Integrated reservoir & production modelling can be extremely beneficial in conducting production forecasting for complex systems consisting of multiple reservoirs, fluid mixing, and complex production networks. These forecasts are even more valuable in scenarios such as sour gas separation, compression, and re-injection into multiple oil and gas reservoirs. The challenges in performing such work include modelling of the reservoir and fluid complexities while correctly representing the reservoir and production network, and optimizing the entire system together. Such tasks have been a daunting proposition for the petroleum industry because of the technical and logistical shortcomings in the traditional approach of integrated modelling where several tools are coupled together explicitly.
In this study, a new, fully-coupled implicit tool was used to model an onshore middle-Eastern asset with multiple reservoirs, each with unique fluid, and multiple networks. The asset uses complex fluid mixing in the integrated network, multi-stage separation followed by gas sweetening, compression and re-injection of compressed sour gas into the reservoirs. A multi-fidelity approach was used throughout the modelling workflow, incorporating reservoir and production related uncertainty in the forecast and optimization process. For example, some reservoirs were modelled with refined numerical discretized models, some with upscaled numerical discretized models and some with decline curves. A total of 16 reservoirs linked to a single production system were modelled as an integrated system. Unique compositional fluid models were used for each discretized reservoir, as the reservoirs had distinct fluid phase behavior during primary production and under miscible enhanced oil recovery.
Consequently, an integrated system was built that contained numerical discretized and decline curve reservoirs, fluid complexities, high fidelity wells, and a production network with gas separation, sweetening and re-injection. The oil production and recycled gas injection were optimized using rules and long-term forecasts were generated.
Throughout the work, several production and reservoir engineers collaborated and shared data on the single multi-user, multi-disciplinary tool using a relational database – which solved the logistical challenges associated with modelling such large-scale projects where the use of traditional methods can often lead to communication breakdowns and data discontinuities.