Oil reservoirs developed under a natural depletion scheme usually leave behind a significant portion of the discovered oil and while in-fill drilling is a logical choice to recovering more oil from these reservoirs, there is always the challenge of locating the remaining oil and optimally placing new wells within these ‘sweet spots’.
3-Dimensional (3-D) dynamic modelling and 4-D seismic are well developed methods of establishing post-production saturation distributions and a combination of both methods further increase the confidence of the post-production saturation distributions predicted. Even then however, these methods are not infallible. Experimental design methods when integrated into dynamic modelling enable a logical and systematic interpretation/utilisation of all available data to understand the extent and impact of uncertainties, an understanding which can help make robust decisions about these re-developments.
The reservoir of interest (Level 1X) is an oil rim reservoir with a 36-year production history from 6 wells. All the existing completions in the reservoir were used to develop the oil-rim. To evaluate the option of developing the gas cap of the reservoir, a technical review of the oil rim had to be carried out to assess the existing developments and evaluate the possibility of further development opportunities in the oil rim prior to the gas cap blow down.
Over the area of this field, 4-D seismic was not available and while the traditional dynamic reservoir modelling method could be used to plan further development, the oil rim nature of the reservoir coupled with reservoir uncertainties (such as oil-water and gas-oil contacts) meant a second method would give further credence to any analyses of the reservoir and the proposals resulting from such analyses. This paper demonstrates how experimental design, incorporated into dynamic simulation, has been applied to the reservoir study to achieve the following:
Integration of practically all data into the study (not a pre-selected value for each uncertainty)
Allowing available data to guide the history match of the reservoir (History matches evolve/drop out of data) thus increasing confidence in the calibrated (low, mid and high case) models.
Allowing a selection of a high confidence case re-development strategy.