贝叶斯框架下联合AVA反演与统计岩石物理的储层参数估计

Bayesian petrophysical-properties estimation with AVA inversion and statistical rock physics

本文基于贝叶斯理论框架研究联合地震AVA反演和统计岩石物理的储层参数概率分布估计方法。首先通过井资料约束的贝叶斯线性AVA反演从角度域道集估计介质弹性参数的概率分布;然后结合适应目标储层的岩石物理模型,依据统计岩石物理关系从弹性参数进一步反演储层参数的概率分布。上述贝叶斯反演以后验概率分布的形式给出了弹性参数与储层参数的估计,以及由实际数据噪声、介质非均匀性、AVA正演方程与岩石物理方程近似引入的不确定性。通过理论模型数值试验展示本文方法的正确性,并应用于川东北地区碳酸盐岩地层的储层参数分布估计,证实了方法的应用潜力。

Based on Bayesian theory,we present a method combining AVA inversion with statistical rock physics to predict the probabilistic distribution ofpetrophysical-properties.First,we apply a Bayesian linear AVA inversion to estimate elastic parameters from angle-domain gathers constrained by well logging data.Then,combining with target reservoir rock physics model,we invert probabilistic distribution of reservoir parameters from elastic properties according to statistical rock physical relations.The mentioned Bayesian inversion for estimation of elastic and reservoir parameters is addressed by the posterior distribution,which describes various uncertainty related to noise in seismic and rock physical data,heterogeneity in media,approximations of AVA equation and physical models.The approach is tested on synthetic data to show its feasibility,then applied to carbonate reservoir petrophysical-properties estimation in northeast of Sichuan Basin.The example demonstrates the potential applicability of the proposed method.