Resolving Productivity Mismatched with an Integrated NMR Pore to Production Solution

Abstract

Abstract Accurate estimation of reservoir properties such as shale volume, porosity, water saturation, and permeability is vital for reserve estimation and producibility. However, reliance on model-based computations alone can introduce uncertainties if the measured data or core results are biased. This paper addresses the case of Field A in Sabah, Malaysia, where the actual production rate falls significantly below the expectations of the model-based analysis. The observed permeability derived from production modeling based on well-test analysis is considerably lower when compared to log-derived and core-measured permeability profiles. To mitigate these uncertainties and enhance production well modeling, we present a workflow that optimizes the utilization of nuclear magnetic resonance (NMR) from pore to production analysis. The workflow involves deriving a pseudo capillary pressure to determine relative permeability, which serves as an input to reduce production well modeling uncertainty. We discuss four key workflows in this paper. The first workflow focuses on removing the hydrocarbon effect on NMR T2 distribution to reveal the actual rock properties. Subsequently, NMR Factor Analysis, an unsupervised machine learning technique, is applied to the water-saturated T2 distribution to identify clusters representing unique pore body profiles for each cluster. The following workflow utilizes an automated calibration process that incorporates core-MICP data and machine learning-derived clusters as inputs to generate a continuous capillary pressure-saturation height relationship from NMR and pore throat size distribution index for relative permeability calculations. This workflow provides Production Engineers with an alternative to use NMR-derived relative permeability as an input to calculate the KH window (effective permeability multiplied by thickness) and develop a well-calibrated model for zones without pressure build-up tests. The implementation of this workflow enhances the credibility of the model and reduces uncertainty for production optimization.