摘要
基于叠前地震数据振幅随入射角变化特征的叠前反演方法是获得储层弹性参数估计的重要手段。常规贝叶斯叠前反演假设弹性参数反射系数满足某一特定分布,利用该分布作为先验信息约束反演,且分布一经确定不再随不同地震道变化,但在弹性参数反射系数的分布横向变化较大或不满足既定分布时会产生误差。为此,通过引入自动相关判别先验,提出了一种贝叶斯叠前同时反演方法,将弹性参数反射系数的先验信息融合到反演中,该先验信息不再对弹性参数反射系数的总体分布作出假设,且随着地震数据的变化自适应改变,使弹性参数反射系数的反演结果具有更好的准确性、更符合横向变化的地质特征,并且更好地反映岩性边界;在构建反演目标函数时,加入了趋势约束算子,提高了反演的稳定性;采用同时反演策略,避免了基于弹性阻抗的叠前反演过程中存在转换误差。单道弹性参数模型和经典推覆体2D模型测试结果证实了方法的可行性和正确性,实际地震数据测试结果证实了方法在提高分辨率和反映地质体特征及边界上的优势。
The pre-stack inversion method based on the amplitude versus angle(AVA)of pre-stack seismic data is an important method to estimate the elastic parameter of subsurface reservoirs.The conventional Bayesian pre-stack inversion assumes that the reflection coefficient of the elastic parameter satisfies a certain distribution,and uses this distribution as prior information to constrain the inversion.Once the distribution is determined,it will be used consistently with different seismic traces,and errors will occur when the lateral changes in the distribution occur or the distribution does not meet the given distribution.To address this issue,a Bayesian pre-stack simultaneous inversion method was proposed by introducing a priori automatic relevance determination(ARD).Using a priori ARD,the reflection coefficients can refer to independent parameter values,and no assumptions are made about the overall distribution of the reflection coefficient of the elastic parameter.The priori ARD can change adaptively with different seismic traces,so that the inversion results can conform to the lateral variation of the geological characteristics.To improve the stability of the inversion,a trend constraint operator was adopted when constructing the inversion objective function.A simultaneous inversion was adopted to avoid the conversion error in the pre-stack inversion based on the elastic impedance.Tests on both synthetic and field data verified the effectiveness of the method in depicting geological structures and lithologic boundaries.
作者
纪永祯
朱立华
林正良
胡华锋
JI Yongzhen;ZHU Lihua;LIN Zhengliang;HU Huafeng(Sinopec Geophysical Research Institute,Nanjing 211103,China)
出处
《石油物探》
EI
CSCD
北大核心
2020年第4期572-582,共11页
Geophysical Prospecting For Petroleum
基金
中国博士后科学基金资助项目(2019M662005)
国家科技重大专项“不同缝洞储集体地震识别与预测技术”(2016ZX05014-001)
国家科技重大专项示范工程“缝洞系统结构刻画及描述技术完善”(2016ZX05053-001)共同资助。
关键词
稀疏约束
自动相关判别先验
叠前同时反演
趋势约束
块状边界
自适应先验
贝叶斯反演
sparse constraint
automatic relevance determination(ARD)
prestack simultaneous inversion
trend constraint
layer boundary
adaptive priori
Bayesian inversion