🔢 mathematics

Semi-Implicit Central scheme for Hyperbolic Systems of Balance Laws with Relaxed Source Term

This paper proposes and validates a novel semi-implicit IMEX finite volume central scheme that effectively handles stiff source terms in hyperbolic balance laws by combining spatial midpoint, temporal trapezoidal, and backward semi-implicit Taylor expansion rules, thereby ensuring stability, robustness near discontinuities, and second-order accuracy for both one- and two-dimensional systems.

Original authors: Sudipta Sahu, Emanuele Macca, Rathan Samala

Published 2026-03-30

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Sudipta Sahu, Emanuele Macca, Rathan Samala

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 trying to predict how a crowd of people moves through a busy train station. Some people are walking normally (the "flow"), but others are suddenly stopping to check their phones or tie their shoes (the "source").

In the world of physics and engineering, this is similar to solving hyperbolic systems of balance laws. These are equations that describe how things like water, air, or heat move and change over time.

The problem arises when the "stopping" part happens incredibly fast. In math terms, this is called a stiff source term. It's like if the people checking their phones did it so fast that your computer simulation had to take a tiny, tiny step every single time just to keep up. If you tried to simulate this with a standard method, your computer would have to wait a million years to finish a simulation that should take seconds.

This paper introduces a new, clever way to solve these problems called the CS-EBT2 scheme. Here is how it works, broken down into simple concepts:

1. The Problem: The "Speed Trap"

Think of the equations as a car driving down a highway.

The Flow (Convective Flux): This is the car driving at a steady speed.
The Stiff Source: This is a sudden, violent brake that happens thousands of times a second.

If you use a standard "explicit" method (like looking at the road only once per second), you will crash because you missed the brakes. To avoid crashing, you have to look at the road a million times a second. This is slow and inefficient.

2. The Solution: The "Smart Co-Pilot" (Semi-Implicit)

The authors created a new method that acts like a smart co-pilot. Instead of just looking at the road, the co-pilot predicts the future.

The "Explicit" Part: For the normal driving (the flow), the car just keeps going as usual.
The "Implicit" Part: For the sudden brakes (the stiff source), the co-pilot says, "I know you're going to brake hard, so let's calculate the result of that braking before we actually move."

By doing this, the car doesn't need to take tiny steps. It can take big, confident strides while still knowing exactly where the brakes are. This is called a Semi-Implicit approach.

3. The Secret Sauce: The "Trick" (Modified Trapezoidal Rule)

How does the co-pilot predict the future so accurately without doing heavy math?
The authors use a mathematical trick called a Modified Trapezoidal Rule combined with a Backward Taylor Expansion.

Analogy: Imagine you are trying to guess the temperature of a cup of coffee in 5 minutes.
- A simple guess might be: "It's hot now, so it will be slightly less hot later."
- This new method is like saying: "I know the coffee cools down based on a specific curve. I'll look at the current temperature, guess the future temperature, and then use a special formula to correct my guess instantly."

This allows the computer to handle the "stiff" (fast) parts of the equation without getting confused or crashing, all while keeping the math simple enough to run quickly.

4. The "Shape-Shifter" (Asymptotic Preserving)

One of the coolest features of this method is that it is Asymptotic Preserving (AP).

Analogy: Imagine a video game character that can run fast or walk slowly.
- When the "stiffness" is low (the brakes are weak), the character runs normally.
- When the "stiffness" is high (the brakes are super strong), the character doesn't freeze or glitch. Instead, the character smoothly transforms into a different mode (like walking on a treadmill) that perfectly matches the physics of the situation.

This means the same computer code works perfectly whether the physics are slow and easy, or fast and chaotic. You don't have to rewrite the code or change the settings; it just adapts automatically.

5. Testing the New Car

The authors didn't just build the car; they took it for a spin on five different tracks:

Jin-Xin Model: A basic test track for fluid flow.
Shallow Water: Simulating floods and waves.
Broadwell Model: A complex gas dynamics test.
Euler with Heat: Simulating air with temperature changes.
Euler with Friction: Simulating air slowing down due to drag.

They also tested it in 2D (like a flat map instead of a straight line). In every test, their new method was:

Faster: It could take bigger steps (larger time steps).
More Accurate: It didn't lose precision even when things got messy.
Robust: It didn't crash when the data was jagged or discontinuous (like a shockwave).

The Bottom Line

This paper presents a new "Swiss Army Knife" for scientists who simulate fluid dynamics, weather, or explosions. It solves the age-old problem of "stiff" equations by using a clever prediction trick that lets computers run simulations much faster without sacrificing accuracy. It's like upgrading from a bicycle that needs to stop at every red light to a self-driving car that knows exactly how to navigate traffic jams.

1. Problem Statement

The paper addresses the numerical challenges associated with hyperbolic systems of balance laws containing stiff source terms (relaxation terms). These systems arise in various physical applications, including fluid dynamics, geophysical modeling, and kinetic theory.

The Core Conflict: Explicit numerical schemes are computationally efficient but suffer from severe time-step restrictions (CFL conditions) when the source term is stiff (i.e., the relaxation parameter $\varepsilon \to 0$ ). Conversely, fully implicit schemes are unconditionally stable but computationally expensive due to the need to solve large, nonlinear systems at every time step.
The Goal: Develop a numerical method that is asymptotic-preserving (AP) and asymptotic-accurate (AA). This means the scheme must remain stable and accurate as $\varepsilon \to 0$ (capturing the correct equilibrium limit) without requiring the time step to scale with $\varepsilon$ , while maintaining high-order accuracy and avoiding the complexity of Riemann solvers.

2. Methodology

The authors propose a novel Semi-Implicit Central Scheme, termed CS-EBT2 (Central Scheme with Euler Backward Taylor, second-order). The method is a finite volume extension of the Nessyahu-Tadmor (NT) central scheme.

Key Components:

Finite Volume Framework: The solution is represented by cell averages. To avoid Riemann solvers, the scheme uses central differencing for convective fluxes.
Staggered Grids: Following the NT approach, the scheme evolves solutions on staggered control volumes ( $[x_i, x_{i+1}]$ ) to enhance stability and dissipation properties without upwinding.
Semi-Implicit Source Treatment:
- Convective Fluxes: Treated explicitly using a midpoint rule in time and piecewise-linear reconstruction (with MinMod limiter) in space.
- Stiff Source Terms: Treated semi-implicitly. The authors derive a new integration strategy for the source term integral over a space-time cell.
- Derivation: The source term integral is approximated by:
  - Spatial integration via the midpoint rule.
  - Temporal integration via the trapezoidal rule.
  - A backward semi-implicit Taylor expansion is used to express the source term at the current time step ( $t_n$ ) in terms of the future time step ( $t_{n+1}$ ) and its derivative. This linearizes the implicit part, allowing for an efficient solution without solving large coupled systems.
Predictor-Corrector Structure:
- Predictor: Solves for intermediate values at $t_{n+1/2}$ using a semi-implicit update for the source.
- Corrector: Updates the cell averages at $t_{n+1}$ on the staggered grid, incorporating the flux differences and the semi-implicit source term approximation.

Key Features:

Locality: The implicit treatment of the source term is local to each cell. Unlike many IMEX schemes that couple neighboring cells, this method decouples the nonlinear systems, requiring only the solution of small systems (size equal to the number of conserved variables) per cell.
No Riemann Solvers: The central nature of the scheme eliminates the need for expensive Riemann solvers.

3. Key Contributions

Novel Finite Volume Formulation: The paper adapts a modified trapezoidal rule for source term integration (previously used in finite difference frameworks) into a finite volume setting, ensuring conservation and consistency.
Asymptotic Properties:
- AP (Asymptotic Preserving): As $\varepsilon \to 0$ , the scheme naturally degenerates to a consistent discretization of the equilibrium (reduced) model without modifying the time step.
- AA (Asymptotic Accurate): The scheme maintains second-order accuracy uniformly as $\varepsilon \to 0$ . Theoretical analysis confirms that the order of accuracy does not degrade in the stiff limit.
Stability Analysis: The authors provide a rigorous linear stability analysis, proving the scheme is A-stable and L-stable for the stiff source term. The stability region is shown to be comparable to, or larger than, standard IMEX-RK2 schemes, allowing for larger CFL numbers (up to 0.9 in tests).
Efficiency: By avoiding Riemann solvers and large coupled linear systems, the method offers a computationally efficient alternative to standard IMEX Runge-Kutta schemes.

4. Numerical Results and Validation

The method was validated against five benchmark models in both 1D and 2D:

Jin-Xin Relaxation Model:
- Tested with smooth (well-prepared and unprepared) and discontinuous initial data.
- Results: Achieved second-order convergence for both variables $u$ and $v$ in the stiff regime ( $\varepsilon = 10^{-10}$ ). The scheme captured sharp discontinuities without spurious oscillations.
Shallow Water Model:
- Tested with smooth and non-smooth (unprepared) data.
- Results: Successfully captured wave steepening and relaxation dynamics, matching high-resolution reference solutions.
Broadwell Model:
- Validated for smooth and discontinuous profiles.
- Results: Confirmed second-order accuracy across a wide range of $\varepsilon$ ( $10^{-14}$ to $1$).
Euler Equations with Heat Transfer:
- Tested with discontinuous initial data in the stiff heat-transfer regime.
- Results: Accurately resolved shock waves and thermal relaxation.
Euler System with Stiff Friction:
- Tested both full and isentropic Euler systems with large friction coefficients.
- Results: The scheme correctly captured the diffusive limit (Darcy-type law) and maintained stability for large friction parameters ( $\alpha = 10^8$ ).
2D Extensions:
- Applied to 2D Jin-Xin and 2D Euler Jin-Xin-type relaxation systems (Riemann problems).
- Results: Demonstrated robustness and accuracy in multidimensional settings with complex wave interactions.

5. Significance and Conclusion

The CS-EBT2 scheme represents a significant advancement in the numerical simulation of stiff hyperbolic systems. Its primary significance lies in:

Robustness: It handles both smooth and discontinuous solutions in stiff regimes without the restrictive time-step limitations of explicit methods.
Efficiency: The local nature of the implicit solve and the avoidance of Riemann solvers make it highly suitable for parallel computing and large-scale simulations.
Simplicity: It achieves second-order AP/AA accuracy with only two implicit stages, simplifying implementation compared to higher-order IMEX schemes that often require deferred correction or multiple stages.

The authors conclude that the method is a reliable, efficient, and accurate tool for solving stiff balance laws in 1D and 2D. Future work will focus on extending the scheme to higher-order accuracy (beyond second order) and incorporating adaptive mesh refinement.