ABSTRACT
The main petrophysical challenges in carbonate reservoirs are often to define meaningful rock types, then to establish robust permeability and saturation models for these rock types, as well as to develop a realistic estimation of irreducible water saturation (Swirr). Realistic Swirr estimation is important for predicting production behavior (expected development of water cut) and thus ultimately for planning the future development scheme of a discovery.
In this study, we present the 2014 Alta discovery, located in the south-western Barents Sea. More than 50% of the expected hydrocarbon resources reside within complex carbonate reservoirs of Permo-Carboniferous age that display highly variable rock properties. Initial screening revealed that primary rock textures and pore geometries were for a large part overprinted by diagenetic processes. Hence, a better control on the reservoir's diagenetic evolution will be needed to apply a full-scale rock typing workflow. In the meantime, it was decided to proceed with a simplified reservoir characterization approach based on the main stratigraphic building blocks. Sufficient core coverage allowed for using permeability measurements on core samples as direct input to a 3D reservoir model. A customized core analysis program, using whole core samples, was designed to characterize the effect of large-scale vuggy pores.
For modelling water saturation, a workflow based on Thomeer hyperbola was developed that describes Mercury Injection Capillary Pressure (MICP) curves. The results adequately specify the saturation in all the stratigraphic building blocks. However, saturation uncertainty in the reservoir is high due to a highly variable cementation factor (m), unknown wettability and the presence of residual oil below the current Free Water Level (FWL). The Alta structure has been, and still is, leaking gas, causing the FWL to rise over time. To address the otherwise underestimated volumes in the transition zone above the current FWL, a deeper pseudo-FWL was created and used as input to the saturation height function.
Despite log-based water saturation (Archie) and core measurements (Dean Stark) indicating more than 80% water saturation for less permeable reservoir rocks within the oil leg, production tests did not produce water at normal rates. This clearly demonstrated the need to distinguish "nonproductive" pore systems (with capillary-bound fluids; in this case water) from pore systems contributing to production ("free" fluids).
A large MICP dataset confirmed that most reservoir rocks exhibit a mix of different pore types and pore throat diameters. To model this accurately, porosity partitioning in non-productive micro-porosity and free-porosity using the NMR logs was performed. Calibrating appropriate T2-cutoffs by matching core MICP to NMR logs in these heterogeneous rocks is seriously hampered by the large difference in sample size. Applying both MICP and NMR measurements to a subset of core plugs helped resolving this challenge. Comparing the corresponding free porosity to total porosity revealed near-linear relationships for different reservoir rocks.
For irreducible water saturation (Swirr), Swbound is calculated using the NMR-based free porosity. Swbound is considered to be a close approximation of Swirr. The resulting full-field simulation showed a significantly improved match between model output and recorded well test data.
