Water Saturation Prediction in 3D Models: Paradigm Change From Free Water Levels to Interfacial Normalized Free Water Levels

In estimating hydrocarbon reserves, saturation prediction plays a key role. To predict hydrocarbon saturation away from the wells Saturation Height Models (SHM) are prepared and deployed, with the goal of taking into consideration rock quality variability in a lateral sense. One of the key inputs in the SHM is the Free Water Level(s). When dealing with only one hydrocarbon type the FWLs determination is a straightforward procedure following directly from the driving fluid gradients. On the other hand, when dealing with an oil rim/gas cap using the FWLs as derivFed from gradients can lead to predicted water saturation peculiarities. By using a FGWL (Free Gas water Level defined as zero capillary pressure between gas and water) next to the FOWL (zero capillary pressure between oil and water) one can observe a discontinuity in the predicted water saturation at the top of the oil rim (at the Gas Oil Contact - GOC). This effect can have a significant effect on field- wide hydrocarbon volumes. Looking for data supporting this predicted peculiarity at the top of the oil rim has not led to positive results, meaning the conventional modeling approach might be incorrect. A parallel workflow has been developed proving that using FWLs might lead to erroneous results in the case of an oil rim. The workflow revolves around calculating the capillary pressure without the default use of the FGWL. The conventional workflows revolve around building a SHM under the Pc_IFT1 approach (IFT1 is a normalization process eliminating involved fluids properties variability- like oil-water/mercury -air , etc), and applying it to the derived FWLs. This leads to errors identified as related to the normalized capillary pressure input. The magnitude of the errors is capillary pressure and rock properties dependent but can underestimate the hydrocarbon saturation at the bottom of the gas cap by as much as 25%. In this paper we demonstrate the correct approach relies on the use of a normalized to IFT1 Free Water Levels (ex: FGWL_IFT1) instead of "just" Free water Levels. This property is consistent with the fundamental assumption used in deriving multi-fluid SHM. As a result, for example, no discontinuity in Pc_IFT1 at the top of the oil rim is present, a satisfying physical solution as well as addressing a modelling shortcoming away from the well. The proposed solution can be easily applied in the 3D models, where the hydrocarbon volumes are estimated. The implementation is just a straightforward correction to the FGWL depending on the fluid properties.