随机边界的波场重建技术研究

摘要

The anti-time migration technique on two-way wave equation is more suitable for complex structure imaging.However,large amount of computation and high memory requirement will make it unsuitable for large-volume seismic imaging.It's better to solve one of the storage problems through stochastic boundary conditions.Starting from the wave equation,this paper uses the commonly staggered-grid and high-order finite difference method to discretize the second order of time and the second order of space.Then,it obtains the stochastic boundary model.Finally,the wavefield is reconstructed with absorbing the boundary.We think that the wavefield reconstruction with the data in absorbing boundary grid is similar to the result of forward modeling.Under the condition of random boundary,we know that wavefield reconstruction can get better effect and prove its feasibility.%基于双程波动方程的逆时偏移技术更适合于复杂构造成像,但是,计算量大和所需存储量高使其无法适用于大数据量的地震成像.随机边界条件思想较好地解决了其中的存储问题.从波动方程出发,采用常用的交错网格高阶有限差分方法对其进行时间2阶、空间2N阶离散,进而得到随机边界模型.最后得到,在吸收边界的情况下波场重建,用吸收边界网格数以内的数据进行波场重建与正演模拟得到的结果相差不大.随机边界条件下,波场重建可以得到更好的效果,证明了其可行性.