利用4D地震数据校正储层静态模型的集合卡尔曼滤波方法

Reservoir static model updating by 4D seismic data using ensemble Kalman Filter method

储层静态模型是利用已知数据,结合先验性认识对未知储层空间进行插值预测,由此得到的结果与实际生产观测数据以及4D地震观测数据之间存在较大差异。基于集合卡尔曼滤波方法,通过观测数据反推系统模型的状态向量,对储层静态模型加以校正使得校正后的储层静态模型和观测数据之间差异最小化。选择4D地震属性差异作为观测数据,通过合理地抽取观测点,提高了运算效率。模型试验表明,校正后的静态模型能够较好地反映储层非均质性,并且与4D地震数据有较好的一致性。

The unknown reservoir properties are usually interpolated by using known data and priori knowledge. The results of the static model from this process could not agree well with production history and 4D seismic response. Taking the observation data as state vector, ensemble Kalman Filter （EnKF）is able to update the model and to minimize the mismatch of observed data and simulated data. Selecting 4D seismic attributes difference as observation data, the calculated efficiency is improved by choosing rational observation point. The updated models can represent heterogeneity of the reservoir and match the 4D seismic data well.