A differentiable programming framework tunes POD-Galerkin tensors using a hybrid trajectory-plus-energy-conservation loss to stabilize chaotic flow ROMs, achieving accuracy and stability with 20 modes on a Re=30,000 lid-driven cavity where classical methods need 80.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
physics.comp-ph 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
A Differentiable Programming Framework for Accurate and Stable Reduced-Order Modeling of Chaotic Flows
A differentiable programming framework tunes POD-Galerkin tensors using a hybrid trajectory-plus-energy-conservation loss to stabilize chaotic flow ROMs, achieving accuracy and stability with 20 modes on a Re=30,000 lid-driven cavity where classical methods need 80.