【摘要】 笔者提出了基于Metropolis抽样的弹性阻抗随机反演方法,该方法是一种基于蒙特卡洛的非线性反演方法,能够有效地融合测井资料中的高频信息,提高反演结果的分辨率。应用贝叶斯理论框架,首先通过快速傅里叶滑动平均模拟算法(fast Fourier transform-moving average,FFT-MA)和逐渐变形算法(gradual deformation method,GDM)得到基于地质统计学的先验信息,然后用Metropolis抽样算法对后验概率密度进行抽样,得到反演问题的解。其中FFT-MA模拟作为一种高效的频率域模拟方法,结合GDM后,在保持模拟空间结构不变的前提下,可以连续修改储层模型,直至满足实际观测地震记录。数值试验表明:FFT-MA模拟有效地提高了计算效率,融入GDM更新算法,可以保证反演结果有效地收敛,并且反演结果与理论模型吻合较好,具有较高的分辨率,该方法采用两步法反演弹性参数,运算速度较快。更多还原

【Abstract】 Stochastic inversion of elastic impedance based on Metropolis sampling algorithm is a Monte Carlo based strategy for nonlinear inversion,which can effectively integrate the high-frequency information of well-logging data and have a higher resolution. This method is formulated in the Bayesian framework. Firstly,the priori information can be obtained through Fast Fourier Transform-Moving Average( FFT-MA) and Gradual Deformation Method( GDM). Then Metropolis algorithm is employed so as to obtain an exhaustive characterization of the posteriori probability density.FFT-MA is a kind of efficient simulation method. Combined with GDM,it can constantly modify the reservoir model and keep the spatial structure unchanged until it matches the observed seismic data.According to the numerical calculations,it can be concluded that FFT-MA simulation can reduce the time consumption.Combined with GDM updating algorithm,the inversion results can converge rapidly,and the final results match the model well and have a higher resolution.In addition,this method adopts two-step method to invert elastic parameters,so it improves the computational efficiency to some extent.更多还原