A fluctuating lattice Boltzmann formulation based on orthogonal central moments

This paper presents a fluctuating lattice Boltzmann formulation based on orthogonal central moments that introduces stochastic forcing directly in the central moment space to exactly satisfy the fluctuation-dissipation theorem, ensuring thermal equilibrium, Galilean invariance, and enhanced numerical stability even in the over-relaxation regime.

The Gist

Imagine you are trying to simulate a cup of hot coffee on a computer. If you zoom in far enough, you don't see a smooth, flowing liquid. Instead, you see billions of tiny, jittery molecules bumping into each other, dancing to the rhythm of heat. This jitteriness is called thermal fluctuation.

For a long time, computer simulations of fluids (like water or air) were like a choreographed dance where everyone moved in perfect, predictable lines. They were great for big, smooth flows, but they missed the tiny, chaotic "jitters" that happen in the real world, especially at small scales (like inside a cell or a microchip).

This paper introduces a new way to simulate these jitters using a method called Lattice Boltzmann. Think of Lattice Boltzmann as a grid of tiny buckets, where each bucket holds a crowd of virtual particles. The particles move from bucket to bucket and bounce off each other (collide).

Here is the simple breakdown of what the authors did and why it matters:

1. The Problem: The "Rigid" Dance

Previous methods tried to add these random jitters by just throwing random noise into the system. But this was like trying to fix a wobbly table by just shaking it harder.

• The Issue: When you add randomness, you have to be very careful not to break the laws of physics (like conservation of energy).
• The Old Way: Imagine a group of dancers where everyone is holding hands in a giant, tangled knot. If one person stumbles (a fluctuation), it pulls on everyone else in a messy, unpredictable way. This made the simulation unstable, especially when the fluid was very thin (low viscosity) or moving fast.

2. The Solution: The "Solo" Dancers

The authors realized that to handle these jitters correctly, they needed to change the "dance floor" itself. Instead of tracking the particles directly, they decided to track Central Moments.

The Analogy: The Orchestra

• Old Method (Raw Moments): Imagine an orchestra where every musician is playing a different instrument, but they are all slightly out of sync. If the violinist gets nervous (a fluctuation), the drummer gets nervous too, and the whole band sounds chaotic. It's hard to tell who is doing what.
• New Method (Orthogonal Central Moments): The authors reorganized the orchestra. They grouped the musicians so that the Violin section, the Drum section, and the Flute section are completely independent.
  • If the Violins get jittery, the Drums don't care.
  • If the Drums get loud, the Flutes stay calm.
  • Because they are orthogonal (independent), the computer can calculate exactly how much "noise" each section needs without messing up the others.

3. Why "Central" Moments?

Usually, simulations measure movement relative to a fixed point (like a wall). But fluids move!

• The Analogy: Imagine you are on a train. If you measure how fast the coffee in your cup is sloshing relative to the station outside, the numbers are huge and confusing because the train is moving.
• The Fix: The authors measure the sloshing relative to the train itself (the local fluid velocity). This is what "Central Moments" means. It removes the "train movement" so they can see the true, chaotic jitter of the coffee. This makes the simulation much more stable and accurate, even when the fluid is moving fast.

4. The Magic Result: The "Perfect Balance"

The most important part of this paper is that they proved their new method perfectly follows the Fluctuation-Dissipation Theorem.

• Translation: In physics, "dissipation" is friction (energy loss), and "fluctuation" is the jitter (energy gain). Nature always balances these two. If you have friction, you must have jitter to keep the temperature right.
• The Achievement: Their new method automatically balances friction and jitter perfectly, like a thermostat that never fails. Even better, it works in "over-relaxed" regimes (where the fluid is very slippery and hard to simulate). The old methods would crash and give errors in these situations, but this new method stays stable.

5. What Did They Test?

They ran a series of tests to prove it works:

• The "Zero Noise" Test: They turned off the jitters, and the simulation behaved exactly like a standard, smooth fluid. (Good!)
• The "Heat" Test: They added heat. The fluid started jiggling exactly as much as physics predicts.
• The "Stress Test": They tried to simulate very slippery fluids (low viscosity). The old methods crashed, but their new method kept dancing perfectly.
• The "Shape" Test: They simulated fluids in weirdly shaped boxes (not just squares). The jitters remained perfectly balanced in all directions, proving the method doesn't have a "favorite" direction.

The Bottom Line

The authors have built a super-stable, physics-perfect engine for simulating fluids that jitter.

By organizing the simulation into independent, "solo" groups (orthogonal central moments) and measuring movement relative to the flow itself, they created a system that:

1. Never breaks the laws of thermodynamics.
2. Doesn't crash when the fluid is very thin or fast.
3. Is ready to be used for complex real-world problems, like blood flow in tiny capillaries or the behavior of nanomaterials.

It's like upgrading from a shaky, hand-cranked radio to a high-definition digital stream that never skips a beat, no matter how chaotic the music gets.

Technical Summary

1. Problem Statement

Thermal fluctuations are intrinsic to fluid dynamics at mesoscopic and microscopic scales, influencing transport coefficients, equilibrium properties, and spatio-temporal correlations. While the Lattice Boltzmann Method (LBM) is a powerful tool for mesoscopic fluid simulation, incorporating thermal fluctuations (Fluctuating LBM or FLBM) presents significant challenges:

• Thermodynamic Consistency: Numerical schemes must satisfy the Fluctuation-Dissipation Theorem (FDT) to ensure correct equilibrium statistics (e.g., equipartition of kinetic energy).
• Galilean Invariance: The formulation must remain invariant under Galilean transformations, which is difficult when noise covariance depends on local velocity.
• Numerical Stability: Existing fluctuating formulations, particularly those based on single-relaxation-time (BGK) or raw-moment Multi-Relaxation-Time (MRT) models, often suffer from instability in low-viscosity (high Reynolds number) regimes or the over-relaxation limit ($\tau \to 0.5$).
• Mode Coupling: In non-orthogonal bases, equilibrium covariance matrices are non-diagonal, requiring complex correlated noise to satisfy FDT. This obscures physical interpretation and can lead to spurious couplings between hydrodynamic and kinetic modes.

2. Methodology

The authors propose a systematic framework for constructing FLBM schemes based on Orthogonal Central Moments (CMs). The core methodology involves:

• Central Moment Space: The collision step is formulated in a reference frame moving with the local fluid velocity. This inherently improves Galilean invariance and suppresses spurious velocity-dependent couplings found in raw-moment formulations.
• Orthogonal Basis Construction: The authors employ a specific orthogonal basis for the central moments (e.g., for D2Q9 and D3Q27). This ensures that the equilibrium covariance matrix of the kinetic modes is strictly diagonal.
• Hermite-Consistent Equilibrium: The equilibrium distribution is constructed via a Hermite expansion up to the maximum order supported by the lattice (e.g., 4th order for D2Q9, 6th for D3Q27).
  • Key Insight: With this specific pairing of orthogonal basis and high-order equilibrium, all non-conserved central moments vanish identically at equilibrium ($k^{eq}_j = 0$ for $j > 0$). This creates a clean separation between the deterministic equilibrium structure (confined to the conserved subspace) and stochastic fluctuations.
• Stochastic Forcing:
  • Noise is introduced directly in the central moment space.
  • Because the basis is orthogonal and the equilibrium covariance is diagonal, each non-conserved mode can be treated as an independent discrete Ornstein–Uhlenbeck process.
  • The noise amplitude is derived mode-by-mode to satisfy the FDT exactly at the lattice level, without requiring correlated noise or velocity-dependent covariance matrices.
• Lattice Agnosticism: The framework is demonstrated on D2Q9, D3Q27, and D3Q19 lattices, proving that the approach relies on the structural pairing of the basis and equilibrium rather than a specific velocity set.

3. Key Contributions

1. Systematic Derivation: A general procedure for incorporating thermal fluctuations into CM-based LBM, ensuring exact compliance with the FDT and conservation laws.
2. Diagonal Covariance & Independence: The demonstration that orthogonal central moments are necessary (not just advantageous) to achieve a diagonal equilibrium covariance matrix. This allows for statistically independent stochastic forcing, preserving Galilean invariance and physical interpretability.
3. Enhanced Stability: The formulation remains stable and well-posed in the over-relaxation regime (near $\tau = 0.5$), a critical limitation of traditional fluctuating BGK schemes which often suffer from numerical breakdown in this region.
4. Explicit Implementations: The paper provides explicit collision operators and transformation matrices for D2Q9 and D3Q27, along with a D3Q19 example, and releases open-source MATLAB and C++ (Kokkos) codes for reproducibility.

4. Results

The authors validated the formulation through seven systematic numerical tests:

• Deterministic Regression: Confirmed that with zero noise ($k_B T = 0$), the scheme reduces exactly to the deterministic CM-LBM, preserving mass/momentum conservation and recovering second-order accuracy (verified via Taylor-Green vortex decay).
• Equipartition of Kinetic Energy: Verified that the system relaxes to a state where $\langle u^2_\alpha \rangle = k_B T / \rho_0$. The scheme showed high accuracy with biases $< 0.3\%$ and no systematic drift.
• Scaling Laws:
  • Thermal Energy: Velocity variance scaled linearly with $k_B T$.
  • Density: Velocity variance scaled inversely with $\rho_0$.
• Relaxation Time Robustness (Crucial Finding):
  • The proposed method maintained accurate equipartition across a wide range of relaxation times ($\tau \in [0.5, 100]$).
  • In contrast, the standard fluctuating BGK scheme exhibited a sharp numerical breakdown and unbounded fluctuations as $\tau \to 0.5$. The CM-based method remained stable even in the immediate vicinity of the stability limit.
• Isotropy in Anisotropic Domains: Tests on 3D domains with varying aspect ratios confirmed that velocity fluctuations remained isotropic and homogeneous, with no directional bias introduced by the geometry or lattice discretization.
• Distribution Statistics: Probability density functions of velocity components collapsed onto a unit-variance Gaussian, confirming correct thermalization of hydrodynamic modes.

5. Significance

This work establishes Orthogonal Central Moments as the natural and robust framework for fluctuating hydrodynamics in LBM.

• Theoretical Clarity: It resolves the issue of mode coupling in fluctuating schemes by ensuring that the deterministic equilibrium structure does not contaminate the kinetic subspace, allowing for a transparent application of the FDT.
• Practical Utility: The method enables stable simulations of thermally fluctuating flows in regimes previously inaccessible to standard FLBM (specifically low-viscosity/high-Reynolds number flows where $\tau$ is close to 0.5).
• Future Applications: The framework provides a solid foundation for extending fluctuating hydrodynamics to complex multiphase, multi-component, and fully thermal flows, where numerical stability and thermodynamic consistency are paramount.

In summary, the paper demonstrates that by aligning the kinetic mode structure, dissipation, and stochastic forcing through an orthogonal central-moment basis, one can achieve a fluctuating LBM that is both thermodynamically exact and numerically robust.