In sandstone reservoirs, one of the most important challenges is the effective upscaling of a fine scale model. A conventional upscaling process is not adequate when significant percentage of shale distribution exists in the reservoir. In the fine scale model, significant discontinuity exists in sand bodies. Some sand bodies are connected to the wells, and some are not. As the fine scale model is upscaled, some discontinuous sand bodies are combined with other connected sands. This results in two potential problems: the connected volume to the existing wells increases, thus making the production from those wells more optimistic, and the production from in fill wells do not show as much additional potential since some of the new volumes which should have been connected to the new well are already drained by the existing wells.
A new procedure1 is developed to overcome this problem. In a new procedure, we first determine the connected sand volume to the existing wells. We remove the dis-connected sand bodies from the fine scale model. We then upscale the model to a desired level. We simulate the flow performance till a desired time when either new wells are drilled or some well are shut-in. At that point, we start from the fine scale model again and determine new connected volume due to additional of new wells. We combine the new virgin sands with depleted sands in the upscaled model and determine the saturation and pressure in the upscaled model using an appropriate material balance technique. We re-start the simulation using newly connected volume till we reach a point of drilling additional wells. The key difference between the proposed method and the existing methods, is our ability to add new hydrocarbon volumes (as well as new conductivity) in the model as a function of time.
The proposed method was applied to a giant oil field in Siberia which is in turbite environment with large amounts of discontinuous sand bodies. We were able to demonstrate the advantage of the proposed method by comparing the performance of upscaled simulation model to the fine scale geologic model. As the percentage of sand decreases in a given reservoir, the difference between the conventional and a proposed method becomes significant. Using the new approach, we would be able to evaluate the infill potential much more accurately.
The traditional approach of reservoir modeling involves developing a fine scale geo-cellular model which incorporates the small scale, static, uncertainties. Once such description is created, it is upscaled to appropriate scale so that it can be flow simulated to understand the dynamic performance of the reservoir. It is difficult to flow simulate a fine scale model since computationally it can be demanding.
In recent studies, investigators have tried to model sandstone reservoirs using this approach. Some sandstone reservoirs tend to be highly discontinuous with sand and shale dispersed within the reservoir. The continuity of the sand bodies may not be precisely known and reflects one of the uncertainties in the static models. The problem becomes even more complex if the reservoir is several meters thick and contains sand bodies which are thin. When fine scale model is constructed, it may contain large number of vertical layers to account for thin sands present in the reservoir.
Upscaling of such models is very difficult because of the following reasons:
If the model is coarsened (upscaled) significantly, many of the sands which are discontinuous, will connect to other sands, and as a result, this will
increase the "connected" pore volume to the well bore,
change the transmissibility (connectivity) distribution in-between wells.
As a result, there will be discrepancies between fine scale recoveries and coarse scale model recoveries of hydrocarbons.