🔢 mathematics

On Affordable High-Order Entropy-Conservative/Stable and Well-Balanced Methods for Nonconservative Hyperbolic Systems

This paper proposes a specific formulation of entropy-preserving fluctuations for nonconservative hyperbolic systems that enables the algorithmic construction of high-order, well-balanced, and robust numerical schemes within both finite volume and summation-by-parts frameworks, as demonstrated through new methods for the compressible Euler and dispersive shallow-water equations.

Original authors: Marco Artiano, Hendrik Ranocha

Published 2026-03-20

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Marco Artiano, Hendrik Ranocha

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are a chef trying to bake the perfect cake. You have a recipe (the laws of physics) that tells you exactly how ingredients should mix and move. But sometimes, the recipe is tricky. It involves ingredients that don't just sit still; they interact in complex ways that aren't perfectly "conservative" (meaning they don't just add up neatly like counting apples).

This paper is about a team of mathematicians (Marco Artiano and Hendrik Ranocha) who invented a new, smarter way to bake these tricky cakes on a computer. They call their method "Affordable High-Order Entropy-Conservative/Stable and Well-Balanced Methods."

That sounds like a mouthful, so let's break it down using some everyday analogies.

1. The Problem: The "Leaky" Recipe

In physics, many things (like air flowing around a plane or water in a river) follow strict rules. Usually, we can calculate these using "Conservation Laws" (like Conservation of Energy: energy can't be created or destroyed, only changed).

However, some systems are Nonconservative. Think of it like a recipe where you have to add a pinch of salt while stirring, and the amount of salt depends on how fast you're stirring. It's messy.

The Old Way: Previous computer methods tried to solve this by taking a long, winding path through the math (integrals). It was like trying to measure the distance of a hike by walking every single step of the trail. It was accurate but incredibly slow and expensive (computationally "expensive").
The Risk: If the computer gets the math slightly wrong, the simulation can blow up. It might create fake energy out of nowhere or lose real energy, leading to a "crash" or a nonsensical result (like a simulation of a storm that suddenly turns into a vacuum).

2. The Solution: The "Magic Ledger" (Entropy)

The authors use a concept called Entropy. In simple terms, think of entropy as the universe's "Ledger" or "Accounting System."

In a smooth, perfect world, the ledger balances perfectly (Entropy is conserved).
In a messy, real world with shocks or turbulence, the ledger can only stay the same or get "messier" (Entropy increases), but it can never magically decrease.

The goal of this paper is to build a computer code that never cheats the ledger. It ensures that the computer simulation respects the laws of physics so strictly that it can't accidentally create energy or break the rules.

3. The "Affordable" Trick: No More Walking the Trail

The authors' big breakthrough is making this "Ledger-keeping" affordable.

The Old Way: To check the ledger, you had to do complex integrals (mathematical summations) that were hard to calculate.
The New Way: They found a way to check the ledger using simple algebra (basic addition and multiplication).
- Analogy: Instead of walking the entire hiking trail to measure the distance, they realized you could just look at the start and end points and use a simple formula to get the exact answer instantly.
- They call this "Algorithmic Construction." It's like having a magic calculator that instantly tells you the right number to write in the ledger without doing all the heavy lifting.

4. The "Well-Balanced" Feature: The Perfectly Still Lake

One specific problem in physics is simulating a Lake at Rest.

Imagine a lake that is perfectly still. The water is deep in some places and shallow in others because of the bottom terrain.
The Problem: Old computer methods often get confused by the changing depth. They might think the water is moving when it's actually still, creating fake waves that ruin the simulation.
The Fix: The authors' method is "Well-Balanced." It knows the difference between "real movement" and "just sitting on a slope." It keeps the lake perfectly still, just like nature intended, even on a curvy, bumpy computer grid.

5. The "High-Order" and "Curved Mesh" Magic

High-Order: Imagine drawing a circle. A low-order method draws it with straight lines (it looks like a hexagon). A High-Order method draws it with so many tiny, smooth curves that it looks like a perfect circle. This paper allows these super-smooth, high-precision calculations to work on the tricky "Nonconservative" systems.
Curved Meshes: Real-world objects (like a plane wing or the Earth) aren't perfect squares. They are curved. The authors showed how to stretch their "perfect circle" math onto these curved, warped shapes without losing accuracy.

Summary: What Did They Actually Do?

They wrote a new "instruction manual" for computer simulations.

Simplified the Math: They turned complex, slow integrals into simple, fast algebra.
Guaranteed Safety: They proved that if you follow their instructions, the computer will never break the laws of physics (Entropy is preserved).
Handled the Messy Stuff: They applied this to systems that were previously too hard to simulate accurately, like gas flowing with internal energy or water with complex waves.
Tested It: They ran simulations of everything from simple waves to complex atmospheric storms, and the results were stable, accurate, and fast.

In a nutshell: They built a faster, safer, and more accurate way to simulate complex physical systems on computers, ensuring that the digital world respects the same rules as our real world, without needing a supercomputer to do the heavy math.

1. Problem Statement

The paper addresses the challenge of developing robust, high-order numerical methods for nonconservative hyperbolic systems of the form:
$\partial_t u + \partial_x f(u) + H(u)\partial_x g(u) = 0$
While entropy-conserving (EC) and entropy-stable (ES) methods are well-established for classical conservation laws (where $H=0$ ) based on Tadmor's theory, extending these to nonconservative systems is difficult.

Existing Limitations: Previous approaches (e.g., Castro et al., 2013) rely on path-conservative formulations involving integrals along specific paths connecting states. These integrals are often analytically intractable or computationally expensive ("unaffordable") to evaluate exactly.
The Gap: There is a lack of a systematic, algebraic framework that allows for the construction of affordable (integral-free) EC/ES fluxes for general nonconservative systems, particularly on high-order Summation-by-Parts (SBP) operators and curvilinear meshes. Furthermore, the necessity of entropy convexity in previous derivations limits the applicability to certain functionals (e.g., pressure equilibrium preservation).

2. Methodology

The authors propose a framework based on purely algebraic characterizations of entropy conservation, removing the need for path integrals and relaxing the requirement for entropy convexity.

A. Algebraic Characterization of Entropy Conservation

Generalization of Tadmor's Condition: The paper proves that Tadmor's "shuffle condition" is not only sufficient but also necessary for entropy conservation, even for non-convex functionals. This allows the derivation of conservation relations for non-standard variables (e.g., total energy with potential temperature).
Fluctuation Form: For nonconservative systems, the authors analyze the system in fluctuation form:
$\partial_t u_i + \frac{1}{\Delta x_i} (D^+ + D^-) = 0$
They derive necessary and sufficient algebraic conditions for the fluctuations $D^\pm$ to ensure entropy conservation.

B. Four Semi-Discrete Forms

The paper systematically analyzes four specific semi-discretizations for the nonconservative term $H(u)\partial_x g(u)$ :

First Form: Pointwise evaluation of $H$ with central differences of $g$ .
Second Form: $H$ evaluated at the cell center, $g$ approximated by a numerical flux.
Third Form: $H$ and $g$ both evaluated via numerical fluxes (split form).
Fourth Form: A flux-differencing form involving $(Hg)$ and $H$ .

For each form, the authors derive specific algebraic conditions that the numerical fluxes must satisfy to preserve entropy. They demonstrate that a linear combination of the first and third forms (parameterized by $\alpha$ ) offers the most flexibility for constructing affordable schemes.

C. Extension to High-Order SBP and Curvilinear Meshes

SBP Operators: The low-order finite volume results are extended to high-order Discontinuous Galerkin Spectral Element Methods (DGSEMs) and general SBP operators. The volume terms are constructed using flux-differencing techniques that mimic the continuous product rule.
Curvilinear Coordinates: The framework is extended to curved meshes using the "free-stream preservation" property. The metric terms are discretized to satisfy discrete metric identities, ensuring that the entropy analysis holds on general geometries.
Well-Balancedness: The authors derive algebraic conditions under which these high-order schemes preserve steady states (e.g., lake-at-rest). They prove that if the underlying two-point fluxes are well-balanced, the high-order SBP extension on curved meshes automatically inherits this property.

3. Key Contributions

Necessity and Sufficiency without Convexity: The paper establishes that Tadmor's algebraic condition is necessary and sufficient for entropy conservation without requiring the entropy function to be convex. This broadens the scope to include non-convex functionals like pressure equilibrium preservation.
Affordable Flux Construction: By avoiding path integrals, the authors provide a recipe for constructing "affordable" EC/ES fluxes. The method relies on solving algebraic equations for the numerical fluxes rather than evaluating complex integrals.
Systematic Classification: The paper provides a complete characterization of four distinct semi-discrete forms for nonconservative terms, linking each to specific algebraic entropy conditions.
Unified Framework for Well-Balancedness: A general condition is derived that extends well-balanced properties from low-order finite volume methods to high-order SBP methods on curvilinear meshes.
New Numerical Fluxes: The framework is used to derive new, provably entropy-conserving/stable fluxes for:
- Variable-coefficient advection.
- Coupled Burgers' equations.
- General polynomial equations.
- Shallow-water equations (including dispersive Sainte-Marie models).
- Ideal Magnetohydrodynamics (MHD) with GLM divergence cleaning.
- Compressible Euler equations using internal energy as a thermodynamic variable (a challenging case due to non-conservative pressure work terms).

4. Results

Numerical experiments were conducted using the Julia-based Trixi.jl framework with DGSEMs:

Entropy Conservation: Tests on variable-coefficient advection, general polynomial equations, and the hyperbolized Sainte-Marie equations showed that the proposed schemes conserve entropy (and total energy) up to machine precision ( $O(10^{-12})$ to $O(10^{-15})$ ) across various polynomial degrees ( $p=1$ to $5$).
Well-Balancedness: The scheme successfully preserved the "lake-at-rest" steady state on warped, curvilinear meshes with discontinuous bottom topography up to $T=100$ , with errors remaining at machine precision.
Accuracy: Convergence studies for the 2D compressible Euler equations demonstrated the expected order of accuracy for the high-order SBP schemes.
Robustness: A long-term simulation (20 days) of the 3D baroclinic instability benchmark using the entropy-stable Euler flux showed that the scheme maintains entropy stability over long integration times, preventing numerical blow-up in under-resolved turbulent flows.

5. Significance

This work represents a significant advancement in the numerical analysis of hyperbolic systems:

Algorithmic Construction: It moves away from heuristic or integral-based flux construction toward a systematic, algebraic procedure that is easier to implement and automate.
Broad Applicability: By decoupling the entropy condition from entropy convexity and path integrals, the method applies to a wider class of physical models, including those with complex equations of state and non-conservative source terms.
High-Order Geometry: It bridges the gap between low-order well-balanced schemes and high-order methods on complex, curved geometries, which is crucial for geophysical fluid dynamics (atmospheric and oceanic modeling).
Practical Impact: The derived fluxes for the Euler equations with internal energy are particularly valuable for atmospheric modeling, where preserving pressure equilibrium and total energy is critical for stability and physical fidelity.

In summary, the paper provides a rigorous, flexible, and computationally efficient framework for designing structure-preserving numerical methods for a wide range of challenging nonconservative hyperbolic systems.