因具有较高的分辨率,VSP资料常用于揭示储层及钻井附近的复杂构造。本文提出一种弹性迭代偏移法。复杂的地质构造对它几乎没什么制约性。在每次迭代过程中,本方法自动对P波和S波的宏观速度模型进行调整和改进。它避开了波场分离(上/下行波及P/S波),并且简化了大多数方法中小倾角地层的假设。 本偏移法是建立于弹性反演理论基础之上。波场外推法的基础是用高阶、稀网格有限差分法去解弹性双程波动方程。在每次迭代中,用梯度法对宏观模型进行调整,其中梯度通过正演模拟场与反向传播的剩余场相关计算而得。偏移法的首次迭代相当于弹性逆时偏移,它的成像条件类似于Claerbout原理。P波和S波反射都能用于成像。 数值实例(合成的和实际的有偏VSP资料)都表明,增加迭代次数可提高图像质量。与单一用声波能量所成的图像相比,用P波和S波能量的成像具有更丰富的井信息和更高的空间分辨率。 在实例中,可以看到用此方法能对相对较复杂的地质构造进行成像 其中与井相交的一条断层和一个小地堑能清晰地辩认出。 Because of its high resolution, VSP data are often used to reveal complex structures in the reservoir and near the wellbore. This paper presents a flexible iterative migration method. Complex geologic formations have little or no restraint on it. During each iteration, this method automatically adjusts and improves the macro-velocity model of P-wave and S-wave. It avoids wavefield separation (up / down-going P / S waves) and simplifies the assumption of low dip formations in most methods. The offset method is based on the theory of elastic inversion. Wavefield extrapolation method is based on the high order, thin grid finite difference method to solve the elastic two-way wave equation. In each iteration, the macro model is adjusted using the gradient method, where the gradient is calculated by correlating the forward modeling field with the backpropagation residual field. The first iteration of the offset method is equivalent to elastic inverse time migration, and its imaging conditions are similar to the Claerbout principle. Both P-wave and S-wave reflections can be used for imaging. Numerical examples (both synthetic and actual biased VSP data) show that increasing the number of iterations improves the image quality. Imaging with P-wave and S-wave energies has richer well information and higher spatial resolution than a single image with acoustic energy. In this example, one can see that relatively complex geological formations can be imaged using this method, where a fault intersecting a well and a small graben can be clearly identified.