Year: 2025
Type: article
Abstract: Process‐based geological modeling replicates the physical laws governing depositional and diagenetic processes, generating geologically realistic models. However, these models sometimes fail to align with well log, seismic, and other spatial data. Conditioning process–based geological models to such data is essential, but has posed challenges for decades. Traditional conditioning methods involve trial‐and‐error or automated inverse procedures to fine‐tune input parameters. While these methods can enhance model calibration, they often fall short of achieving fully satisfactory results. This paper introduces a novel approach that integrates a multiscale regionalized (MR) multiple‐point statistics (MPSs) method with process‐based geological modeling to improve well data conditioning. The proposed workflow, termed MPS‐MR, leverages statistical inference of geological patterns derived from replicated events within localized regions of a training image (TI). During stochastic simulation, each unsampled location is associated with a localized region of the TI. As these regions are location‐dependent, the MPS‐MR method overcomes the stationarity requirement of the TI, a limitation of many existing MPS techniques. The MPS‐MR captures and reproduces geological patterns and trends across multiple scales. Its integration with process‐based modeling offers an effective solution for well data conditioning, while preserving geological patterns and trends. This is demonstrated through a case study involving high‐sinuosity fluvial channel modeling.
Source: Journal of GeoEnergy
DOI: 10.1155/jge5/1988687
Authors: Yupeng Li, Lin Ying Hu
Institution: Saudi Aramco (United States)
Language: English
Subfield: Geochemistry and Petrology
Field: Earth and Planetary Sciences
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid:
Funder: Saudi Aramco
OpenAlex ID: W4412464311