Abstract
Gas reservoirs in the Nile Delta of Egypt are characterized vertically by its thin beds of sands and shale and laterally by severe variations in facies. These challenges in the static modeling have a strong impact in the dynamic modeling which can be summarized in the following points. First, the vertical sequence of sands and shale leads to the difficulty in detecting a single gas-water contact in the field. Second, the vertical heterogeneity leads to the use of fine gridding especially in the vertical direction to accurately simulate the fluid flow in the reservoir. Third, the lateral variation in facies forces to use different saturation functions at different parts of the reservoir.
The dynamic behavior of pressure and production performance from few wells (total seven wells) producing from this field show severe vertical discrepancy in pressure, gas and water production. This adds another challenge in the dynamic modeling and leads to dividing the field into three main reservoirs that are completely isolated with each has different reservoir and production characters.
Due to these challenges, we developed an unconventional approach to model this field and estimate its gas in place that honors both static and dynamic data. First, we used the concept of initialization by enumeration instead of the conventional approach of initialization by equilibration that requires accurate detection of gas-water contact. The initial pressures are obtained from MDT (Modular Dynamic Tester) data and the water saturation are obtained from the petrophysical analysis in initial wells and populated in the reservoir to honor the seismic and the production data. Second, we used low vertical permeability to simulate the vertical variations in sand-shale sequence. This also helps in reducing the severity of the phases being not in equilibrium due to the initialization by enumeration. Third, we used fine gridding to capture the heterogeneous variation in property. To reduce the CPU time for running one single fine grid model, we divided it into three different models each represent one reservoir with running each model separately. Both production and pressure data confirms that these reservoirs are completely isolated and no wells are producing commingled from these reservoirs thus separating them will not have any effect on its performance.
This work summarizes the workflow we developed in the dynamic modeling of this field, the history match approach we used to calibrate the model and finally the suggested optimum development plan to increase its reserves.
