A Carbonate Case Study Investigation Using a Petrophysical Rock Type Workflow Using Deterministic to Machine Learning and Introducing Multi-Well Capillary Saturation Height Modeling.

Abstract

Abstract The objective of this paper is to propose a practical approach to implement Petrophysical Rock Typing (PRT) Techniques and validate them through Advanced Flow Unit Analysis, Multi-Well Saturation Height Modeling and Machine Learning. This customized workflow is targeting Carbonates Reservoirs and show casing a multi-well synthetic dataset for a carbonate reservoir from the Arabian Gulf area. The objectives are to provide a consistent approach to identify the initial Petrophysical Rock Typing (PRT) interpretation based on geological pore geometry analysis, multi-well-Advanced Flow Unit Analysis and refine PRT results integrating traditional deterministic techniques. PRT results are validated using statistical methods and Multi-Well Saturation Height Modeling and compared to Machine Learning. The key steps include: – Geological, MICP, core analysis and pore throat calibration. – Verify PRT with Deterministic Depth Domain Techniques. – Verify and model PRT using probalistic supervised analysis (IPSOM and HRA). – Predict PRT with Multi-Well Non-Normalized Modified Lorenz. – Validate PRT using Saturation Height Modeling. – Validate PRT using Multi-Well Saturation Height Modeling – Compare PRT results with Machine Learning. – Introduce Multi-Well Saturation Height Modeling The proposed workflow combines MICP geological rock analysis, Advanced Flow Unit Analysis along with deterministic and probalistic PRT methods. Verification and validation processes provide a method to evaluate the uncertainty and consistency of the results. This case study is based on a carbonate reservoir located in the Arabian Gulf region. This approach targeted a key well dataset (4-6 wells) that includes a poor well, a moderate well and a good performing well, which are located at different structural positions across the structure (Crest, Mid-range, and Flank). The implemented workflow provides the key inputs to validated Petrophysical Rock Typing interpretation for carbonate reservoirs. Geologic rock analysis, MICP and pore geometry analysis are completed as soon as the data is available. This information provides pore throat size calibration and pore geometry PRT profiles. PRT is predicted using a Multi-Well Composite Diagnostic Non-Normalized Modified Lorenz Cross refined and modified by integrating traditional Deterministic Pore Throat indicator techniques. This enables, a consistent integration of several techniques to characterize the petrophysical rock types from a Geological, Petrophysical and Reservoir behavior standpoints. Furthermore, multi-well Saturation Height Modeling feeds the workflow with robust validation methods taking into consideration the impact of the transition zone. Additional verification includes statistical based (probalistic) methods and machine learning techniques. This practical case study provides guidance on how and when to use the various PRT methods, so people can make an informed selection. An important step in the process is comparing PRT results from multiple methods. Machine learning was used to predict PRT as a validation process and Multi-Well PRT based capillary saturation height modeling is introduced in this paper. The proposed workflow combines MICP geological rock analysis, Advanced Flow Unit Analysis along with deterministic and probalistic PRT methods. Verification and validation processes provide a method to evaluate the uncertainty and consistency of the results. This case study is based on a carbonate reservoir located in the Arabian Gulf region.