Abstract
Abstract and Overview
The objective of this extended abstract is to highlight how integrated studies can be enhanced and reduce uncertainty by applying these new methods including multi-well saturation height functions, advanced petrophysical rock typing methods, advanced core-log integration, machine learning & neural network algorithms. We include a generic reservoir case study showing the improved workflow.
A major challenge in most studies is how to propagate fluid saturation in a 3D model. Using the new Multi-Well Saturation Height Function method allows for improved cored-log integration of Petrophysical Rock Types (PRT) using Open Hole logs. To tackle this issue, an integrated workflow has been deployed which included the following steps (Gunter et.al. 2018, 2020-2022).
Lithofacies, pore geometry, pore types, core data and fluid contacts must be understood. Core-log integration to calibrate core and rock properties from deterministic to probalistic methods should be used and results compared. Petrophysical Rock Types definition and validation at the cored wells are extended to the non-core wells using Statistical/Probabilistic methods. A multi-well Saturation Height Modeling (SHM) is implemented based on Petrophysical Rock Types and matching fluid saturations from well logs. This multi-well approach allows a better estimation of Free Water level, improved understanding of reservoir compartmentalization and reduced uncertainty over traditional single well saturation height models.
The case study shows how this new workflow and application provides an efficient distribution of fluid saturations based on capillary theory, fluid contacts, petrophysical rock types, and pore geometry. Successful 3D models must have an excellent geological representation of the reservoir system including thin section information, mineral composition calibration, pore geometry, capillary properties, and flow capacity of reservoir units. Core log integration must be validated to identify cutoffs and then define petrophysical rock types and permeability equations.
The above enables, from statistical standpoint, successful selection of the inputs logs that would discriminates between the defined rock types through a Linear Discriminant Analysis, which is a critical point in the performance rock types propagation in the non-cored intervals/ wells.
Utilization of multi-well saturation height model improves results, the 3D distribution of the fluids and identifies fluid contacts with less uncertainty, than single well based methods.