The Nemba Field is located approximately 62 km west of the coast of Cabinda, Angola. The reservoir for the field is the Lower Cretaceous (Albian) Pinda Formation. Field OOIP is estimated to be 819 MMB, of which approximately 317 MMB is expected to be ultimately recovered.

The Pinda Formation comprises a thick succession of highly heterogeneous mixed carbonate and elastic sediments that were deposited in environments ranging from shallow marine shelf to non-marine coastal plain. The dominant depositional environment was that of a low sediment supply barrier island system with multiple, small sources of fluvial input, extensive lagoons, open bays, and tidal fiats in the coastal plain. Wave-dominated, sand-rich shoreface/barrier island systems developed in regions in close proximity to clastic input sources. Nearshore carbonate shoal complexes developed in shoreline regions that were more distal from the sources of clastic input. The lagoonal and coastal plain sediments can also be either siliciclastic or carbonate in composition.

The reservoir is characterized by a thick succession of vertically stacked flow units which vary in thickness dramatically throughout the field due to differences in accommodation space during deposition. Impermeable layers between the flow units also vary in thickness, or, in many cases, pinch out stratigraphically. This results in flow units that are segregated vertically in one area of the field being stacked and in vertical pressure communication in other areas of the field. Variations in the three-dimensional geometries of sediments deposited in open marine shoreface, carbonate shoal, tidal channel/inlet, and bay-head deltaic systems also present potential modeling problems.

Accurate geostatistical modeling and reservoir simulation required an innovative integration of stratigraphy, depositional facies analysis, and lithofacies-based petrophysics into the model. The “Gocad” software was used for geostatistical modeling. Total model size was 9 million cells with areal grid dimensions of 100 by 100 metres, and maximum vertical layer thickness two feet. The reservoir was divided into intervals defined by field-wide flooding surfaces or exposure surfaces in order to group together flow units deposited in similar depositional environments. The correlated stratigraphic layers were mapped field wide using 3-D seismic data. The sediments within each interval were divided by composition into lithofacies groups, and each lithofacies group was assigned several possible variogram lengths (correlation lengths). The division into composition-based lithofacies allowed for mineral-based petrophysical analysis of the data set. Lithofacies models were constructed using “Categorical Sequential Indicator Simulation.” This model was used to define regions for each rock group which were filled with petrophysical properties consistent with the individual lithofacies, stratigraphic layer, and depth in the reservoir. Multivariate statistical analysis of the petrophysical and core data sets allowed for the construction of synthetic permeability, porosity and water saturation logs for uncored wells. Log permeability was used as the primary modeling parameter and populated in each region by layer using “Guassian Sequential Simulation.” Porosity was computed from log permeability using “Correlated Probability Field Cloud Transforms.” Separate scattergrams were constructed for each rock group in each stratigraphic interval. Water saturation was modeled from evaluated wire-line data using defined relationships between log permeability, porosity and depth above oil/water contact.

For reservoir simulation the geostatistical model was scaled up by a factor of 100 using dynamic flow simulation which preserves fractional flow behavior. The effectiveness of the scaleup was confirmed by a comparison of flow performance between the fine and coarse grid models in several cross sections. Following scaleup the fault planes and scaled properties were loaded into ECL Grid and fault transmiscibilities computed prior to input of all data into “Chevron's Enhanced Application Reservoir Simulation” software.

AAPG Search and Discovery Article #90937©1998 AAPG Annual Convention and Exhibition, Salt Lake City, Utah