Novel Static Rock Typing Approach for Static Modeling of a Highly Heterogenous Reservoir from a Giant Carbonate Oil Field, Offshore Abu Dhabi

Abstract

Abstract Sediments in inner ramp environments are susceptible to sea-level changes, which often increase the heterogeneity of the original texture of the carbonate rocks. Additionally, inner ramp settings can encompass supratidal settings, which may include the distribution of evaporitic dolomites, further increasing the complexity of the reservoir. Furthermore, after deposition, carbonate rocks undergo diagenesis, which further complicates the characteristics of the reservoir. Decomposing these heterogeneities and complex geological elements to construct accurate rock types is crucial for creating precise geological models. The study was undertaken with the intention of developing a model that was adaptable and robust, capable of seamlessly transitioning between static and dynamic models to promptly correct any issues encountered in the dynamic model. With this motivation, we first developed rock types by integrating geological and petrophysical interpretations. These rock types were then utilized to construct a model where porosity, permeability, water saturation (Sw), saturation number (SATNUM), and relative permeability were interconnected in a cohesive chain. The first step included detailed descriptions of cores and thin sections from thirty-nine cored wells to define the lithofacies distribution within the reservoir. Subsequently, interpretations of the depositional environment were made based on these lithofacies distributions. Candidate Rock Types (CRTs) were then defined by linking lithofacies with diagenetic processes. Concurrently, Petrophysical Groups (PGs) were identified from porosity-permeability (Phi-K) and Mercury Injection Capillary Pressure (MICP) data using machine learning. A challenge encountered in the PG process was the existence of different MICP groups within the same Phi-K area due to the existence of limestone and dolomite in this reservoir. Therefore, by revisiting the geological interpretation and pre-segregating limestone and dolomite, we successfully categorized these distinct MICP groups within PGs. Following this, Static Rock Types (SRTs) were generated by reconciling the trends and clusters of CRTs and PGs, which were then predicted in un-cored wells using defined boundaries in the Phi-K domain. In the 3D reservoir model, SRTs were modeled based on the probability maps and Vertical Proportion Curves (VPCs) for each reservoir zone. The final SRTs were designed and optimized to achieve reasonable predictability in uncored wells, which was confirmed by blind tests and found to be useful in controlling the distribution of SRTs within the reservoir model. The spatial distribution of SRTs was also verified by comparing it with depositional environment trends and diagenesis trends. SRTs were then employed to constrain the petrophysical properties and saturation modeling. Accurately capturing all the heterogeneities in the reservoir is crucial for constructing a reliable 3D reservoir model that honors the reservoir’s flow behavior. The strong integration of geology and petrophysics from the initial steps enabled the successful execution of this SRT workflow, leading to the construction of a robust 3D reservoir model that captures both the geological understanding and flow behavior of the reservoir.