基于BISQ模型的储层物性参数贝叶斯反演方法

Bayesian inversion based on BISQ model for reservoir physical properties

对于物性参数反演这类存在多解性的问题,常见的方法往往难以高效、高精度地获得最优解。因此,根据贝叶斯统计学,采用高斯先验模型引入测井数据约束,并发展储层物性参数贝叶斯反演方法,以克服反演的多解性。其次,在BISQ理论和等效流体理论的基础上建立了岩石物理正演模型,该模型反映了孔隙介质中固体和流体的多种物理作用,可精细地刻画储层中地震波的传播过程,进而提升反演结果的可靠性。此外,由于确定性优化算法通常只有局部收敛能力,无法收敛到反演问题的全局最优解,所以引入具有全局收敛能力的杂交遗传算法,以提高反演方法的精度和效率。最终,利用所提反演方法预测中国东部A油田目标区域的孔隙度和含水饱和度,获得了较好效果。

For problems with multiple solutions such as the physical property inversion,common methods are usually difficult to obtain optimal solutions with high efficiency and accuracy.Therefore,according to the Bayesian statistics,we use the Gaussian prior model to introduce the constraints from well‑log data and develop a Bayesian inversion method for reservoir physical properties to overcome the multi‑solution problem.Next,a petrophysical forward model is developed based on the BISQ and effective fluid theories,which can reflect the multiple physical effects of solids and fluid in the porous medium and finely characterize the propagation process of seismic waves in the reservoir,thus enhancing the reliability of inversion results.In addition,because deterministic optimization algorithms usually have only local convergence capability and fail to converge to the global optimal solution of the inversion problem,hybrid genetic algorithms with global convergence capability are introduced to improve the accuracy and efficiency of the inversion method.Finally,the proposed inversion method is used to predict the porosity and water saturation in oil field A in eastern China,and positive results are obtained.