A Novel Workflow to Model Grain Size and Capillary Pressure by Integrating Nuclear Magnetic Resonance Well Logs and Core Data: A Case Study for an Offshore Gulf of Mexico Oil Reservoir

Abstract

Abstract Subsurface properties derived from core testing are generally considered to be ground truth during the life cycle of field exploitation. The core acquisition is expensive, and analysis is time-consuming, and as a result, not many wells in a field are cored. Therefore, it is a common practice to integrate core and log data to develop a petrophysical model for uncored wells. But due to the complexities involved in integrating advanced multi-dimensional measurements, the value of the log data such as nuclear magnetic resonance (NMR) is not fully realized. In this paper, we will present a novel automated workflow to calibrate NMR well logs with MICP and Grain Size Distribution (GSD) data and apply calibration blindly to NMR logs in wells without core data to validate the models. First, NMR factor analysis and fluid substitution are performed on NMR well logs to 1) remove any hydrocarbon effect to simulate 100% water-bearing T2 distributions, 2) decompose T2 distributions into poro-factors representing underlying pore size distributions, and 3) determine NMR-specific textural facies over the depth intervals of wells with and without MICP and GSD data. Mean poro-factors from 100% water-bearing T2 distributions are then automatically calibrated against mean pore throat size distributions from MICP and mean grain size distributions in every facies for wells where such core data are available. Calibration models then can be easily applied across all wells with common textural facies derived using only NMR data. These workflows are applied to an oil reservoir in the Gulf of Mexico. This oil field is located in the southern Green Canyon protraction area of the Gulf of Mexico, in more than 4,500 feet of water. The main reservoir sandstones are early Miocene in age and were deposited by turbidity currents in a basin floor fan slope to an abyssal plain setting. These deposits overlaid autochthonous salt which eventually evacuated, leading to a structurally complex salt-cored anticline with four-way dip closure. The workflow, including calibration and application, provides consistent and continuous results across wells for both MICP and GSD using NMR T2 distributions. It includes MICP equivalent capillary pressures, pore-throat-size distributions, water saturation, Swanson permeability, entry pressure and irreducible pressure. Saturation height functions are derived by facies to incorporate into reservoir models. It also includes equivalent grain size distributions, Wentworth grain size volumes such as clay, silt, and sand, the distribution numbers (D10, D50, and D90), sorting coefficient, uniformity coefficient, and fines volume. An independent GSD model was also developed using routine core analysis (RCA) data to establish a relationship between porosity, permeability, and mean grain size. Both the NMR GSD model and the RCA GSD model demonstrated consistent results with the available core data in validation datasets. A proposed everyday workflow automates the NMR MICP and GSD core-log integration process to quickly propagate expensive core knowledge across all wells in a field. Calibration ensures that no single or variable scale factor is assumed across the range of T2, but rather facies-wise models are completely data-driven. Also, the workflow eliminates the need to accurately depth match core and log data. Calibration models can be stored, shared among experts, and can be applied post-logging or in real-time for quick decision-making related to reservoir engineering, completions, and production.