This paper discusses about a systematic approach of probabilistic dynamic modeling which was used to assess subsurface uncertainties, and to forecast range of ultimate recovery (P90, P50, and P10) for infill and EOR opportunity (immiscible Water-Alternating-Gas, iWAG) in a mature offshore oilfield.
The field is very complex, with multi-layered and highly compartmentalized reservoirs and has been on production under natural depletion since the last 47 years. It is in "mid-life crisis", facing problems of high watercut and GOR, with bypassed oil areas and pressure depletion, and therefore in need for redevelopment to increase its recovery factor. However, despite the long production history, subsurface uncertainties remain high due to its reservoirs complexity and limited reservoir data.
To assess the subsurface uncertainties and their impact on field redevelopment plan, a fine-scaled sector model was built for a focused area of the field with substantial remaining opportunity. Multiple static realizations were used to represent different reservoir connectivity that can be highly influential on the ultimate recovery, especially for iWAG injection. Along with these multi-static realizations, key dynamic uncertainties such as fault transmissibility, kv/kh, aquifer properties, etc. were also incorporated. It is very important to note that uncertainty parameters may vary between history matching and forecast period, especially when there is a change in drive mechanisms. Therefore, additional uncertainty parameters such as residual oil saturation (Sorw, Sorg), in the case for iWAG scenario, have also been included from the very beginning to ensure seamless transition from history matching up to forecasting.
Another big challenge to conduct probabilistic dynamic modeling in a mature field is, to generate multiple equally probable history matched realizations for subsequent proxy modeling and probabilistic forecasting. The uncertainty and optimization (U&O) modeling tools, including experimental designs, have made this possible and proven to be the most efficient way to speed up the modeling process. It has helped to generate the physical P90, P50, and P10 dynamic models that could be used to assess the robustness of various development options, hence able to deliver the optimum strategies while managing subsurface uncertainties.
All these proved to be a comprehensive probabilistic workflow to firm up several sizeable infill and injection portfolios in this complex mature field, with potential to increase recovery factor by 8% to 13%. Most importantly, decision makers were given a range of outcomes instead of a deterministic forecast, allowing proper risks mitigation plans to be devised to deliver a robust project definition and superior performance.