Rospo Mare is a fractured karstic heavy oil carbonate reservoir. Reservoir pressure has been constant during the entire production, supported by a strong hydrodynamic aquifer.
Most of producing wells are horizontal and completed in open hole. They produce at critical rate to prevent water production, due to the lack of water treatment facilities.
This paper focuses on a new history match approach, considering the complexities given by the described conditions.
A traditional history match approach meets criticalities due to the lack of matching parameters, since the lack of water production, pressure differences and negligible gas production, due to a very low GOR.
A reservoir modeling with history match was already developed in 2015 and an upgrade was necessary.
To carry on the update, correct the former history match and obtain a predictive model, a reverse engineering approach was applied running the simulations in forecast mode. The time between the end of the former history and the present time was simulated in such manner.
Analysis of wellbore storage from well tests revealed values much higher than the theoretic ones computed considering wellbore geometry without fractures. This is explained considering the karst and the fracture system around the wells connected directly to the wellbore.
Because of this a decrease in well productivity is observed not only as water arrives in the wellbore but also in the fracture connected, even with no water production at the surface. To model this phenomenon, especially when water coning occurs, the area around the well drain was considered during simulation, and a factor reducing well productivity was introduced and applied as water saturation exceeding a threshold. This factor has been calibrated well by well as history match parameter.
The threshold applied to water saturation around the wellbore succeeded to represent this behavior. Production data were represented, and the history match was achieved.
After two years of production, results from the model were compared with effective production data. The model was found again consistent with the data, validating the approach.
Using the forecast mode for reverse engineering successfully allowed to validate the predictiveness of the model, in absence of other constraining data.
The obtained model introduces a new way of considering and modeling wellbore and production behavior in fracture carbonates with karst. It allows a good representation of the physical behavior of the field and provides a useful tool for piloting the field and assisting future decisions.
