Abstract
The reservoir in South Oman – discovered in 1989 – is a 400m thick, tight, micro-porous silicilyte slab encased in salt at a depth of 4000m. The field contains the unique Athel silicilyte formation, some 4 km below the surface and is fully encased in salt as a result of which the reservoir is over-pressured at 800 bars.
The crestal area has average matrix porosity of 22% and the permeability of the reservoir is extremely low (typically 1 – 100 μD). The oil in this field is very light, sour, low viscosity and volatile (48° API, Pb=260 bar, H2S=1.5mol%, CO2=2.5mol%). The application of massive hydraulic fracturing (2-5 fracs/per well) combined with the favorable oil properties and undersaturated nature of the oil has made economic primary depletion development possible in this low permeability reservoir with initial oil rates in the range of 100-600 m3/day. Current oil recovery factor is 4.2% and expected to reach 10% through continued primary depletion.
MGI via hydraulically fractured wells is seen as possible EOR method to improve field recovery. A pilot injection phase consisting of 2 inverted 5-spot patterns in the crest of the reservoir will start Q2-2015 followed by a 2nd phase of crestal MGI by 2026. The purpose of the current dynamic modeling is to assess the 2nd phase of crestal MGI.
The modeling challenges are complex reservoir characteristics (vertical and areal heterogeneity), tight reservoir nature, hydraulically fracced wells, sour content and MGI. The subsurface complexity requires a detailed static model (areally and vertically) combined with preferably a compositional dynamic model. Obviously, run time of a compositional dynamic model is increasing dramatically with increasing complexity which is limiting the practical applications.
Several solutions to optimize the running time and accuracy have been investigated (i.e. explicit or implicit modeling of fracs, rate or THP control, black oil versus EOS models, various up-scaling options, sector versus full field models), tested and calibrated with historical data. In addition, history matching of more than 14 years production history of 43 wells with 25 fracs/well was another significant challenge. Different approaches, e.g. single/multiple case deterministic (MultiRun) and Design of Experiments (DoE) approach have been investigated for both history match and forecasting.
This paper covers a fit for purpose modeling approach, using Shell’s in-house simulation tool MORES, with powerful functionalities that enable modeling of the subsurface complexity in order to make the field development decision on the MGI extension. In addition, successful application of uncertainty modeling and assessment by using a classic approach (i.e. deterministic models) and a stochastic approach (i.e. DoE using assisted HM methods) will also be covered in the paper.