【作者】 陈常乐；

【导师】 刘洋；

【作者基本信息】 吉林大学 ， 固体地球物理学， 2012， 硕士

【摘要】 随着油气勘探开发程度的日益深入，非常规油气资源逐渐成为国家油气战略的重要组成部分，中国的火山岩储层具有很大的生油潜力，尤其在东北松辽盆地，火山岩广泛发育，如何有效地识别火山岩储层具有重要意义。我国火山岩储层和油气藏的勘探开发始于20世纪70年代，于20世纪90年代中后期得到突飞猛进的发展。目前，我国几乎所有的主要油气盆地内部都发现了火山岩和火山岩储层。在现今的油气勘探过程中，火山岩油气藏的描述主要是对有利储层和含气性的预测。而对火山岩油气藏描述的基础是火山岩地层的识别，如何进行火山岩地层的识别是一个重要的研究焦点。因此，针对火山岩油气藏，利用各种地球物理方法技术开展火山岩储层识别与预测具有重要的现实意义。由于火山岩储层所固有的岩性以及储层空间极其复杂，使得火山岩储层的预测非常困难。地震波属性是一种有效的储层特征分析工具，是研究火山岩空间分布的参数。频率、地震倾角、曲率等属性是利用地震数据识别火山岩的敏感特征。在本次研究中，作者提出利用地震波局部属性的方法来进行火山岩的识别，首先考虑瞬时频率，对实际地震数据进行瞬时频率的求取之后发现剖面上有大量负频率值，在传统意义下，这暂时是不符合实际情况的。于是，引入局部频率，局部频率通过整形正则化来确定平滑半径，仅需要平滑半径一个参数即可以求取局部频率。对地震剖面求取的局部频率，没有野值存在，能够反映单个数据点周围局部范围的频率信息。通过平面波分解滤波器(PWD)求取局部倾角，作者推导近似的同相轴局部曲率，通过对理论模型的计算并进行误差计算，结果表明，近似的同相轴局部曲率可以很好的逼近真实值，具有较高的计算精度。通过计算实际地震剖面的局部频率，在局部频率数据体上圈定低频的区域范围，对同一剖面计算曲率，在曲率数据体上将高曲率值区域圈出，在地震剖面上将这两个区域进行叠合显示，能够自动识别出火山岩体可能的分布区域，为进一步识别火山岩储层提供一定的指导作用。作者对实际地震数据的不同地震波属性之间做了交汇图，并且对数据进行筛选，把火山岩可能发育的部分截取出来，针对这部分地震数据进行瞬时振幅、局部频率、局部相位、局部倾角的计算，并对不同属性之间进行交汇显示，试图分析不同属性的相关性。结果表明，在交汇图中火山岩没有明显相关属性，传统的交汇图方式很难对储层的预测进行指导，因此为下一步的研究探索指出了方向。更多还原

【Abstract】 With the deepening of oil and gas exploration and development, unconventionaloil and gas resources is becoming an important part of the national oil and gasstrategy, the volcanic reservoirs in China has great potential for oil generation,especially in the northeast of Songliao Basin, volcanic rocks are widely developed,and how effective to identify the volcanic reservoir is significant. Volcanic reservoirexploration and development began in the1970s, the rapid development appeared inthe late1990s. At present, most major oil and gas basin interior show the reservoir ofvolcanic rocks or volcanic rocks.In today’s oil and gas exploration process, the description of the volcanic rockreservoirs is aiming at the favorable reservoir and gas-bearing properties forecasts.Describing on the volcanic rock reservoir is based on the identification of the volcanicrocks, volcanic strata of recognition is an important research focus. Therefore,according to the volcanic rock reservoir, using a variety of geophysical methods tocarry out the volcanic reservoir identification and prediction has important practicalmeaning. Because the volcanic reservoir lithology and inherent reservoir space isextremely complex, the prediction of volcanic rock reservoirs is very difficult.Seismic attribute is an effective reservoir characterization tool and is the keyparameter for study of the spatial distribution of volcanic rocks. The local seismicfrequency, local seismic dip, local curvature of event, and other properties aresensitive features to identify the volcanic rocks in seismic data.In this study, the author proposes to use local seismic attributes for theidentification of volcanic rocks. First of all, I consider the instantaneous frequencywhich always shows a large number of negative frequency values in the section, it isobviously unreasonable. Thus I introduce the local frequency. Local frequencydetermine the smoothing radius by shaping regularization, smooth radius is only oneparameter that can constrain the underdetermined problem.I used plane-wave destruction filters (PWD) to calculate local dip and deducedthe approximate local curvature of event by using local dip. Numerical tests show thatthe local curvature of event can be a good approximation of the true value and has high accuracy.By computing the local frequency of the actual seismic profiles, I draw thelow-frequency region from the body of the local frequency data. I also calculated thecurvature of the same profile and get the high curvature region. The two attributes byusing overlay display can identify the distribution of volcanic body and furtherprovided the guidance for the identification of the volcanic reservoir.Cross plot between the different seismic attributes, including the instantaneousamplitude, the local frequency, local phase, and local dip, was developed to find therelationship between two different attributes. However, the results show that, in thecrossplot map, there is no clear trend for guidance of the volcanic reservoir prediction.更多还原