基于统计学的随机地震反演在储层预测中的应用——以蜀南某地区嘉陵江组储层预测为例

APPLICATION OF STATISTICS-BASED SEISMIC STOCHASTIC INVERSION TO RESERVOIR PREDICTION——IN CASE OF RESERVOIR PREDICTION OF JIALINGJIANG FORMATION IN SOUTH SICHUAN

在钻井资料较少的勘探开发初期,现今流行的地震反演方法难以准确预测储层。如递推反演分辨率不高;基于模型反演和随机模拟地震反演技术依赖大量井资料,在井资料较少的情况下,反演结果可靠性不高等。基于统计学的概率密度函数,随机地震反演引用测井资料参与反演约束,能有效地提高反演结果分辨率,并能充分考虑地下地质的随机特性,使反演结果更符合实际地质情况。同时,通过对测井资料概率密度函数的求取来实现反演,解决了随机模拟地震反演的变差函数在井资料较少,甚至在工区内井分布不均匀的情况下,变得极不稳定而使反演结果可靠性不高的问题,能较为准确地预测出储层低速异常带的分布。该方法在蜀南嘉陵江组储层预测中的成功应用表明,在薄储层且勘探初期钻井资料较少,且现今流行的叠后反演方法不能准确预测储层的情况下该方法能取得很好的应用效果。

The popular methods of seismic inversion are difficult to accurately predict reservoir with little drilling data during Early exploration and development.For example,the resolution of recursive inversion is low,also the results of model-based inversion and random simulation inversion are low reliable.Well data is used to constrain the process of inversion to increase the inversion resolution in statistics-based seismic stochastic inversion.In the same time,the random characters of subsurface geology are fully considered,which made the inversion results more consistent with practical geological scenario.The application of probability density function of well data can help to solve the problem of low reliability of variation function when the data is inadequate and uneven.In addition,the method also precisely predicts the distribution of low-velocity anomaly area.The satisfying result in reservoir prediction of Jialingjiang formation indicates that this method can get a good effect when popular post-stack inversion can not precisely predict thin reservoir without enough drilling information during the early exploration.