Rayleigh-Taylor Unstable Flames: Thin and Thick

This paper investigates Rayleigh-Taylor unstable flames through large-scale Boussinesq model simulations and direct measurements, revealing that while self-generated turbulence can thicken these flames, their internal structure remains distinct from classical turbulent flames, thereby highlighting the need for specialized subgrid models in practical applications.

The Gist

Imagine a flame not as a simple, steady candle flickering in the wind, but as a chaotic, living thing fighting a war on two fronts. This is the story of Rayleigh-Taylor (RT) unstable flames, a phenomenon that happens when a heavy fluid (like oil) sits on top of a lighter one (like water) and gravity tries to swap their places. If you add fire to the mix, things get incredibly complicated.

This paper by E.P. Hicks investigates what happens when these flames try to burn in such a chaotic environment. The author asks a simple but profound question: Can we understand these wild, self-made chaotic flames using the same rules we use for normal, turbulent flames?

Here is the breakdown of the research using everyday analogies.

1. The Setup: A Heavy Blanket on a Light Pillow

Imagine you have a heavy, wet blanket (the fuel) sitting on top of a light, fluffy pillow (the ash). Gravity wants the heavy blanket to fall through the pillow. As it tries to sink, it creates "bubbles" of pillow rising up and "spikes" of blanket falling down. This is the Rayleigh-Taylor instability.

Now, imagine the blanket is on fire. As the fire burns, it creates its own wind and turbulence. The paper explores the tug-of-war between:

• Gravity: Trying to stretch the flame into long, thin spikes and bubbles.
• The Fire's Own Wind: The turbulence created by the flame itself, which tries to scramble and thicken the flame.
• The Burning: Trying to consume the fuel.

2. The Big Surprise: The "Thin Front, Fat Back" Flame

For a long time, scientists thought these flames were just "thin" and fast, like a stretched-out rubber band. But Hicks discovered something new: Sometimes, these flames get "fat."

Think of a flame like a sandwich.

• Normal Turbulent Flames: When wind blows on a normal flame, it wrinkles the front (the bread facing the wind), making the whole sandwich thicker from the front.
• RT Unstable Flames: These flames behave differently. The "front" of the flame (the side facing the heavy fuel) gets stretched so thin by gravity that it becomes razor-sharp. However, the "back" of the flame (the side facing the ash) gets blown apart and thickened by its own turbulence.

The Analogy: Imagine a runner sprinting.

• A normal turbulent flame is like a runner wearing a heavy, wind-blown cape that makes them look wider as they run.
• An RT unstable flame is like a runner who is being pulled forward by a giant elastic band (gravity) that stretches their body thin, while a chaotic swarm of bees (turbulence) is buzzing around their legs, making their lower half look messy and thick. The result is a weird, lopsided shape: thin in front, thick in back.

3. The "Thickening" Threshold

The paper found that these flames only get "fat" (thickened) under specific conditions. It's like a recipe:

• If the flame is naturally very thin (high reaction speed) or the fluid is "thick" (high viscosity/Prandtl number), the flame stays thin and fast.
• But if the flame is naturally a bit "fluffier" (lower reaction speed) and the fluid is "runny" (low Prandtl number), the self-generated turbulence wins. The flame gets thickened, but it slows down.

The Takeaway: You can't just assume these flames act like normal turbulent flames. They have a unique "personality" that depends on how "runny" the fluid is and how "fast" the chemical reaction is.

4. Why Old Maps Don't Work

Scientists have "maps" (called regime diagrams) that predict how flames behave in turbulence. These maps are like weather forecasts for fire.

• The Problem: When the author plotted these RT flames on the old maps, they didn't fit. The old maps assumed the flame would thicken from the front, but these flames thicken from the back. They also didn't account for the "runniness" of the fluid (Prandtl number).
• The Solution: The author drew a new map. This new map includes the specific properties of the flame and the fluid, allowing scientists to predict whether a flame will be "thin and fast" or "thick and slow."

5. Why Does This Matter? (The Real World)

This isn't just abstract math; it affects real-world technology and even the universe:

• Airplane Engines: New engines use centrifugal force to mix fuel and air. If engineers use the wrong model (assuming the flame is thin when it's actually thick), the engine might be inefficient or dangerous.
• New Fuels: As we switch to fuels like ammonia or new refrigerants, we need to know how they burn in storage tanks. If a flame gets thickened by its own turbulence, it might burn slower or faster than expected, changing the explosion risk.
• Supernovae: When a star explodes (Type Ia supernova), it's essentially a giant RT unstable flame. Understanding if the flame gets "thick" helps astrophysicists understand how these stars explode and why they are so bright.

The Bottom Line

The paper concludes that RT unstable flames are unique creatures. They are not just "turbulent flames" with a different name. They are a three-way battle between gravity stretching them thin, turbulence scrambling them thick, and the fire trying to burn.

If you want to model them, you can't use the old rulebook. You need a new one that understands that these flames can be thin in the front and fat in the back, and that their behavior depends heavily on how "runny" the fluid is. It's a reminder that in nature, fire doesn't always behave the way we expect!

Technical Summary

Technical Summary: Rayleigh-Taylor Unstable Flames: Thin and Thick

1. Problem Statement
Rayleigh-Taylor (RT) unstable flames occur at the interface between heavy fuel and light ash layers, a phenomenon critical to diverse systems ranging from aviation turbine engines and alternative fuel storage to Type Ia supernovae. A fundamental challenge in modeling these systems is determining whether RT unstable flames can be described using existing subgrid models derived from turbulent combustion theory.

Traditional turbulent combustion assumes a flame consuming turbulent fuel, where turbulence and burning compete in a two-way interaction. In contrast, RT unstable flames generate their own turbulence via baroclinic vorticity as the instability stretches the flame front. Previous work (Hicks 2015, 2019) suggested RT flames are "thin" (thinner than laminar) and distinct from turbulent flames because RT stretching overwhelms turbulent thickening. However, other studies suggested unconfined RT mixing zones might thicken over time. The core problem is reconciling these views: Can RT flames be thickened by their self-generated turbulence, and if so, does their structure resemble classical turbulent combustion regimes?

2. Methodology
The author conducted a large-scale parameter study using Direct Numerical Simulations (DNS) and nearly-DNS of Boussinesq model flames. The study utilized the spectral element code Nek5000 to resolve the full turbulent cascade.

• Governing Equations: The simulations solved the Boussinesq equations with a model reaction term, isolating RT instability effects while excluding shocks and Landau-Darrieus instabilities.
• Control Parameters:
  • Gravity ($G$) and Domain Width ($L$): Fixed at $G=32, L=32$ (based on previous work showing significant deviation from standard RT speed models).
  • Prandtl Number ($Pr$): Varied from 0.5 to 8. This is critical as terrestrial applications have $Pr \approx 1$, while Type Ia supernovae have extremely low $Pr \approx 10^{-5}$.
  • Reaction Model ($p$-flames): A novel family of traveling wave solutions was developed using a parameterized reaction rate $R(T) = (p+1)T^{p+1}(1-T^p)$. By varying $p$, the author created flames with the same laminar flame speed ($s_0=1$) but different physical laminar flame widths ($\delta$). Lower $p$ yields thicker flames; higher $p$ yields thinner flames.
• Measurement Strategy:
  • Flame Width: Measured between temperature contours $T=0.1$ and $T=0.9$, normalized by the laminar width ($\delta_N$).
  • Internal Structure: The flame was sliced into isovolumes to measure local thickness and surface area correlations of temperature contours.
  • Regime Analysis: The study reformulated standard turbulent combustion non-dimensional numbers (Reynolds $Re$, Damköhler $Da$, Karlovitz $Ka$) to account for $p$ and $Pr$. Specifically, new "extended" Karlovitz numbers ($Ka_x, Ka_w, Ka_c$) were derived to test if self-generated turbulence drives thickening.

3. Key Contributions

• Discovery of "Thick+Slow" Regime: The paper identifies a regime where RT unstable flames become thicker than laminar and propagate slower than predicted by standard RT speed models. This occurs at low $p$ (thick laminar flames) and low $Pr$.
• Novel Internal Flame Structure: Unlike turbulent flames which thicken from the preheat (front) layer, RT unstable flames exhibit a "thin-front, thick-back" structure. The RT instability thins the front, while self-generated turbulence thickens the rear (ash side).
• Reformulation of Regime Diagrams: The study demonstrates that classical and experimental turbulent combustion regime diagrams fail to classify RT flames because they ignore $Pr$ and the specific laminar flame width ($p$). A new regime diagram based on an extended Karlovitz number ($Ka_x$) is proposed.
• Resolution of Contradictory Literature: The work reconciles previous findings: RT flames are "thin" in regimes where RT stretching dominates (high $p$, high $Pr$) but can become "thick" when self-generated turbulence dominates (low $p$, low $Pr$).

4. Key Results

• Thickening Mechanism: The transition from "thin+fast" to "thick+slow" is driven by self-generated turbulence. The data collapses well when plotted against the extended Karlovitz number ($Ka_x$), which accounts for $Pr$ and the physical flame width.
  • Thresholds: Flames remain thin for $Ka_x \lesssim 60$. A transition region exists for $60 \lesssim Ka_x \lesssim 80$. Flames become thicker than laminar for$Ka_x \gtrsim 80$.
• Structural Asymmetry:
  • Thin Flames: Temperature contours evolve coherently as a single unit stretched by RT instability.
  • Thick Flames: The rear contours ($T=0.7, 0.9$) desynchronize from the front contours due to turbulent mixing, while the front remains thin.
• Regime Diagram Failure: Standard diagrams (Peters 2000, Skiba et al. 2018) incorrectly place these flames in the "thin reaction zones" regime. The new diagram, using axes of $Pr$ and a modified energy transfer rate, correctly separates thin, transitional, and thick regimes.
• Distributed Burning: The study argues that a full transition to distributed burning (where turbulence completely dominates propagation) is unlikely. The RT instability is required to generate the turbulence; if the flame thickens too much, turbulence production is quenched, and the RT instability thins the flame again, creating a cyclical dynamic. However, at very high $Ka_x$ and $G$, a "thin-front, very thick brush" structure resembling distributed burning may exist.

5. Significance and Implications

• Modeling Aviation and Fuel Safety: For terrestrial applications (e.g., ammonia or R-32 refrigerants), using flamelet models based on turbulent combustion assumptions may be inaccurate if the flame enters the "thick+slow" regime. Models must account for $Pr$ and the specific reaction kinetics ($p$).
• Type Ia Supernovae: The findings suggest that as white dwarf flames propagate outward and density decreases (effectively lowering $p$ and $Pr$), they may thicken due to self-generated turbulence. This could create conditions favorable for Deflagration-to-Detonation Transition (DDT) via mechanisms like the Zel'dovich gradient, even without pre-existing turbulence.
• Numerical Best Practices: The paper warns against using artificially thickened reaction models (like the standard KPP model with large $\delta$) in simulations. If $Pr$ is low, these models may produce artificially slow, overly thickened flames that do not reflect physical reality. Accurate reaction models and careful selection of $Pr$ are essential.
• Theoretical Advancement: The work establishes that RT unstable flames are governed by a three-way competition (RT instability, self-generated turbulence, and burning) rather than the two-way competition of standard turbulent combustion, necessitating new theoretical frameworks and subgrid models.