Generating saturation functions (capillary pressure or J-functions), to initialize complex carbonate reservoir, always presents challenge to Petrophysicists and reservoir engineers, especially with no SCAL data available or any existing rock properties trends such as porosity-permeability relationship which are used to assign saturation function in 3D models. The proposed method requires initial water saturation (Swi) distribution in 3D model in hand (first stage) and then the Swi distribution recalculated more accurately (second stage) using group of capillary pressure (Pc) curves based on Swi intervals. Calculating Swi distribution in two stages should not impose any limitations since in the first stage Swi distribution can be estimated by many ways such as J-functions (4), group of Pc curves based on porosity intervals (part of this work). Otherwise, it may be estimated by any girding software which uses Swi log-data, but this method is not recommended since the software would give erroneous estimation of Swi in the areas of no well control such as in the transition zone area.
The proposed method calculates the Swi distribution in two stages, first the Swi calculated by Pc curves based on porosity rages, then in the second stage the Swi recalculated using group of Pcs based on water saturation ranges.
A script file was written to differentiate between these regions and assign saturation numbers (SATNUMs) for each saturation region which used in the simulator to initialize the dynamic model. Shuaiba reservoir is presented as study case here to demonstrate the capability of the proposed method. The proposed procedure can be to initialize huge complex carbonate reservoir such as Shuaiba formation in the United Arab Emirates.
The initial water saturation profile from log data matched the water saturation calculated by the dynamic model in 90% of the wells (more than 100 wells at initial water saturation).
The proposed procedure eliminates the tedious efforts to find rock property trends such as porosity-permeability correlations, which in many carbonate reservoirs may not exit, in order to assign a saturation function for each porosity/permeability range(s).
The difference in initial oil in place (STOOIP) calculations between the static model (40 million cells) and the dynamic model (2.7 million cells) is less than 2 percent. It can be used also in many heterogeneous reservoirs to reduce the uncertainty in STOOIP and thus reserves estimations.