Numerical model for pellet rocket acceleration in PELOTON

This paper presents a validated numerical model within the PELOTON code that simulates pellet rocket acceleration in thermonuclear fusion devices by accounting for ablation cloud asymmetry and plasma gradients, demonstrating consistency with JET experimental trajectories and revealing reduced deviation for deuterium-neon composite pellets.

The Gist

Imagine a tiny, super-cold snowball (a "pellet") being shot into a giant, swirling oven of super-hot gas (plasma) inside a fusion reactor. This isn't just a simple collision; it's a high-speed dance where the snowball tries to survive while the oven tries to melt it.

This paper describes a new computer simulation called PELOTON that acts like a high-definition movie director for this dance. Its main job is to figure out why these snowballs don't just melt in a straight line, but instead get pushed sideways, accelerating like a rocket.

Here is the breakdown of what the paper found, using simple analogies:

1. The "Rocket" Effect: Why the Snowball Moves Sideways

Usually, if you blow air on a balloon, it pushes it away. But here, the "air" is actually a stream of invisible, super-fast electrons from the hot plasma.

• The Setup: As the snowball enters the oven, it starts melting, creating a thick, cold cloud of gas around it.
• The Twist: The oven has a magnetic field that is stronger on one side (the "High-Field Side" or HFS) and weaker on the other (the "Low-Field Side" or LFS).
• The Analogy: Imagine the snowball is a person standing in a crowd. On one side (HFS), the crowd is dense and chaotic, making it hard for the "heat" (electrons) to reach the person. On the other side (LFS), the crowd is thinner, so the heat hits the person harder.
• The Result: Because the heat hits the LFS harder, the gas cloud on that side gets hotter and pushes back harder. This creates a pressure difference. The snowball gets squeezed from the hot side and pushed toward the cool side. It's like a rocket being pushed by the exhaust, but in reverse: the pressure behind it (on the LFS) is higher than the pressure in front of it, shoving it sideways.

2. The Computer Model: PELOTON

The authors built a 3D simulation to track this. Think of PELOTON as a super-accurate weather forecast for the inside of the reactor.

• It tracks the snowball as it melts.
• It calculates how the cold gas cloud forms and moves.
• It accounts for the fact that the cloud isn't uniform; it's "charged up" differently on different sides, which changes how the hot electrons hit it.
• They tested this model against real experiments at JET (a famous fusion lab in the UK) and found that their computer predictions matched the real-life snowball paths almost perfectly.

3. The "Shattered" Snowball (SPI)

Sometimes, instead of one big snowball, they shoot in a "shattered pellet" (SPI). Imagine throwing a handful of ice chips instead of one block.

• The Cloud Overlap: If two ice chips are close together, their gas clouds can bump into each other. The paper found that if they are side-by-side, the bottom one gets pushed harder. If they are lined up one behind the other on the same magnetic path, they actually pull toward each other because they block the heat from hitting the back of the front one.
• The Neon Mix: They tried adding a tiny bit of neon gas (like a different flavor of ice) to the snowball. This made the gas cloud cooler and slower. While the "rocket push" still happened, it was weaker. Interestingly, in real experiments, this didn't seem to change the path much, likely because the neon caused other big changes in the plasma that masked the effect.

4. The "Scaling Law": A Recipe for Prediction

The team analyzed hundreds of simulations to create a simple "recipe" (a scaling law).

• The Recipe: The strength of the sideways push depends mostly on how hot the plasma is and how dense it is.
• The Surprise: The size of the snowball (the pellet radius) barely matters! A tiny chip and a big chunk get pushed with roughly the same force per unit of mass. This is a huge simplification for scientists trying to predict how these pellets will behave.

5. What This Means for the Future

The paper concludes that this model is ready to be used for the next giant fusion machine, ITER.

• They plan to use this "rocket physics" to predict how shattered pellets will behave in ITER's massive plasma.
• They want to refine the model to include how the plasma particles spread out (diffusion) to make the predictions even more accurate.

In a nutshell: The paper explains that when cold pellets melt in a fusion reactor, they get pushed sideways by an invisible "wind" of heat that hits them unevenly. The authors built a computer model that predicts this push perfectly, showing that the size of the pellet doesn't matter much, but the temperature and density of the plasma do. This helps scientists understand how to safely inject fuel into future fusion power plants.

Technical Summary

Technical Summary: Numerical Model for Pellet Rocket Acceleration in PELOTON

Problem Statement
The injection of cryogenic pellets or shattered pellet injection (SPI) fragments into thermonuclear fusion devices is subject to "rocket acceleration," a phenomenon where fragments deviate from their intended trajectories. This acceleration arises from the asymmetric heating of the ablation cloud surrounding the pellet. As hot plasma electrons stream along magnetic field lines, the ablation cloud ionizes and drifts across field lines due to magnetic field curvature (grad-B drift). This drift creates an asymmetry: the high-field-side (HFS) of the cloud provides less electrostatic shielding and shorter path lengths for incoming electrons compared to the low-field-side (LFS). Consequently, the HFS experiences higher electron energy deposition and pressure, accelerating the pellet toward the LFS. While observed experimentally, existing models often treat the driving pressure difference as an adjustable parameter or rely on semi-analytical approximations that may lack realism in higher-order terms.

Methodology
The authors developed a direct numerical simulation model for pellet rocket acceleration implemented within PELOTON, a 3D Lagrangian particle code. The methodology integrates several key physical components:

1. Non-Uniform Cloud Charging: Unlike previous versions of PELOTON that assumed a constant electrostatic shielding potential, this model solves for a non-uniform shielding potential ($\Phi$) across the ablation cloud. This is achieved by solving a nonlinear equation for the shielding potential along magnetic field lines, accounting for the albedo effect and collisional backscattering. This variation leads to different effective electron densities ($n_{eff}$) and heat depositions on the HFS versus the LFS.
2. MHD and Ablation Physics: The evolution of the cold, partially ionized ablation cloud is governed by low Magnetic Reynolds number magnetohydrodynamic (MHD) equations in Lagrangian coordinates. The model includes:
   • Surface ablation boundary conditions where heat flux drives sublimation.
   • Grad-B drift dynamics, modeled via a simplified transverse drift acceleration equation driven by the pressure difference integral.
   • An advanced electron heat deposition model based on solutions to the 3-D linearized Fokker-Planck kinetic equation, incorporating opacity and shielding effects.
   • A multi-species equation of state solver for deuterium-neon mixtures, utilizing Saha equations to determine ionization fractions, temperature, and sound speed.
3. Plasma Cooling Module: A new module was developed to provide evolving background plasma states during SPI simulations. This module discretizes the plasma into flux volumes, distributes ablated material, and accounts for ionization losses, Ohmic heating, and ion-electron energy exchange. It was validated against JOREK and INDEX simulations.
4. Validation and Simulation: The model was applied to simulate SPI experiments in the JET tokamak. Simulations included pure deuterium and deuterium-neon (0.5% neon) mixtures. The code tracked fragment trajectories, ablation rates, and rocket acceleration forces, comparing results against experimental data (specifically JET #96874).

Key Contributions and Results

• Direct Numerical Simulation of Rocket Acceleration: The paper presents a first-principles numerical model that calculates the rocket acceleration force directly from the pressure difference ($P_{HFS} - P_{LFS}$) generated by asymmetric electron heating, rather than relying on empirical fitting parameters.
• Scaling Law for JET: Based on 20 simulation data points covering JET-relevant plasma parameters ($T_e \in [450, 2000]$ eV, $n_e \in [6\times10^{19}, 3\times10^{20}]$ m$^{-3}$), the authors derived a scaling law for the pressure difference driving the acceleration:
  $$\Delta P [\text{bar}] = 1.0 \left( \frac{T_e [\text{keV}]}{2} \right)^{1.26} \left( \frac{n_e [1/m^3]}{10^{20}} \right)^{0.4}$$
  The fit shows a mean absolute relative error of approximately 13%. The study found that $\Delta P$ is practically independent of the pellet radius within the tested range.
• Validation with JET Experiments: PELOTON simulations of deuterium SPI trajectories, incorporating the calculated rocket acceleration, showed consistency with experimentally measured trajectories in JET, particularly regarding the deviation toward the low-field side.
• Neon Mixture Effects: Simulations of deuterium-neon pellets (0.5% neon) revealed that the neon component reduces the ablation cloud temperature, longitudinal expansion velocity, and grad-B drift velocity. While this results in a smaller but still measurable rocket acceleration compared to pure deuterium, the authors note that in actual JET experiments, the rocket effect on neon trajectories is likely masked by global MHD events triggered by neon radiation.
• Fragment Interactions: The study quantified the impact of cloud overlap. Fragments separated transversely by 2–4 cm showed modified acceleration due to pressure redistribution. Fragments aligned along the same magnetic field line experienced strong mutual attraction due to partial screening of the electron heat flux, though the authors note the probability of such alignment in real SPI plumes is low.
• Plasma Gradients: The model demonstrated that strong local plasma gradients can amplify rocket acceleration by a factor of ~3, but these effects are transient, acting only during short intervals as fragments cross deposition boundaries.

Significance and Claims
The paper claims to provide a comprehensive numerical framework that captures the essential physics of pellet rocket acceleration, specifically the role of non-uniform electrostatic charging and grad-B drift. The authors emphasize that their model moves beyond semi-empirical approaches by resolving the pressure asymmetry directly.

The significance of the work lies in its ability to:

1. Reproduce experimental SPI trajectories in JET with quantitative fidelity.
2. Provide a predictive scaling law for the pressure difference driving rocket acceleration, which can be utilized in other codes.
3. Demonstrate that while rocket acceleration is a significant physical mechanism, its observable impact on trajectories can be modulated by pellet composition (e.g., neon addition) and background plasma gradients.

The authors remain modest regarding the immediate application to ITER, stating that future work is required to develop specific scaling laws for ITER plasmas and to refine the cooling module to include plasma diffusion and MHD effects. They explicitly state that the reported scaling law is specific to the JET parameter space and magnetic geometry.