【作者】 王孝；

【导师】 贺振华； 雍学善；

【作者基本信息】 成都理工大学 ， 地球探测与信息技术， 2011， 博士

【摘要】 中国西部地区实际地形和地表结构复杂,静校正问题是地震成像的瓶颈。随着石油地震勘探向黄土塬、沙漠、山地等复杂地区发展,这个瓶颈成了急需解决的任务。为此,本文对西部复杂地表条件下静校正方法进行了较全面的研究,并取得了有意义的进展。第一、西部复杂地表条件下的静校正,需要解决的首要问题是近地表速度结构建模。因为在这些地区,小折射、微测井等实测资料直接建立模型的精度已经难以满足勘探需要,折射波静校正在无稳定折射层条件下的应用效果受到局限,层析静校正方法虽能适应这种复杂地表条件,但也受限于实际观测数据的影响。为此,本文研究了一种基于实测资料约束的层析反演近地表建模方法。提出了采用小折射、微测井等地学数据,借助克里金办法约束反演近地表建模,提高近地表建模精度,实现了约束层析反演静校正方法。该方法提高了模型建立的精度,为复杂地表区更加准确的近地表模型建立提供了新方法。第二、需要采用非地表一致性静校正方法。目前,生产中使用的各种静校正方法,基本上都以地表一致性假设为基础。假设近地表低速带介质的速度远低于其下伏地层的速度,地震波在低速带内垂直传播,不同类型的地震波在低速带内都沿同一路径传播,即在同一炮点或检波点处,所有地震波的时间延迟相同。但在西部复杂区,地形起伏剧烈,低速带速度纵横向变化很大,使得地震波到达不同检波点的时间延迟变得复杂,实际的近地表介质结构与地表一致性假设的情况差异很大。这样,就在地表一致性静校正与实际静校正之间产生了较大偏差,既出现了“静校不静”的问题,我们把这种偏差称为非地表一致性静校正。本文研究了这种偏差产生的原因,提出了减小这种偏差的非地表一致性静校正方法。第三、研究了配套的解决高频成像问题多种静校正量分离和重构技术,见到了明显的应用效果。第四、对长期以来困扰地球物理界的长波长静校正问题进行了初步探讨,并提出了一种解决长波长静校正问题的空间固定排列法,实现了交互确定长波长静校正边界来解决长波长静校正问题,见到了初步效果。最后对基于波场延拓的基准面校正进行了尝试,并开展了模型试验。更多还原

【Abstract】 The factual surface and structure are very complex in the western of China , static correction is the choke in seismic imaging. This problem is becoming the chiefly task with the development of seismic exploration in Loess Plateau Tableland, desert and Mountainland. Research on static correction in the western with complex surface conditionhas been done and some progress has been achieved in this article.Firstly, near surface velocity modeling is the vital problem in static correction in the western with complex surface condition. In the western the model built with small refraction and micro-well logging data directly can not satisfy seismic exploration requirements. The effect of refraction wave static correction is limited by the condition that there are no stable refracting layers, the adaptability of tomography static correction is better but is limited by the factual observation data. So this article researched on the method of near surface modeling with tomography inversion constrained by the factual data using Kriging method, by which the near-surface modeling precision is improved, The model precision has been improved with this method, so the near surface model can be built more exactly in complex surface area.Secondly, a method of non-surface-consistency static correction is necessary. Now most of static correction methods used widely in production are based on the assumption of surface consistency that the velocity of near surface in lower-velocity region is assumed far lowered than that of the down layer, seismic wave is spreading vertically in the region, seismic wave of different types is spreading along the same path in the region, that means the time-delays of seismic wave are same on the same shot point or same receiver point. But in the complex western, acute topography undulation, sharp lateral and longitudinal velocity variations in low velocity regions result in the time-delays of seismic wave arriving at different receiver are complex, the actual near-surface structure is very different from that based on the assumption of surface consistency. So big deviations were produced between the actual static correction and based on the surface consistency. We called the deviation as non-surface-consistency static correction. The reason brings this deviation has been researched in this article, a method of non-surface-consistency static correction has been put forward to deduce the deviation.Thirdly, some methods of department and reconstruction of static correction value have been researched as auxiliary technology, and obvious effect has been achieved.Fourthly, this article also covers some preliminary research on long wavelength static corrections that has been confusing geophysicists for a long time, and put forward a spatial fixed arrangement method for long wavelength static correction, The problems in long wavelength static correction are solved by inactively picking the static correction boundary, good effect with this method has been achieved. The research on datum plane static correction based on wave field continuation also has been covered, its feasibility has been proved by model experiment in this article.更多还原