二维弹性多波时空域高斯束偏移方法

随着我国勘探开发难度逐步增大，勘探目标开始向裂缝油气藏、岩性油气藏等复杂探区转移，研究高精度、适应性强的多波多分量深度偏移算法在后续的地震解释、属性分析及储层预测中具有重要意义.针对多波多分量地震数据，本文提出了一种二维弹性波时空域高斯束偏移方法.时空域高斯束沿中心射线传播时能够面向成像目标描述局部波场，且对振幅和频率可调制的Gabor基函数有天然的适应性，因而将基于Gabor分解的子波重构方法应用于震源波场构建，从而得到任意点源函数产生的时空域高斯束波场.该方法由于直接在时间域进行计算，可以避开频率域中出现的假频和边缘截断效应等问题.基于各向同性弹性波动方程的Kirchhoff-Helmholtz积分解，利用矢量时空域高斯束传播算子构建格林函数和格林位移张量，并结合上行射线追踪策略，实现了检波点波场的反向延拓.针对矢量波成像问题，本文借鉴弹性波逆时偏移方法从矢量延拓波场中分离出纯纵波分量和纯横波分量，进而采用修改后的内积成像条件产生具有明确物理意义的PP、PS成像结果，避免了转换波成像的极性反转问题.最后利用简单两层模型和不含盐体构造的部分Sigsbee2a模型的成像结果，并将其与应用近似纵横波成像条件、标量和矢量势成像条件的偏移剖面进行对比，验证了本文方法的正确性和有效性.

Elastic space-time Gaussian beam method for seismic depth imaging

As the difficulty of exploration and development in China gradually increases the exploration target begins to transfer to complex exploration areas such as fractured reservoirs and lithologic reservoirs. And it is of vital significance for subsequent seismic interpretation, seismic attribute analysis, and reservoir prediction to develop a multi-wave, multi-component depth migration algorithm which is highly accurate and strongly robust. In order to process common-shot multi-wave and multi-component seismic data, we develop a 2D elastic imaging method using vector space-time Gaussian beam propagation operators. The space-time Gaussian beam can describe the local wave-field of target-oriented areas over the whole central ray path and naturally adapt to the Gabor basis function which is modulated in terms of amplitude and frequency. Therefore, the source wavelet refactored method based on Gabor decomposition is applied to forward wavefield construction and then we can obtain a space-time Gaussian beam wave-field generated by an arbitrary point source function. This method can avoid problems such as aliasing and edge truncation effects occurring in the frequency domain due to its direct calculation in the time domain. Adhering to the framework of the acoustic space-time Gaussian beam method, we perform the up-going ray tracing from subsurface imaging points to the receiver surface for the construction of the backward wavefield. The P-and S-wave-fields are simultaneously extrapolated based on the Kirchhoff-Helmholtz integral solution of the isotropic elastodynamic equation. The Green's functions and Green's displacement tensors are represented with vector space-time Gaussian beam summation. We separate the extrapolated wavefields into compressional and shear modes using the Helmholtz decomposition, and then implement a modified dot-product imaging. This approach permits to produce clear PP images and avoid polarity reversal problems for PS images. Numerical experiments on a simple two-layer model and a partial Sigsbee2a model without salt structure have verified the accuracy and validity of the proposed method.