论文部分内容阅读
象澳大利亚西北大陆架那样的复杂地质构造,常规地震处理不能正确地成像,断裂引起的横向变化使得下伏构造也不能完全成像,而对于这种构造,叠前深度偏移却是一种正确的成像方法,但是,它需要非常精确的速度场信息。叠前时间偏移和叠后偏移是较快的方法,不过它们的假设前提通常有问题。我们需要一些规则来识别何种情况使用较简单的方法,以及确定各类偏移方法所需的速度模型。 依据澳大利亚西北大陆架Oliver测线的地质情况(具有60％的横向速度差),我们生成了一个合成地震数据集。时间、深度算法既用于速度模型平滑程度的评估,又适合于解释平滑程度的评估。 用平滑的速度做偏移可引起两种不希望有的效应,即同相轴纵向错位和振幅衰减。对于不同的平滑度,这两种效应可以作为横向位置的函数进行测量,以评价在Oliver模型上偏移的效果。 在叠前深度偏移中,当我们在200m的横向范围内平滑速度时,2800m处的反射层将产生70m的深度误差和60％的振幅衰减。而在叠前时间偏移中,同样的平滑将产生40ms的时间误差及50％的振幅衰减。在纵向上平滑时,叠前深度偏移存在45m的深度误差和40％的振幅衰减,叠前时间偏移有着类似的误差。
As with the complex geological formations in the northwestern continental shelf of Australia, conventional seismic processing can not correctly image and the lateral changes caused by the fractures make it impossible for the underlying tectonics to fully image, whereas prestack depth migration is a correct image for this configuration However, it requires very precise velocity field information. Prestack time migration and poststack migration are faster methods, but their assumptions are often problematic. We need some rules to identify when to use simpler methods, and to determine the speed models needed for various types of offset methods. Based on the geology of the Oliver line of the Northwest Shelf in Australia (with 60% lateral velocity difference), we generated a synthetic seismic data set. Time and depth algorithms are not only used to evaluate the smoothness of velocity model, but also to explain the degree of smoothness. Offsetting at a smooth speed can cause two undesirable effects, namely longitudinal misalignment of the in-phase axis and amplitude attenuation. For different degrees of smoothness, these two effects can be measured as a function of lateral position to evaluate the effect of the offset on the Oliver model. In the pre-stack depth migration, the reflective layer at 2800 m will have a depth error of 70 m and an amplitude attenuation of 60% when we smooth the velocity in a lateral range of 200 m. In prestack time migration, the same smoothness will result in a 40 ms time error and 50% amplitude attenuation. When longitudinally smooth, there is a 45m depth error and a 40% amplitude attenuation for the prestack depth migration, with similar errors in prestack time migration.