Treffeisen, Torben ; Henk, Andreas (2021)
AI-Based Comparison and Upscaling of Detailed to Homogenized Fault Zone Representations in Reservoir-Scale Hydro-Mechanical Simulations.
The 55th U.S. Rock Mechanics/Geomechanics Symposium. Houston, Texas, USA (18.06.2021-25.06.2021)
Konferenzveröffentlichung, Bibliographie

Kurzbeschreibung (Abstract)

Coupled hydro-mechanical simulations are by now a standard tool for reservoir modeling. However, challenges remain regarding the proper implementation of faults into such reservoir-scale numerical models which result mainly from the scale difference between the element size of the numerical grid and the internal heterogeneity of real fault zones. In this study, we present an AI-based upscaling approach from detailed fault zones to homogenized fault properties in finite element models. The detailed fault zone model contains a fault core with shear bands and host rock lenses as well as accompanying damage zones with a fracture density decreasing with distance from the fault core. This detailed fault zone model is automatically compared by using the Structural Similarity Index and difference images with a database of 432 models describing the fault zone as one uniform material homogenizing the different fault zone units. The goal is to introduce a new upscaling workflow of hydro-mechanical fault zone properties in reservoir-scale simulations through a database driven approach to represent the fault in a numerical simulation with an optimal grid size while maintaining the bulk effect of a detailed fault zone description on fluid flow, stress and deformation and thus, improving the predictive power of reservoir-scale simulations.

Frage zum Eintrag

Redaktionelle Details anzeigen