Fast Supremizer Method on Penalty-Based Reduced-Order Modeling for Incompressible Flows
Authors:
- 17 December 2025
- journal article
- Published by Society for Industrial & Applied Mathematics (SIAM) in SIAM Journal on Numerical Analysis
- Vol. 63 (6) , 2483-2511
Abstract
The supremizer method enriches the reduced velocity basis for pressure recovery in incompressible flows, ensuring the inf-sup condition in the reduced space. In the full-order model, a small penalty term is often introduced to prevent spurious modes [Y. He, Math. Comp., 74 (2005), pp. 1201–1216] and is also essential for accuracy in the proper orthogonal decomposition–based reduced-order model [ A.-L. Gerner and K. Veroy, Math. Models Methods Appl. Sci., 21 (2011), pp. 2103–2134 ]. However, coupling pressure and velocity, along with the supremizer basis, significantly increases the computational costs in both offline and online phases. We find that the primary role of supremizers is to improve stability, rather than velocity accuracy. We propose a novel method using several supremizers for the velocity basis, decoupling the penalized system to solve for velocity. The full set of supremizers is then used to recover pressure. This strategy reduces the computational cost while maintaining stability and accuracy. We derive error estimates using a supremizer-augmented projection operator, which depend on the inf-sup constant rather than on the inverse of the penalty coefficient. We also develop two new supremizer construction options satisfying the inf-sup condition, one of which avoids solving the full-order equations for obtaining supremizer basis, further reducing offline costs. Numerical experiments demonstrate the effectiveness of the proposed method. For comparable accuracy, CPU time tests show that the online computational cost is reduced by about [Formula: see text], and the offline assembly cost by [Formula: see text], compared to [Y. He, Math. Comp., 74 (2005), pp. 1201–1216].Keywords
Funding Information
- Agence Nationale de la Recherche (ANR-22-CE46-0005)
- Nouvelle Aquitaine Region
This publication has 26 references indexed in Scilit:
- A POD-Based Reduced Order Method Applied to Goda Time-Splitting SchemeJournal of Scientific Computing, 2025
- A phase field method for convective phase change problem preserving maximum bound principleApplied Numerical Mathematics, 2024
- An efficient time-splitting approximation of the Navier–Stokes equations with LPS modelingApplied Mathematics and Computation, 2018
- Supremizer stabilization of POD–Galerkin approximation of parametrized steady incompressible Navier–Stokes equationsInternational Journal for Numerical Methods in Engineering, 2014
- Variational multiscale proper orthogonal decomposition: Navier‐stokes equationsNumerical Methods for Partial Differential Equations, 2013
- Reduced basis approximation and a posteriori error estimation for Stokes flows in parametrized geometries: roles of the inf-sup stability constantsNumerische Mathematik, 2013
- Numerical solution of parametrized Navier–Stokes equations by reduced basis methodsNumerical Methods for Partial Differential Equations, 2007
- Reference values for drag and lift of a two‐dimensional time‐dependent flow around a cylinderInternational Journal for Numerical Methods in Fluids, 2004
- Galerkin proper orthogonal decomposition methods for parabolic problemsNumerische Mathematik, 2001
- Numerical Approximation of Partial Differential Equations1994