A positivity preserving and entropy stable nodal discontinuous Galerkin scheme for ideal MHD equations

This paper develops a new nodal discontinuous Galerkin scheme for ideal MHD equations that simultaneously ensures divergence-free properties, positivity preservation, and entropy stability by combining an HLL numerical flux with a locally divergence-free projection and an essentially oscillation-free damping mechanism.

The Gist

Imagine you are trying to simulate the birth of a star or the swirling plasma inside a fusion reactor using a supercomputer. To do this, you use mathematical equations called MHD (Magnetohydrodynamics). These equations describe how a "soup" of electrically charged particles (plasma) moves and interacts with magnetic fields.

However, simulating this "soup" is like trying to predict the movement of a thousand angry, swirling hurricanes at once. It is incredibly difficult because the math often "breaks" in three specific ways.

This paper introduces a new mathematical "recipe" (a numerical scheme) that fixes these three problems. Here is the breakdown:

1. The Three "Glitch" Problems

Problem A: The "Ghost Magnet" (Divergence Error)
In nature, magnetic fields are very strict: they always form closed loops. You can’t have a "North Pole" of a magnet sitting by itself without a "South Pole." In computer simulations, however, math errors often create "ghost magnets"—tiny, fake magnetic poles that pop up out of nowhere. These ghosts act like little explosions, causing the whole simulation to crash.

• The Fix: The authors added a "Cleanup Crew" (called a Locally Divergence-Free projection) that constantly sweeps the simulation to make sure no ghost magnets are left behind.

Problem B: The "Negative Soup" (Positivity Preservation)
In the real world, things like density and pressure can never be negative. You can have zero density (a vacuum), but you can't have negative density. But computers are literal-minded; if a math calculation overshoots a tiny bit, it might say the density is -0.001. The moment that happens, the math "breaks," and the simulation turns into digital nonsense.

• The Fix: They installed a "Safety Guardrail" (a Positivity Preserving limiter). If the math starts heading toward a negative number, the guardrail gently nudges it back into the realm of physical reality.

Problem C: The "Chaos vs. Order" (Entropy Stability)
Physics has a rule called the Second Law of Thermodynamics: things naturally move from order to disorder (entropy increases). In a simulation, if the math doesn't respect this, you might see "unphysical" behavior—like a shockwave suddenly turning into a smooth, calm wave, which is impossible in real life.

• The Fix: They designed a "Smart Traffic Controller" (an Entropy Stable flux). This ensures that energy is dissipated correctly, just like how friction slows down a sliding box, keeping the simulation's "chaos" realistic.

---

2. The "Secret Sauce": How they did it

The authors didn't just invent one new thing; they combined the best parts of two different existing methods.

Think of it like building a high-performance car.

• Method 1 was like a car with a perfect, smooth engine (it was great at handling smooth flows) but had terrible brakes (it crashed when it hit a bump/shockwave).
• Method 2 was like a rugged off-road Jeep (it could handle bumps) but was very bumpy and jittery to drive.

The authors created a Hybrid Supercar. They used a sophisticated "engine" (the Nodal Discontinuous Galerkin method) and paired it with a high-tech "braking and suspension system" (the HLL flux and LDF projection).

3. Does it actually work? (The Test Drive)

To prove their "car" works, they put it through several extreme "obstacle courses":

• The Vortex: Making the plasma spin like a whirlpool.
• The Blast Wave: Simulating a massive explosion with huge pressure differences.
• The Astrophysical Jet: Simulating a high-speed beam of plasma shooting through space at incredible speeds.

The Result: The simulation didn't crash, it didn't create "ghost magnets," it didn't produce negative density, and it followed the laws of physics perfectly.

Summary for a Non-Scientist

The paper provides a more robust, "smarter" way for computers to simulate the complex, violent movements of plasma. It ensures the math stays physically realistic—preventing "impossible" things like negative matter or fake magnetic poles—even when simulating massive explosions or cosmic jets.

Technical Summary

Technical Summary: A Positivity Preserving and Entropy Stable Nodal Discontinuous Galerkin Scheme for Ideal MHD Equations

Authors: Yue Wu and Chi-Wang Shu
Subject: Computational Magnetohydrodynamics (MHD), High-Order Numerical Methods

---

1. The Problem: Challenges in Ideal MHD Simulations

The ideal MHD equations describe the motion of perfectly conducting, quasi-neutral plasmas. Numerically solving these equations is notoriously difficult due to three interconnected challenges:

• Magnetic Divergence-Free Condition ($\nabla \cdot \mathbf{B} = 0$): This is an inherent physical constraint (no magnetic monopoles). Violating this condition leads to unphysical solutions and numerical instability.
• Positivity Preservation (PP): In high-speed or high-Mach number flows, numerical oscillations near shocks can cause density ($\rho$) or pressure ($p$) to become non-positive, leading to immediate code failure.
• Entropy Stability (ES): Because MHD equations are nonlinear hyperbolic conservation laws, weak solutions are not unique. A robust scheme must satisfy the second law of thermodynamics (entropy inequality) to ensure the numerical solution converges to the physically relevant solution.

Existing Discontinuous Galerkin (DG) methods typically solve one or two of these issues (e.g., modal versions are divergence-free/PP, while nodal versions are entropy stable), but a unified nodal scheme addressing all three has been lacking.

2. Methodology: The Proposed PP-ES-DG Scheme

The authors develop a nodal DG scheme that integrates several advanced mathematical frameworks to achieve a "triple threat" of robustness.

A. Mathematical Framework:

• Godunov–Powell (GP) Source Term: The scheme uses the GP source term to handle the magnetic divergence. This term allows the scheme to satisfy the second law of thermodynamics even when $\nabla \cdot \mathbf{B}$ is non-zero, providing numerical stability against divergence errors.
• Summation-By-Parts (SBP) Property: The scheme utilizes Legendre–Gauss–Lobatto (LGL) nodes and a flux-differencing framework to ensure the discrete operators mimic the integration-by-parts property, which is foundational for entropy stability.

B. Key Algorithmic Components:

1. Entropy Stable (ES) HLL Numerical Flux: The authors derive a specific HLL (Harten-Lax-van Leer) flux. Unlike standard penalization methods, they provide a rigorous bound for signal speed estimates that guarantees entropy stability while simultaneously allowing for positivity preservation.
2. Locally Divergence-Free (LDF) Projection: To ensure the PP limiter works correctly, the magnetic field must be locally divergence-free. The authors implement an element-wise post-processing step using a Piola transformation and a saddle-point solver to project the magnetic field into a divergence-free space.
3. Hybrid Limiting Strategy: The scheme follows a specific sequence to maintain stability without losing accuracy:
   • Step 1: LDF Projection (to fix $\nabla \cdot \mathbf{B}$).
   • Step 2: OEDG Damping (Essentially Oscillation-Free damping to suppress Gibbs oscillations near shocks).
   • Step 3: Zhang–Shu PP Limiter (To strictly enforce $\rho > 0$ and $p > 0$).

3. Key Contributions

• Unified Nodal Scheme: The primary contribution is the first nodal DG scheme that is simultaneously Positivity Preserving, Entropy Stable, and Locally Divergence-Free.
• New Signal Speed Estimates: The derivation of HLL signal speed bounds that satisfy the entropy inequality for the Godunov–Powell MHD system.
• Robust Algorithmic Pipeline: A proven sequence of LDF projection $\rightarrow$ OEDG damping $\rightarrow$ PP limiting that maintains the divergence-free property throughout the limiting process.

4. Results and Validation

The authors validated the scheme through a battery of rigorous numerical experiments:

• Smooth Flows (Alfvén Wave): Demonstrated high-order (up to $k+1$) convergence rates in smooth regions.
• Discontinuities & Shocks (Yee–Sjögreen, Orszag–Tang, Cloud-Shock): Successfully captured complex shock-vortex interactions and rotational discontinuities without unphysical oscillations or entropy violations.
• Extreme Conditions (Blast Wave & Astrophysical Jet): Proved robustness in low-plasma-beta regimes (where magnetic pressure vastly exceeds thermal pressure) and high-Mach number flows (Mach 800), where most schemes fail due to negative pressure.
• Entropy Evolution: Confirmed that the total mathematical entropy decreases over time, as required by the second law of thermodynamics.

5. Significance

This work provides a highly robust tool for plasma physics simulations. By solving the convergence of divergence-free, positivity-preserving, and entropy-stable properties, the authors have created a scheme capable of handling the most extreme scenarios in astrophysics (e.g., jets, supernovae) and fusion technology, where traditional high-order methods often break down.