基于贝叶斯分类的储层物性参数联合反演方法

Joint inversion of petrophysical parameters based on Bayesian classification

应用地球物理资料进行储层物性参数反演是储层预测及综合评价的重要步骤。基于贝叶斯分类的储层物性参数联合反演方法综合应用统计岩石物理模型和蒙特卡罗仿真模拟技术,在贝叶斯反演框架下,基于贝叶斯分类算法计算储层物性参数后验概率分布,实现多种储层物性参数的联合反演。该方法不需要进行复杂的模型初始化,而是通过统计岩石物理模型建立储层物性参数与岩石弹性参数之间的关系,进而与叠前地震反演相结合,不仅能模拟地球物理随机特性,还能解决常规物性参数反演方法对测井资料过度依赖的问题。海上某区块实际资料应用结果表明,该方法能为储层精细描述提供多种物性参数,并可对反演结果的误差进行定量评价。

Petrophysical parameters inversion based on geophysical data is an important step in reservoir prediction and comprehensive evaluation.We proposed a new petrophysical parameter inversion method based on Bayesian classification.This method involves multiple theories and techniques,such as statistical rock physical model,Monte Carlo simulation technology.The posterior probability distribution of petrophysical parameters is calculated based on Bayesian classification algorithm and can realize multiple petrophysical parameters joint inversion based on Bayesian inversion framework.This method does not need to carry out complex model initialization.It obtains the relationship between petrophysical parameters and elastic parameters through statistical rock physics model,and then combined with pre-stack seismic inversion.This method can not only simulate geophysical stochastic properties,but also can well solve the problem of excessive dependence on log data in conventional petrophysical parameters inversion methods.The practice shows that it can provide various petrophysical parameters for fine reservoir description and quantitatively evaluate the errors of inversion result.