Replace Human Intelligence with Fast and Smart Geometric Reasoning and Graph Neural Network to Accelerate Next Gen ModSim Workflows

Quadros, William Roshan, Winovich, Nickolas, Ernst, Corey Devan et al. · 2025

Read Replace Human Intelligence with Fast and Smart Geometric Reasoning and Graph Neural Network to Accelerate Next Gen ModSim Workflows by Quadros, William Roshan, Winovich, Nickolas, Ernst, Corey Devan et al. (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/doe-osti-2997308-replace-human-intelligence-with-fast-and.txt.

Description

Sandia National Laboratories report SAND2025-11995R covers a laboratory project by William Quadros, Nick Winovich, Corey Ernst, and Nya Hardy on teaching software to carve up engineering models. Before a part can be simulated, experts split its computer model into blocks that mesh cleanly with brick-shaped elements. That judgment step is slow, variable, and hard to scale for high-stakes simulations.

The pipeline shrinks a 3D model to a 2D skeleton, then to a graph of surface and curve nodes. Each node is labeled with the cutting action an expert would take in the Cubit tool. A graph neural network trained on those labels predicts five surface actions and three curve actions on new shapes it never saw. Training used an 80/20 split and 2,000 passes over the data. The aim is expert speed without losing expert judgment.

Catalog record

CREATORQuadros, William Roshan, Winovich, Nickolas, Ernst, Corey Devan et al.
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE43 KB
LICENSEPublic Domain
RIGHTSPUBLIC DOMAIN
SOURCEDOE OSTI

Source: https://www.osti.gov/biblio/2997308

Browse the full library · About the Tower of Babel