Hydrodynamic simulation-derived training images to improve braided-river reservoir architecture characterization

Multiple-point geostatistics (MPS) can reproduce complex geological patterns, but its performance depends on representative training images (TIs). We developed a workflow that converts forward hydrodynamic simulations of braided-river deposition into three-dimensional TIs for reservoir-architecture modeling. Five candidate TIs were reconstructed from time-resolved erosion–deposition increments and classified into channel, mid-channel-bar, floodplain, and bar-margin elements. They were evaluated using cross-sectional architecture, fractal dimension, vertical transition probability, and vertical facies thickness, with interpreted well sequences providing quantitative constraints. The selected TI was used for MPS modeling and compared with sequential indicator simulation (SIS) and object-based modeling (OBM) across ten realizations per method. MPS preserved TI-consistent channel–bar geometries and target proportions, whereas SIS produced fragmented architectures and OBM generated excessive vertical amalgamation. Maximum connected mid-channel-bar thickness was 17.45–20.64 m for MPS, compared with 9.89–22.87 m for SIS and 31.18–50.02 m for OBM. At two held-out wells excluded from conditioning and TI selection, ensemble modal architectural-element accuracy was 76% and 78% for MPS, exceeding SIS (70% and 63%) and OBM (63% and 65%). The workflow links physically based TI construction, quantitative TI selection, and independently tested reservoir prediction, providing a reproducible approach for characterizing interwell architecture while avoiding unsupported inference about dynamic-flow performance in braided-river reservoirs.

Authors

Institutions

Publication Details

Journal
Scientific Reports
Published
2026-11-19
DOI
https://doi.org/10.1038/s41598-026-80635-6
Primary Topic
Geological Modeling and Analysis
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Hydrodynamic simulation-derived training images to improve braided-river reservoir architecture characterization

Lihua Ren, Xiangbo Zhang, jiang chao, Chengyan Lin et al.
Scientific Reports
Geological Modeling and Analysis
article

Hydrodynamic simulation-derived training images to improve braided-river reservoir architecture characterization

Lihua Ren, Xiangbo Zhang, jiang chao, Chengyan Lin, Suheng Li, Huafeng Liu
article en

Abstract

Multiple-point geostatistics (MPS) can reproduce complex geological patterns, but its performance depends on representative training images (TIs). We developed a workflow that converts forward hydrodynamic simulations of braided-river deposition into three-dimensional TIs for reservoir-architecture modeling. Five candidate TIs were reconstructed from time-resolved erosion–deposition increments and classified into channel, mid-channel-bar, floodplain, and bar-margin elements. They were evaluated using cross-sectional architecture, fractal dimension, vertical transition probability, and vertical facies thickness, with interpreted well sequences providing quantitative constraints. The selected TI was used for MPS modeling and compared with sequential indicator simulation (SIS) and object-based modeling (OBM) across ten realizations per method. MPS preserved TI-consistent channel–bar geometries and target proportions, whereas SIS produced fragmented architectures and OBM generated excessive vertical amalgamation. Maximum connected mid-channel-bar thickness was 17.45–20.64 m for MPS, compared with 9.89–22.87 m for SIS and 31.18–50.02 m for OBM. At two held-out wells excluded from conditioning and TI selection, ensemble modal architectural-element accuracy was 76% and 78% for MPS, exceeding SIS (70% and 63%) and OBM (63% and 65%). The workflow links physically based TI construction, quantitative TI selection, and independently tested reservoir prediction, providing a reproducible approach for characterizing interwell architecture while avoiding unsupported inference about dynamic-flow performance in braided-river reservoirs.

Scientific Reports
China University of Petroleum, East China (CN)
Geological Modeling and Analysis
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon