【作者】 王康宁；

【导师】 段云卿；

【作者基本信息】 中国地质大学（北京） ， 地球探测与信息技术， 2010， 硕士

【摘要】 随着油气勘探程度的不断深入,寻找新的油气愈加困难,勘探重点已经从构造油气藏等较容易发现的油气藏逐渐过渡到以岩性、薄层等复杂油气藏的勘探为主。由于地震数据分辨率的有限性,使得储层预测的精度和识别薄层的能力受到限制。而反演技术可以充分结合测井数据垂向上的高分辨力和地震数据的横向连续性,已经成为为储层预测过程中最重要的一种技术手段。地震反演根据应用资料的不同又分为叠后和叠前反演,而叠后反演作为一种较为成熟的方法目前应用更为广泛。本文调研了现有各种叠后反演方法的发展情况,并选取了其中道积分反演,递推反演,基于模型反演及基于模式识别算法的分频反演进行了深入的理论学习及各自适用情况优缺点分析,并选取具有代表性的几种方法在塔河油田卡拉沙依组碎屑岩储层进行实际应用研究。塔河油田碎屑岩的勘探开发很好的潜力和前景,但由于储层复杂、地震资料品质不能满足要求等原因,勘探开发的前进步伐受到限制。针对该研究工区特征,完成了地震资料零相位化处理,测井资料处理、地层对比划分、精细构造解释、基于自然电位目标曲线重构等一系列工作并在此基础上进行了稀疏脉冲反演、地质统计学反演和分频反演实验,对比了几种不同方法的实现结果并进行分析。通过实际应用对比,稀疏脉冲反演结果分辨率较低,与地震数据的相关性好,反演结果稳定,但不能达到分辨薄层的要求。该方法适用于勘探初期井数据较少的情况,并且其结果可在随机反演过程中起到横向上的约束作用。地质统计学反演引入了地质统计学的信息,将地震反演方法与随机模拟理论进行充分的结合,得到的结果可以分辨薄层,综合了地震数据的横向分辨率和测井数据的垂向分辨率的优势。该方法具有模型化现象,与地震相关性较低,其结果具有多解性,得到的是多个等概率结果,需结合实际地质情况进行风险估计。分频反演是结合小波分解技术,支持向量机算法,及AVF原理等多种理论的一种较新的反演方法,得到的结果不受子波的影响,与地震数据相关性好,分辨率较高,能适应薄储层的要求,对提高该区储层预测精度具有一定的价值。更多还原

【Abstract】 As the exploration of oil and gas become more and more extensive, it become more and more difficult to find new oil and gas, the emphasis of exploration has been transferred from the reservoirs easy to find such as structural reservoirs to complicated reservoirs such as lithology and thin bed. Because of the limitation of seismic data resolution, the accuracy of reservoir prediction and the ability of identifying thin reservoirs are limited. The inversion technology can connect the high resolution of the vertical logging data and the lateral continuity of the seismic data, so it has become the most important technology in reservoir prediction.The seismic inversion can be divided in to post-stack and pre-stack inversion according to the different application data, and the post-stack inversion has been used more extensively for its maturity. The original studies the development situation of various post-stack inversion methods and choose the integral inversion, recursive inversion, inversion based on model and the frequency-divided inversion to do extensive theoretical study and relative merits analysis, and choose representative methods to do practical application research in Kalashayi series clastic reservoirs in Tahe oil field.The exploration of Tahe oil field has great potential and foreground, but because of the complexity of reservoir and the seismic data quality cannot satisfy the requirement, the development of exploration is limited. According to the characters of the research work area, finish the zero-phasing process of the seismic data, logging data process, layer comparison division, fine structure interpretation, target curve rebuilt based on the SP curve. And based on these work finish the sparse spike inversion, geostatistics inversion and frequency-divided inversion experiment, compare the different results of the various methods and do analysis.Through comparison of the practical applications, we can find that the sparse spike inversion has low resolution, high correlation with seismic data and stable inversion result, but cannot meet the requirement of flaggy resolution. This method is proper for the situation in which there is less logging data and the result can be the vertical limitation of the geostatistics inversion process. The geostatistics inversion method in take the information of geostatistics, connect thoroughly the seismic inversion method and Stochastic simulation Theory, the results can be used in flaggy resolution, which integrate the lateral resolution of seismic data and the vertical resolution of logging data. This method has the model, low correlation with seism, and the result has multiplicity, can lead to several random results, so the risk estimation is needed according to practical geology, Frequency division inversion is a new method that connects the wavelets decomposition technology, supporting several theories such as the vector machine algorithm and AVF principle, the results are not influenced by the wavelets, having high correlation with seismic data, high resolution, meeting the requirement of flaggy reservoir. So the method has high value of raising the accuracy of the reservoir.更多还原