advanced methods to resolve thelithology challenge with facies identification. The method combines core analysis,conventional logs, outcrop studies, and seismic data to build multiple-ranks sequence andthen predict the sedimentary distribution. However, the analysis results are frequentlyconstrained by the low resolution of the seismic and conventional log data and by limitedcore data.In this paper, we propose a new workflow for high-resolution sequence-stratigraphyanalysis by integrating borehole resistivity images. First, the borehole resistivity images arecalibrated with core data; then the depositional facies are identified from calibratedresistivity-image data combined with multiple-domain data through supervised neuralnetwork method. Second, multiple-ranks base levels are identified from seismic andborehole resistivity image data; and the thickness change of cross bedding and depostionalfacies column are used to classify the strata stacking patterns. Third, the ancient topographymap was built from sequence-stratigaphy analysis and then the vugs and Karst distributionwas summarized within a sequence stratigraphy framework. Finally, the sweet spot waspredicted combining with fracture evaluation from seismic structure.We present a case study where this new approach was implemented in the Block A ofthe Tazhong uplift in the Tarim basin. Seven different depositional facies were identifiedfrom three wells-core data to build the continue facies column for every well after appliedto additional4wells and non-cored intervals.3different base-levels were classified fromborehole resistivity image and seismic data. Sequence framework was built after stratastacking patterns identified. The thickness of top long term base level was used to analyzethe vugs distribution by integrating with single well secondary porosity results fromborehole micro-resistivity image. The local structural curvature controlled the fracture distribution based on single well fracture parameters and curvature computation. The sweetarea was predicted from vugs and fracture potential distribution. A recently drilled wellconfirmed these results and approved the workflow can be applied to similar thickcarbonate reservoirs in shoal-reef margin of a carbonate platform.更多还原