基于贝叶斯理论的逐次迭代非线性AVA反演方法

Non-Linear Pre-Stack Seismic AVA Inversion Based on Bayesian Theory Using Successive Iteration Method

常规AVA三参数反演方法均基于横波速度与纵波速度之比γ为常数这一假设条件,且常被近似地取为0.5。然而在许多情况下γ并不为常数,而是在横向与纵向都渐变。若一概假定γ等于0.5,反演出的岩性参数势必要偏离真实值,因此有必要合理地选择γ。笔者基于贝叶斯理论,提出逐次迭代非线性AVA的反演方法。该方法把γ看成横向与纵向都渐变的反演初始背景,通过给定初始模型计算初始背景γ,并采用逐次迭代的策略求解该反演问题,解决了关于γ的选取问题以及由于引入变γ值而带来的非线性问题,提高了AVA三参数反演结果的精确度。

Conventional three-term AVA inversion methods are based on the assumption thatγ （the ratio of S-wave velocity to P-wave velocity）is a constant value usually considered to be 0.5,whileγ is horizontally and vertically varied gradually in many cases.The estimated parameters of the inversion is bound to deviate from its true values with γ invariably being 0.5.and the selection γ needs to be reasonably.Based on Bayesian theory,we presents a nonlinear pre-stack seismic AVA inversion using successive iterative method,which considered the ratio’s initial background varying horizontally and vertically and being calculated by the-given initial model,and the nonlinear inversion problem was solved by successive iteration.The proposed method gived a reasonable solution for the selection of γ and solved the nonlinear problem caused by variable ratioγ.And the accuracy and stability of the three-term AVA inversion were improved.