A neural network trained on rheometry data serves as a viscosity closure in a Cahn-Hilliard-Navier-Stokes finite-element solver via ONNX export, validated on literature benchmarks and experimental silicone-ink droplet-rise data.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
physics.flu-dyn 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Neural-Network-based Viscosity Closure for Non-Newtonian Multiphase Flows
A neural network trained on rheometry data serves as a viscosity closure in a Cahn-Hilliard-Navier-Stokes finite-element solver via ONNX export, validated on literature benchmarks and experimental silicone-ink droplet-rise data.