In a giant carbonate reservoir, located offshore Abu Dhabi, permeability and production data are hard to reconciliate. Recent investigations on this reservoir have shown that it is affected by natural fractures. Borehole images, cores and thin-sections allowed characterizing two main fracture sets. However, the scarcity of data, with regards to the size of the reservoir, did not allow building a discrete fracture network. To overcome this gap, modelling fractures in an implicit way appears the best and simplest solution to enhance permeability and match the production data.
Implicit means that fractures are not considered as a discrete object. Instead, at the cell scale, fracture impact on permeability has been translated into multipliers in the x, y, and z directions, superimposed on the matrix permeability. The process of implicit fracture modelling is not straightforward. Building a fracture conceptual model (seismic faults, fault-related fractures, subseismic faults, and background fractures) is the starting point, to understand the big picture. The second step is defining the layer/fracture relationships, and quantifying the interaction between matrix and fractures (transmissibility). The third step is modelling and implementing the permeability modifiers in the 3D geological grid.
The distribution and occurrence of the different features remains difficult to constrain using one single deterministic model. Using multi-realization is powerful tool to get the closest match to PTA.
Deterministic scenarios have been created combining joints and faults, assigning different properties to each of the feature: size, frequency, degree of baffling, degree of permeability anisotropy, vertical transmissibility, mechanical stratigraphy, position to the structure of the reservoir…The permeability field resulting from each scenario has then be compared to PTA to determine which case best matches the dynamic data. The created permeability array is then added to the matrix permeability of the reservoir, enhancing it up to 6 times.
Implicitly modelling fractures through multi-realizations enable to get a solution that match as much as possible the PTA. In the particular case of large reservoir with scarce data, it allows geologists to have an estimate of the fracturing state of the reservoir and how these fractures are contributing to the overall field flow behavior. It is also a way of mimicking the fracture pattern in the case of few data availability, and which ultimately will enhance the quality of the history match