Reservoir Characterization, Geological Modeling, and Reservoir Simulation of the N'Sano Field, Upper Pinda Reservoir

Abstract

Abstract N'Sano field is located offshore of the Angolan province of Cabinda in approximately 250 feet of water. The N'Sano field, discovered in October 1992 with Well 44-08X, is comprised of three non-communicating, oil-bearing horizons: the Upper Pinda reservoir, the Vermelha reservoir, and the Likuoala reservoir. The Upper Pinda reservoir of the N'Sano field was deposited in a flat, cyclical, near-shore marine environment. Consequently, the U. Pinda reservoir is characterized by a highly stratified sequence of sandstones, carbonates, and shales. The vertical correlation lengths of the various reservoir and non-reservoir facies are between 0.5 - 20.0 feet. Due to the stratified nature of the U. Pinda reservoir, a highly detailed reservoir characterization and geological modeling effort was conducted in support of a reservoir simulation study. To incorporate the complex reservoir geology, a facies based approach was used for the reservoir characterization and geological model construction. For the purpose of geological modeling, nine lithological facies were identified from U. Pinda core. All reservoir property relationships (log porosity estimators, porosity-permeability transforms, capillary J-functions) were derived on a facies basis. A 52x68x329 (1,340,144 cell) geocellular model of the U. Pinda reservoir was constructed for the geological study. Geostatistical techniques were used to populate the geocellular model. Vertical variograms were obtained from well logs. The log coverage in the U. Pinda reservoir in N'Sano was determined to be sufficient to provide good estimates of the vertical variograms. Areal variograms were obtained from a reservoir continuity study. Reservoir continuity, correlation lengths, and directionality of the nine lithological facies were obtained from this study and converted into areal variograms. The geocellular model was then upscaled to a 35x45x40 (67,200 total cell/43,010 active cell) reservoir simulation model. In order to preserve the stratified nature of the reservoir, the geological model was upscaled using the dynamic scale-up technique of Durlofsky et al. The simulation model was then history matched with two years of primary recovery and three years of secondary (waterflood) recovery. This paper will discuss in detail all phases of the study including: the reservoir characterization, geological modeling, model scale-up, history matching, and predictions. P. 317