A high order accurate and energy stable continuous Galerkin framework on summation-by-parts form for the incompressible Navier-Stokes equations

This paper presents a high-order, energy-stable Continuous Galerkin finite element framework using a Summation-By-Parts (SBP) formulation and the Simultaneous Approximation Term (SAT) technique to accurately and efficiently solve the incompressible Navier-Stokes equations, even in the presence of discontinuous boundary conditions.

The Gist

Imagine you are trying to simulate how water flows through a complex plumbing system or how air moves over an airplane wing. To do this, scientists use massive mathematical "blueprints" called the Navier-Stokes equations.

The problem is that these equations are incredibly "fussy." If your mathematical model isn't perfect, the simulation can "explode"—meaning it creates wild, unrealistic wobbles (oscillations) or simply crashes because it can't handle the math.

This paper introduces a new, high-tech way to build these mathematical blueprints that is both super accurate and rock-solid stable.

Here is the breakdown of how they did it, using everyday analogies.

---

1. The Problem: The "Corner Clash" (Discontinuous Boundaries)

Imagine you are painting a room. On one wall, you want bright red, but on the adjacent wall, you want deep blue. At the very corner where the two walls meet, there is a sharp, impossible line where red must instantly become blue.

In fluid simulations, this happens in something called a "Lid-Driven Cavity" (imagine a box of water where the top lid is sliding). At the top corners, the water is being pushed hard by the lid, but the side walls are standing still. This "clash" of speeds creates a mathematical "glitch" that usually causes the simulation to vibrate wildly, like a guitar string being plucked too hard.

2. The Solution: The "Soft Hand" Approach (SAT Technique)

Most traditional methods try to force the math to obey the boundary rules strictly—like trying to hammer a square peg into a round hole. If the peg doesn't fit perfectly, the wood cracks. This is called "strong enforcement."

The researchers used a technique called SAT (Simultaneous Approximation Term). Instead of hammering the peg, imagine using a soft, flexible hand to guide the peg into place. If there is a slight mismatch at the corner, the SAT method "negotiates" with the boundary. It says, "I know there's a jump here, so I'll apply a gentle corrective pressure to smooth it out." This prevents those wild, non-physical wobbles and keeps the simulation calm and smooth.

3. The Framework: The "Lego-Style" Precision (SBP-CGFEM)

The researchers used a framework called Summation-By-Parts (SBP) within a Continuous Galerkin (CGFEM) method.

• The CGFEM part is like using high-resolution Lego bricks. Instead of using big, chunky blocks to represent the flow, they use very sophisticated, "high-order" shapes (Lagrange polynomials) that can curve and bend beautifully to match the actual flow of the liquid.
• The SBP part is like a built-in "accounting system." In physics, energy must be conserved (it can't just appear out of nowhere). The SBP method ensures that every bit of mathematical "energy" is accounted for. It’s like having a strict accountant in your simulation who ensures that no "fake energy" is created by the math, which guarantees that the simulation won't spiral out of control (this is what they call "Energy Stability").

4. The Proof: The "Stress Test"

To prove their new method works, they put it through three tests:

1. The Math Test (MMS): They created a "fake" perfect world where they already knew the answer, just to see if their math could find it. It did, with incredible precision.
2. The Corner Test (Lid-Driven Cavity): They ran the "sliding lid" simulation. Even with the violent speed changes at the corners, the simulation stayed smooth and matched real-world physics perfectly.
3. The Obstacle Test (Backward-Facing Step): They simulated water flowing past a sudden step in a pipe. This is a classic "stress test" for fluids because it creates swirling whirlpools (vortices). Their method captured these swirls accurately and efficiently.

Summary

In short, these researchers have built a better, smoother, and more stable mathematical engine for simulating fluids. It’s like moving from a jerky, vibrating old car to a high-performance electric vehicle: it’s more precise, it handles the "sharp turns" of physics much better, and it won't break down when the road gets bumpy.

Technical Summary

Technical Summary: High-Order Accurate and Energy Stable CGFEM Framework for Incompressible Navier-Stokes Equations

1. Problem Statement

The paper addresses the numerical simulation of the Incompressible Navier-Stokes (INS) equations, which govern a wide range of viscous fluid phenomena. Solving these equations presents several long-standing challenges in computational fluid dynamics (CFD):

• The Divergence-Free Constraint: Standard velocity-pressure formulations often lead to saddle-point problems that require strict compatibility conditions (the inf-sup or Ladyzhenskaya-Babuška-Brezzi condition) to ensure solution uniqueness.
• Continuity Requirements: Alternative formulations, such as stream function-based approaches, often require $C^1$-continuous basis functions, which are mathematically and computationally difficult to implement.
• Discontinuous Boundary Conditions: Traditional "strong" enforcement of boundary conditions can trigger non-physical oscillations (Gibbs phenomena) at points of discontinuity, such as the corners of a lid-driven cavity.
• Stability and Accuracy: Maintaining high-order accuracy while ensuring energy stability (preventing unphysical energy growth) is critical for robust simulations.

2. Methodology

The authors propose a novel framework using the Continuous Galerkin Finite Element Method (CGFEM) integrated with the Summation-By-Parts (SBP) and Simultaneous Approximation Term (SAT) techniques.

• Formulation: The INS equations are discretized using a primitive variable formulation (velocity $u, v$ and pressure $p$). To bypass the inf-sup condition and the need for $C^1$ continuity, the authors utilize equal-order Lagrange polynomials (up to 4th order) and a skew-symmetric form of the advective terms.
• SBP-SAT Framework:
  • SBP Operators: The spatial derivatives are constructed using SBP operators, which mimic the integration-by-parts property at the discrete level. This ensures that the discrete operators maintain the same mathematical structure as the continuous ones.
  • SAT (Weak Boundary Enforcement): Instead of forcing boundary conditions directly onto the solution (strong enforcement), the authors use the SAT technique to impose them weakly via penalty terms. This allows the method to handle discontinuous boundary data (like the lid-driven cavity corners) without special regularization or smoothing.
• Temporal Discretization: The semi-discrete system is integrated in time using the second-order Backward-Difference Formula (BDF2), and the resulting nonlinear algebraic systems are solved using Newton’s method.

3. Key Contributions

• Novel Application: To the authors' knowledge, this is the first application of the SBP-SAT technique within a CGFEM framework specifically for the nonlinear incompressible Navier-Stokes equations.
• Energy Stability Proof: The paper provides a rigorous mathematical proof of energy stability at both the continuous and discrete levels. By carefully selecting the penalty matrices ($\Sigma$), the authors prove that the discretization satisfies an energy bound, ensuring the solution remains bounded and physically consistent.
• Handling Discontinuities: The framework provides a systematic way to handle discontinuous boundary conditions without the need for ad hoc smoothing functions (like hyperbolic tangents), which are common in other methods.

4. Results and Validation

The framework was validated through three distinct numerical tests:

• Method of Manufactured Solutions (MMS): This confirmed that the implementation is highly accurate, achieving the theoretical $(k+1)$ or $k$-th order convergence for Lagrange polynomials of degree $k$.
• Lid-Driven Cavity Flow: The method was tested across a wide range of Reynolds numbers ($Re = 100$ to $10,000$). The results showed excellent agreement with benchmark data (Ghia et al.) and, crucially, produced oscillation-free solutions near the discontinuous top corners, demonstrating the effectiveness of the SAT technique.
• Backward-Facing Step Flow: The scheme successfully captured complex flow features, such as primary and secondary recirculation vortices. The results matched reference data even when using a significantly coarser mesh than previous studies, highlighting the method's efficiency.

5. Significance

This research provides a robust, high-order, and mathematically stable framework for incompressible flow simulations. By combining the accuracy of CGFEM with the stability of SBP-SAT, the authors have developed a method that is both computationally efficient and capable of handling the most difficult aspects of fluid mechanics—namely, the pressure-velocity coupling and discontinuous boundary data—without sacrificing physical integrity. This has broad implications for engineering applications ranging from aerospace to biomedical design.