Well-balanced POD-based reduced-order models for finite volume approximation of hyperbolic balance laws Journal: Journal of Computational and Applied Mathematics
I. Gómez-Bueno, E.D. Fernández-Nieto, S. Rubino · 2025
Read Well-balanced POD-based reduced-order models for finite volume approximation of hyperbolic balance laws Journal: Journal of Computational and Applied Mathematics by I. Gómez-Bueno, E.D. Fernández-Nieto, S. Rubino (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/arxiv-2505-03237.txt.
Introduction
Gómez-Bueno and co-authors shrink heavy flow simulations without losing their balance in a 2025 computing paper. Hyperbolic balance laws govern flows where waves race against source terms, like shallow water over uneven beds. A good model must hold still water exactly still, a property called well-balancing. Their reduced models compress full finite-volume runs with proper orthogonal decomposition, patched by interpolation tricks that keep the cost down.
Tests on transport, Burgers, and shallow-water problems beat older time-averaging approaches in speed and accuracy. A theory result justifies building the reduced model on a well-balanced full one. Parameter sweeps and sensitivity checks confirm robustness, the point of a surrogate meant for many-query work.
Catalog record
| CREATOR | I. Gómez-Bueno, E.D. Fernández-Nieto, S. Rubino |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 81 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-03237
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Well-balanced POD-based reduced-order models for finite volume approximation of hyperbolic balance laws Journal: Journal of Computational and Applied Mathematics I. Gómez-Bueno, E.D. Fernández-Nieto, S. Rubino Abstract This paper introduces a reduced-order modeling approach based on finite volume methods for hyperbolic systems, combining Proper Orthogonal Decomposition (POD) with the Discrete Empirical Interpolation Method (DEIM) and Proper Interval Decomposition (PID). Applied to systems such as the transport equation with source term, non-homogeneous Burgers equation, and shallow water equations with non-flat bathymetry and Manning friction, this method achieves significant improvements in computational efficiency and accuracy compared to previous time-averaging techniques. A theoretical result justifying the use of well-balanced Full-Order Models (FOMs) is presented. Numerical experiments validate the approach, demonstrating its accuracy and efficiency. Furthermore, the question of prediction of solutions for systems that depend on some physical parameters is also addressed, and a sensitivity analysis on POD parameters confirms the model’s robustness and efficiency in this case. MSC 76M12 , 35L60 , 65M08 † † highlights: Efficient POD-ROM combines DEIM, PID, and FV methods for hyperbolic systems. Proves ROMs inherit well-balanced property from well-balanced FOMs. DEIM strategy enhances efficiency, accuracy, and reduces spurious oscillations. Predictive ROMs validated for parameterized systems via sensitivity analysis. Applied to transport and Burgers equations with source terms and shallow water system with bottom and Manning friction. 1 Introduction The numerical resolution of hyperbolic equations for realistic applications could require high computational costs, due to the use of very fine grids to properly capture complex shock phenomena or discontinuities, and also integration over long periods in time. Reduced-Order Models (ROMs) applied to numerical design of complex physics problems are a proper mathematical tool that is wide-spreading in the scientific community in the recent years in order to save computational time without so much loss of accuracy with respect to classical methods such as Finite Difference (FD), Finite Element (FE) or Finite Volume (FV) methods [ 29 ] . Among the most popular ROMs approaches, Proper Orthogonal Decomposition (POD) strategy provides optimal (from the energetic point of view) basis or modes to represent the dynamics from a given database (snapshots) obtained by a Full-Order Model (FOM). Onto these reduced bases, a Galerkin projection of the governing equations can be employed to obtain a low-order dynamical system for the basis coefficients. The resulting low-order model is named standard POD-(Galerkin) ROM, which thus consists in the projection of high-fidelity (full-order) representations of physical problems onto low-dimensional spaces of solutions, with a dramatically reduced dimension. These low-dimensional spaces are capable of capturing the dominant characteristics of the solution, their main advantage being that the computations in the low-dimensional space can be done at a reduced computational cost [ 32 , 38 ] . This has led researchers to apply POD-ROMs to a variety of physical and engineering problems, including Computational Fluid Dynamics (CFD) problems in order to model advection-diffusion equations, see e.g. [ 6 , 24 ] , the Navier–Stokes Equations (NSE), see e.g. [ 7 , 8 , 22 , 35 , 13 , 42 ] , and also hyperbolic equations, such as Shallow Water (SW) equations [ 4 , 19 , 40 , 43 , 44 ] . Although POD-ROMs can be very computationally efficient and relatively accurate in some flow configurations, they also present several drawbacks, when applied for instance to strongly nonlinear systems. In this direction, several techniques have been presented for hyperbolic systems, such as the dynamic mode decomposition [ 4 , 37 ] , the Discrete Empirical Interpolation Method (DEIM) [ 14 , 17 , 21 ] , the tensorial POD [ 41 ] , and also non-intrusive hyper-reduction techniques [ 4 ] . Furthermore, to address this issue, a recent paper [ 39 ] uses the time averaging approach in combination with the Proper Interval Decomposition (PID, [ 3 , 30 , 44 ] ) applied to SW using augmented Riemann solvers. Moreover, it should be emphasized that, in the context of hyperbolic problems, reduced order modeling techniques suffer from fundamental limitations when used to predict solutions beyond the training time window. This issue arises due to the highly nonlinear and propagative nature of the solutions to systems of balance laws, which may develop discontinuities or rapidly evolving features whose dynamics are not adequately represented by the dominant modes extracted from limited-time snapshots. Consequently, the reliability of reduced models in extrapolative regimes may be affected, and care should be taken when applying them outside the training interval. In this work, we focus on the use of POD-based ROMs for generic hyperbolic systems with source terms. In particular, we show how the use of DEIM technique to deal with nonlinearities combined with PID outperforms the time averaging approach in combination with PID proposed in [ 39 ] , especially in terms of physical accuracy. These PDE systems have been extensively used to model waves caused by small disturbances in steady solutions. A notable example is tsunami waves in the ocean. A numerical method that can solve all steady states of the PDE exactly, or with improved accuracy, or at least a significant subset of them, is referred to as exactly Well-Balanced (EWB). Additionally, the method is classified as fully exactly well-balanced if it exactly preserves all the stationary solutions of the hyperbolic system (see [ 26 ] ). The design of well-balanced schemes is a highly active area of research, and many studies focus on this topic, specially for the case of finite volume methods. For example, see [ 27 ] for the Burgers equation; [ 1 , 5 , 15 , 16 , 33 , 34 ] for the shallow water equations; and [ 2 , 9 , 12 , 20 , 31 ] for other hyperbolic systems, among others. In this paper we present an original theoretical result of consistency at ROM level, to justify the choice of well-balanced methods to build the FOM, which is independent of the considered hyperbolic balance law. Three examples of well-balanced POD-based methods are presented, for the transport equation with a source term, non-homogeneous Burgers equation and for the SW equations with topography and Manning friction. Finally, following other works focused on reduced order modelling for parametrized problems (see [ 17 ] ), the question of prediction of solutions for physical parameter-dependent systems is also addressed in this paper, as well a sensitivity analysis to specific parameters of the proposed POD-ROM, such as the number of POD modes and time windows in terms of the errors between FOM and ROM solutions and the CPU times they require. The rest of the paper is organized as follows: first, we introduce in Section 2 the finite volume full-order models for general one-dimensional hyperbolic PDE systems, detailing the numerical schemes and discussing the construction of exactly well-balanced schemes. In Subsection 2.1 , we specify the FOMs for three examples of balance laws: the linear transport equation with source term, the Burgers equation with a nonlinear source term, and the shallow water equations with and without Manning friction. In all cases, we use PVM-0 type numerical fluxes [ 11 ] , and for the shallow water case, we additionally apply the HLL scheme [ 28 ] . Section 3 outlines the strategy for