Analysis and Modeling of Fracture-enhanced Production in Lacustrine Carbonate Reservoirs at Kambala Field, Cabinda Province, Angola, West Africa

Scheevel, J.—Scheevel Geo Technologies LLC
Domingos, F. and Nogueira, J.—Sonangol
Fernandes, M.; Skander, L.; Costa, L.; Lomando, A.; Kienast, V.—ChevronTexaco

        The Kambala Field, of offshore Cabinda, Angola produces oil and associated gas from the Lacustrine Toca carbonate facies of the Barremian and Aptian aged Bucomazi Formation. The productive intervals reside at depths more than 10000 feet (3000m) below mean sea level. The field is a structural-stratigraphic trap located on a basement-cored horst block below the extensive Late Aptian Loeme salt interval. 
        The Kambala Toca discovery well (Lower Toca) was drilled in 1971 with production commencing in 1974. The field has produced at various rates as high as 28 MBOPD. Subsequent drilling, material balance and two history-matched full field flow simulation studies have estimated reserves in the range of > 1.0 BBOOIP. Of this resource less than 50 MMBO have been produced, in over 30 years. This represents only about 5% recovery from the 15 field wells over approximately 1500 hectares (~3708 acres). 
        The reservoir setting is the product of the continental rifting during the early history of the Lower Congo Basin prior to complete opening of the south Atlantic. The Barremian and Aptian depositional systems consisted of interconnected lakes that were fringed by algal, oncolitic, and china marls carbonates or sand lacustrine shoreline. The carbonate shorelines tended to occur during periods of shallowing anoxic/euxinic lake stages, where earlier sediments were not fully emergent, whereas, the sandy lake margins dominated during periods of active tectonics and where continuous shorelines provided a connection to clastic source material. The Toca facies name designates the carbonate reservoirs of Bucomazi age, with the name Bucomazi generally reserved to describe the fine-grained very organic-rich pelagic lake deposits that are the lateral equivalent of the Toca. 
        The Toca carbonates at Kambala underwent a complex combination of early sub-aerial exposure, and partial to complete dolomitization. These alteration phases were accompanied by syndepostitional and post-lithification faulting and fracturing, often, associated with the invasion of various fluids including, anoxically-derived lacustrine pore fluids, basement-derived hydrothermal brines, and migrating hydrocarbon liquids and gas. The combined depositional, deformational and fluid-migration history has conspired to create two stacked carbonate sequences (Upper and Lower Toca) each ranging from 75-300+ feet (20-100 meters) thick possessing highly heterogeneous reservoir fabric at all scales of observation. The result is a productive Toca sequence that has highly variable reservoir behavior from well to well. Consequently the field has historically been, at the same time, both rewarding and frustrating to exploit. 
        The development history of the Kambala Toca has been very sporadic. Much of the explanation for this history can be attributed to the highly variable outcomes of the infill and delineation wells over the years. Initial productions (IP’s) have ranged from a few hundred BOPD to as high as 15,000 BOPD on flowing tests. Still other development wells, in the interior of the known productive area have been plugged and abandoned because of failure to produce on test. Estimated ultimate and demonstrated cumulative recovery (per well) are also quite variable from a few 10’s of thousands BO to more than 20 MMBO. History and experience have not been without gain, however. For example higher production rates in recent wells have been achieved through experimentation with massive acid fracturing, and the use 3D seismic data and empirical amplitude evaluation and more recently, advanced methods for creating detailed high resolution 3D seismic products have also led to greater well targeting success. Nevertheless, predictable well production rates still remain elusive. 
        The historical perspective of the elements of reservoir risk and reward at Kambala resulted in the proposal to drill a horizontal well in the Toca in late 2002. This was a bold step in that the proposed well was to be the deepest horizontal well in the Cabinda concession and would also be landed and completed below the regional Loeme salt layer and this would be a first. Because of extended lateral reach within the reservoir, the well was expected to penetrate the entire spectrum of heterogeneity within a single wellbore as opposed to multiple near-vertical penetrations. The well was also designed to target primary observed fracture trends at high angle to the wellbore in order to maximize the fracture system contribution to well production. Dynamic modeling considerations suggested that the well could produce at target rates with lower drawdown and thereby minimize the possibility of drawing water to the wellbore via localized fracture networks. 
        The proposed horizontal well was successfully drilled and completed in late 2002. This well achieved significant drilling milestones in that it is the deepest horizontal well in the Cabinda concession, and was successfully completed below the thick Leome salt layer. More importantly, the well was a development success and is producing very close to the pre-drill predictions. Historical stabilized IP’s for all wells in the field average 4560 BOPD and have a median value of 2650 BOPD. The current stabilized production of the horizontal well is averaging 3500 BOPD. Future horizontal well opportunities are being evaluated based on this success.
        The long production history and historical data gathering on the Kambala Toca field has left us with considerable background for making reliable observations that will aid further predictive modeling of the Toca reservoir. Cores, standard wireline logs, Nuclear Magnetic Resonance logs, image logs, wireline pressure testing (SFT’s), flowmeter production logging and 3D seismic all contribute to our current knowledge of the reservoir properties and architecture. Flow simulation studies with the benefit of the long production history have been insightful in understanding the pressure and fluid saturation history of the field but have historically, and continue to be unreliable in predicting individual well performances prior to drilling. It is expected that the highly heterogeneous reservoir conditions and localization of features relative to the simulation model scale, limit the flow-simulation model’s ability to predict accurately at specific locations, but still allow gross field behaviors to be modeled. 
        The Block 0 working interest partners’ lengthy operating experience combined with high quality data have led to improvements in understanding of the Toca at Kambala. Experience in structural geology, seismic interpretation and processing, petrophysical and depositional systems in the Cabinda area (Block 0 concession), the Democratic Republic of the Congo (to the south) and Republic of the Congo (to the north) add to our understanding the rock- and fluid-history of Kambala field.
        Although many questions remain to be answered, our historical observations and studies continue to be synthesized and the following conclusions can now be made about the Kambala Toca reservoirs:

1. Heterogeneity of the matrix reservoir rocks is high, while the maximum matrix porosities remain in the 2-10% range, generally following the original depositional system trends. 
2. Matrix permeablities associated with these low porosities are too low to support observed flow rates in some most wells. This fact is further quantified using finite difference flow simulations, which require large (up 50X) matrix perm multipliers to match well performance. Large adjustments such as this are indicative of the presence of a secondary permeability system. (Fracture and/or vug network).
3. Fracture observations from image logs and cores indicate that fracture density and total fracture count, from well can be highly variable. There exist both open and mineral-filled fractures, with heavy precipitation of baroque dolomite indicative of episodes of hydrothermal brine invasion. Open fracture orientations are dominantly NNE-SSW at a high angle to the NW-SE horst bounding normal faults that were episodically active during deposition and burial, but other fracture orientations also exist. 
4. North to NNE oriented transverse faulting, both regionally and within the Kambala field is associated with the contemporaneous NW-SE oriented rift-extension. These transverse faults fit well into a system of left-lateral, simple shear whose shear couple is oriented N-S. Such a system, acting throughout the history of the basin would be capable of producing both NW-SE normal faults and potentially the complexity to produce the NNE-SSW open fracture system. The presence of deep transverse faults in the basement may provide a likely conduit for the hydrothermal brines to permeate the reservoir.
5. Vertical pressure communication (observed by wireline pressure testing) appears to be enhanced within each of the two Toca reservoirs. This implies that fracture systems are assisting vertical communication between layered matrix porosity zones. Also, lateral compartmentalization, observed in both in pressure and saturation history of the field, requires that observed NW-SE normal faults must serve as barriers or baffles to flow rather than conductors of flow. This observation stands in contrast to the role of open fractures most of which are at high angles to these normal faults.
6. High resolution and seismic characterization techniques have been used to clarify stratigraphic, structural and reservoir property configurations within the field. New technology in the form of windowed principal component analysis (PCA) has served to enhance the vertical resolution and predictive character of the seismic data and help give insight into additional drilling opportunities by allowing reservoir lithology prediction at the reservoir flow-unit scale.
7. Highly fractured flank wells have resulted in the highest historical IP rates but also tend to entrain reservoir water immediately, resulting in short lives and sacrificing oil in place. Horizontal wells may prove to be the most reliable method for minimizing the risk of heterogeneity, maximizing the intersection with the fracture network and minimizing the near-bore pressure drawdown necessary to drill and complete economic wells in the Toca reliably.

Copyright © 2004. The American Association of Petroleum Geologists. All Rights Reserved.