Joint History Matching of Well Data and Surface Subsidence Observations Using the Ensemble Kalman Filter: A Field Study
Joint History Matching of Well Data and Surface Subsidence Observations Using the Ensemble Kalman Filter: A Field Study
复制标题
使用集成卡尔曼滤波器对井数据和地表沉降观测进行联合历史匹配：现场研究
DOI：
10.2118/141690-ms
复制
发表时间：
2011
期刊：
影响因子：
--
通讯作者：
P. V. Hooff
中科院分区：
文献类型：
--
作者：
F. Wilschut;E. Peters;K. Visser;P. Fokker;P. V. Hooff
由于各种原因，集成卡尔曼滤波（EnKF）用于油藏模型历史匹配的报道数量正在稳步增加。在这里，我们报告利用EnKF的能力来同时处理来自不同来源的观测。虽然传统上只匹配井数据，但我们将地面沉降观测数据与井数据结合使用。地表沉降是由于储层岩石通过地下的力学响应而被压实造成的。压实作用是由于储层衰竭过程中孔隙压力降低造成的。因此，沉降数据包含了储层动压分布的信息。将井资料与地面沉降观测的联合历史拟合应用于Roswinkel气田。该油田已经运行了25年，在此期间，进行了9次平整作业，产生了一组有价值的沉降数据。由于背斜构造中可能存在大量封闭性断裂，储层的分区不确定是该储层的一个重要特征。因此，在本研究中，断层可传递性的估计取代了岩石性质的不确定性。在先前对Roswinkel的研究中，通过反演沉降测量来估计压实场。结果表明，储层中存在多条封闭断层，将油田划分为具有独立压力历史的不同隔室。然而，反演后的生产数据历史匹配并不令人满意。我们现在已经能够证明，在油藏断层可传递性的情况下，可以用合成情况下的生产和地面沉降数据以一致的方式来估计驱动参数。此外，对于实际的Roswinkel案例，联合历史拟合清楚地揭示了两个来源的数据与当前油藏模拟模型之间的差异。作为一种灵活的历史匹配方法，EnKF成功地将地表运动数据与油井生产数据联合进行历史匹配。但更重要的是，它证明了利用互补信息源来改进储层表征的潜力，以及估计驱动参数的可能性。石油工程师学会版权所有。
The number of reported applications of the Ensemble Kalman Filter (EnKF) for history matching reservoir models is increasing steadily for various reasons. Here, we report on exploiting the capability of EnKF to handle observations from different sources simultaneously. While traditionally only well data are matched, we use surface subsidence observations together with well data. Surface subsidence results from compaction of the reservoir rock through the mechanical response of the subsurface. Compaction is caused by decreasing pore pressures during reservoir depletion. Therefore, the subsidence data contains information about dynamic pressure distributions in the reservoir. The joint history matching of well data and surface subsidence observations was applied to the Roswinkel gas field. This field has been operated for 25 years, during which nine leveling campaigns generated a valuable data set of subsidence data. An important feature of the reservoir was the uncertainty about its compartmentalization, due to a large number of possibly sealing faults in the anticlinal structure. Therefore, instead of uncertain rock properties, fault transmissibilities were estimated in this study. In a previous study on Roswinkel, a compaction field was estimated by inverting subsidence measurements. The results indicated several sealing faults in the reservoir, dividing the field into different compartments with independent pressure histories. The post-inversion history match of production data, however, was unsatisfactory. We have now been able to show that estimating the driving parameters, in casu the fault transmissibilities in the reservoir, can be achieved in a consistent way with both production and surface subsidence data for a synthetic case. Furthermore, for the actual Roswinkel case, the joint history match clearly reveals the discreapancy between the data from both sources and the current reservoir simulation model. The joint history match of land surface movement data together with well production data is a success for EnKF as a flexible method for history matching. But more importantly, it demonstrates the potential of using complementary sources of information for improved reservoir characterization, and the possibility of estimating the driving parameters. Copyright 2011, Society of Petroleum Engineers.