Tensor network reduced order models for wall-bounded flows
Martin Kiffner, Dieter Jaksch
Abstract
We introduce a quantum inspired and widely applicable tensor network-based framework for developing reduced order models describing wall-bounded fluid flows. For the examples of the lid-driven and doubly-driven cavities, we find that our method only uses a small fraction of the number of variables parameterizing the solution compared to direct numerical simulation and can improve its runtime by an order of magnitude on comparable hardware. Our work provides a novel path towards efficient high-precision simulations of the Navier-Stokes equation at high Reynolds numbers.
Topics & Concepts
Bounded functionTensor (intrinsic definition)Path (computing)Reynolds numberComputer scienceWork (physics)Order (exchange)Fraction (chemistry)Applied mathematicsMathematical optimizationTopology (electrical circuits)Computational scienceMathematicsPhysicsMathematical analysisMechanicsGeometryTurbulenceProgramming languageThermodynamicsChemistryFinanceCombinatoricsEconomicsOrganic chemistryModel Reduction and Neural NetworksLattice Boltzmann Simulation StudiesTensor decomposition and applications