Pith. sign in

REVIEW 2 major objections 4 minor 32 references

Runaway electron induced explosions of graphite; modeling versus controlled DIII-D experiments

T0 review · 2 major / 4 minor · reviewed 2026-07-11 · grok-4.5

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read An extended thermomechanical model of graphite under runaway-electron impact now reproduces both the onset of failure and the subsequent fragmentation and debris expulsion seen in controlled DIII-D experiments.

desk verdict Solid, first quantitative extension of RE-graphite damage modeling past Rankine onset into full fragmentation and debris; matches two DIII-D shots within large experimental bars. read the letter →

arxiv 2607.03914 v1 pith:LWDH7ON2 submitted 2026-07-04 physics.plasm-ph cond-mat.mtrl-sci

T. Rizzi , A. Kavroulakis , S. Ratynskaia , P. Tolias , M. Tatarakis , E. Kaselouris , V. Dimitriou , E. Hollmann
show 5 more authors
F. Brochard A. Kulachenko M. Beidler C. Lasnier R. A. Pitts
This is my paper · ORCID
classification physics.plasm-phcond-mat.mtrl-sci
keywords runawayelectroninduceddamagePFCfragmentationexplosionsfiniteelementanalysissmoothed-particlehydrodynamicsJohnson-HolmquistmodelgraphiteDIII-D
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Runaway electrons that strike plasma-facing graphite can drive explosive material failure and throw debris. Earlier models stopped at the first appearance of cracks under linear elasticity. This paper adds plasticity through the Johnson-Holmquist constitutive model, switches to an effective-plastic-strain failure criterion, and converts failed finite elements into smoothed-particle-hydrodynamics particles so that fragments can keep moving. The same workflow is then compared with two instrumented DIII-D shots that deposited roughly 10 kJ in about 1 ms. Predicted eroded volumes, depths, surface areas and debris-speed distributions match the measured values closely enough that the authors treat the model as validated for graphite. The result is presented as a necessary intermediate step before the same physics can be applied to tungsten, the material chosen for ITER and most future reactors.

What carries the argument

Adaptive FEM-to-SPH conversion driven by the Johnson-Holmquist constitutive model and an effective-plastic-strain failure criterion of 0.025: once an element exceeds the plastic-strain limit it is replaced by an SPH particle that inherits its mass, velocity and temperature and can thereafter fly free, allowing the simulation to continue past the onset of failure into fragmentation and debris flight.

What would settle it

A new controlled DIII-D or similar shot whose independently measured energy and pitch distributions produce a Geant4 heat map that, when run through the same JH-2 + FEM-SPH pipeline, yields an eroded volume or debris-speed histogram that lies well outside the experimental uncertainties reported for the two benchmark cases.

Watch

Extended reading notes

Core claim

When the Johnson-Holmquist model, an effective-plastic-strain failure threshold of 0.025 and adaptive FEM-to-SPH conversion are combined with Geant4 energy-deposition maps, the resulting thermomechanical simulation reproduces the measured eroded volumes (approximately 130 mm^{3} versus 140 mm^{3} for the 2023 shot and 60 mm^{3} versus 30 mm^{3} for the 2024 shot), the depths of the damage pits and the bulk of the observed debris-speed distributions for both controlled DIII-D graphite explosions.

Load-bearing premise

The runaway-electron energy and pitch angle that are fed into the energy-deposition calculation are treated as known external inputs even though they were themselves chosen to match earlier onset-of-failure or detector data.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

2 major / 4 minor

Summary. The manuscript extends prior one-way coupled linear thermoelasticity + Rankine modeling of graphite under runaway-electron (RE) impact to the nonlinear fragmentation and debris-expulsion stage. It replaces the elastic constitutive law with the Johnson-Holmquist 2 (JH-2) model, adopts an effective-plastic-strain (EPS) failure threshold (ε_fail = 0.025), and couples finite-element analysis to adaptive FEM-to-SPH conversion in LS-DYNA. Energy-deposition maps are obtained from Geant4 using mono-energetic, mono-pitch RE parameters previously reconstructed for two DIII-D shots (Case 1: 3 MeV/25°; Case 2: 4 MeV/10°). The model is shown to reproduce the measured eroded volumes (~130 mm^{3} vs 140 mm^{3}; ~60 mm^{3} vs 30 mm^{3}), depths, surface areas and debris speed distributions of the two controlled explosions, thereby providing a complete thermomechanical description of RE-driven graphite explosions.

Significance. If the reported agreement holds under the stated inputs, the work supplies the first quantitative continuum description of the full RE-induced explosion sequence (onset → fragmentation → debris kinematics) for a brittle, sublimating PFC material. The hybrid FEM-SPH framework and the explicit comparison against instrumented DIII-D data constitute a concrete, falsifiable step toward the stated goal of tungsten modeling. Strengths include the transparent use of external JH-2 parameters, the side-by-side Rankine/JH-2 onset-time consistency, and the open discussion of residual volume and high-speed-tail discrepancies in Sec. 6.

major comments (2)
  1. Sec. 3.1 and Table I: the mono-energetic/mono-pitch RE parameters fed to Geant4 are taken from the discrete set already shown to match either Rankine onset times or TLD maps in prior work by overlapping authors. While the subsequent JH-2/SPH evolution is independent once the deposition map is fixed, the manuscript should quantify (or at least bound) how modest variations around the chosen (E, heta) pairs affect the predicted eroded volume and debris spectrum; without that sensitivity the claim of successful benchmarking remains conditional on the prior reconstruction.
  2. Sec. 5.1 and Table I: Case 2 yields a factor-of-two volume overestimate (~60 mm^{3} simulated vs ~30 mm^{3} measured). The text attributes the discrepancy to a larger wetted area and weaker gradients, yet does not demonstrate that the same deposition map that matched the TLD data also produces the observed surface-area and depth when the EPS criterion is applied. A short parametric check on peak energy density or wetted-area scaling would clarify whether the overestimate is within experimental uncertainty or signals a residual model bias.
minor comments (4)
  1. Fig. 2 caption and panels (g–i): the color bar is saturated at 0.01 while ε_fail = 0.025; a brief note that the saturation is only for visualization would avoid confusion.
  2. Sec. 3.3: the statement that thermal expansion is introduced separately because JH-2 does not couple temperature to volumetric strain is correct, but the precise algorithm (subtract ε_T from total strain at each step) should be stated more explicitly for reproducibility.
  3. Fig. 5: error bars or uncertainty bands on the experimental speed histograms (especially Case 2) would make the comparison more quantitative; the text already notes ~20 % association errors and lower video quality.
  4. Sec. 6: the discussion of SPH birth-location bias is useful; a one-sentence remark on whether a pure mesh-free or discrete-element alternative is planned for tungsten would strengthen the outlook.

Circularity Check

2 steps flagged · score 3.0 of 10

Modest circularity limited to selection of mono-energetic/mono-pitch RE parameters already tuned to prior onset-of-failure or TLD matches by overlapping authors; JH-2/EPS/FEM-SPH fragmentation predictions of volumes and debris remain independent of that choice.

  1. fitted input called prediction [Sec. 3.1 Energy deposition; Table I]
    "In the modeling of Case 1, the 3 MeV & 25 ◦ pitch scenario will be followed, since it leads to accurate predictions for the onset of brittle failure [8]. ... In the modeling of Case 2, the 4 MeV & 10◦ pitch scenario will be followed, since it provides the best match with the TLD data [6]."

    The mono-energetic mono-pitch RE parameters that drive Geant4 energy deposition (and therefore all subsequent thermomechanical evolution) are not independently measured; they are selected from the discrete set already shown to reproduce either Rankine onset times or TLD maps in prior papers by overlapping authors. Once fixed, the same maps are used to claim successful prediction of eroded volumes and debris. The selection step therefore statistically forces consistency with the earlier diagnostics, even though the new nonlinear observables remain partially independent.

  2. self citation load bearing [Sec. 1 Introduction; Sec. 3.1; Sec. 4.1]
    "In our previous work, a KORC-Geant4-COMSOL workflow was demonstrated to be capable of predicting the onset of brittle failure within one-way coupled linear thermo-elasticity and Rankine’s failure theory [8, 9]. ... the 3 MeV & 25 ◦ pitch scenario will be followed, since it leads to accurate predictions for the onset of brittle failure [8]."

    The justification for adopting the particular RE impact parameters (and for claiming continuity of the failure mechanism) rests on the authors’ own prior Rankine-based simulations rather than on an external, independently verified reconstruction. While the present JH-2/SPH extension is new, the load-bearing premise that these parameters correctly represent the experimental RE beam is imported via self-citation.

full rationale

The thermomechanical core (JH-2 constitutive equations, EPS=0.025 failure, adaptive FEM-to-SPH conversion, Geant4 energy maps once RE parameters are fixed) is self-contained, draws material constants from external literature, and is benchmarked against new observables (eroded volumes ~130 vs 140 mm^{3} and ~60 vs 30 mm^{3}, depths, debris speed distributions) that were not used to select the inputs. The only circular element is the discrete choice of RE impact parameters (3 MeV/25° Case 1; 4 MeV/10° Case 2) justified solely by their prior success on related diagnostics in papers with author overlap. That choice is acknowledged in Sec. 3.1 and Table I but is not load-bearing for the paper’s central claim that the nonlinear extension reproduces fragmentation once the deposition map is given. Residual discrepancies are openly attributed to experimental tracking limits and SPH birth-location bias rather than hidden. No self-definitional loops, uniqueness theorems, or ansatz smuggling appear. Score 3 reflects one non-central fitted-input step plus ordinary self-citation of the group’s earlier Rankine onset work.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim rests on standard continuum mechanics plus a small set of literature-sourced material constants and two reconstructed RE beam parameters that were already tuned to earlier diagnostics. No new physical entities are postulated; the free parameters are the usual JH-2 coefficients and the EPS failure threshold taken from graphite literature.

free parameters (3)
  • EPS failure threshold ε_fail = 0.025
    Set to 0.025 for ATJ graphite; controls the moment of FEM-to-SPH conversion and therefore the final eroded volume.
  • JH-2 strength parameters A,B,C,n,m and HEL = literature values from Refs. [22,23]
    Taken from external graphite dynamic-loading references; they set the intact and fractured strength surfaces that govern plastic work and damage accumulation.
  • RE mono-energy and pitch angle = 3 MeV/25° (Case 1); 4 MeV/10° (Case 2)
    Chosen from discrete reconstructed candidates so that Case 1 matches prior Rankine onset and Case 2 matches TLD data; they fix the Geant4 volumetric heat source Q.
assumptions (4)
  • domain assumption One-way thermal-to-mechanical coupling (temperature affects stress via thermal strain and temperature-dependent moduli, but plastic work does not feed back into the heat equation for SPH particles).
    Stated in Sec. 3.4; simplifies the hybrid FEM-SPH scheme but neglects possible local heating of debris.
  • domain assumption Radiative and vaporization cooling are negligible on the 1 ms loading timescale.
    Sec. 3.4; justified by moderate temperatures but becomes questionable once hot debris is ejected.
  • ad hoc to paper Failed elements convert instantaneously to stress-free SPH particles that inherit only nodal velocity and constant temperature.
    Core of the adaptive algorithm (Sec. 3.4); the zero-stress assumption after conversion directly shapes the low-speed part of the debris distribution.
  • standard math Continuum conservation laws (mass, momentum) remain valid for both FEM and SPH representations of ATJ graphite.
    Eqs. (2)-(3); standard Lagrangian continuum mechanics.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Runaway electron induced explosions of graphite; modeling versus controlled DIII-D experiments." pith.science (2026). https://pith.science/paper/LWDH7ON2

@misc{pith2026260703914,
  author       = {Pith},
  title        = {Pith review of: Runaway electron induced explosions of graphite; modeling versus controlled DIII-D experiments},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LWDH7ON2}},
  note         = {Machine review of arXiv:2607.03914}
}
read the original abstract

The state-of-the-art concerning the modeling of the thermomechanical response of graphite to runaway electron (RE) impact is based on one-way coupled linear thermoelasticity combined with Rankine's strength-based failure criterion and limited to the onset of material failure. Here, the predictive capabilities are extended to the nonlinear damage phase characterized by material fragmentation and debris expulsion. This is achieved by introducing plasticity via the Johnson-Holmquist constitutive model, adopting an effective plastic strain-based failure criterion and coupling finite element analysis with smoothed-particle hydrodynamics. The extended thermomechanical model is successfully benchmarked against the results of two controlled RE-induced damage experiments recently carried out in DIII-D. This constitutes an important step towards the final objective of quantitatively describing the thermomechanical response of tungsten to REs.

Figures

Figures reproduced from arXiv: 2607.03914 by the authors.

Figure 1
Figure 1. Case 1 (#191366): (a) post-mortem confocal [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. LS-DYNA coupled FEM-SPH simulation results for Case 1 (#191366): enlarged view of the damaged [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Final RE-induced graphite damage at t = 4 ms; (a) Case 1, #191366, (b) Case 2, #200241. The colormap indicates element displacement. discussed in Sec. 3. In particular, the larger wetted area, smoother spatial gradients and lower peak values of the energy density, lead to a weaker and delayed growth of the strain field. Consequently, a considerably smaller fraction of the material volume satisfies the failure criter… view at source ↗
Figures from the paper (1 more)
Figure 5
Figure 5. Figure 5: The SPH particle (blue) vs the experimen [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

32 extracted references · 1 canonical work pages

  1. [1]

    Pitts, A

    R.A. Pitts, A. Loarte, T. Wauters, M. Dubrov, et al. Plasma-wall interaction impact of the ITER re-baseline. Nucl. Mater. Energy, 42:101854, 2025

    2025

  2. [2]

    R. A. Pitts, X. Bonnin, F. Escourbiac, H. Frerichs, et al. Physics basis for the first ITER tungsten divertor.Nucl. Mater. Energy, 20:100696, 2019

    2019

  3. [3]

    Krieger, S

    K. Krieger, S. Brezinsek, J.W. Coenen, et al. Scrape-off layer and divertor physics: Chapter 5 of the special issue: on the path to tokamak burning plasma operation.Nucl. Fusion, 65(4):043001, 2025

    2025

  4. [4]

    Ratynskaia, M

    S. Ratynskaia, M. Hoelzl, E. Nardon, et al. Runaway electron-induced plasma facing component damage in tokamaks.Plasma Phys. Control. Fusion, 2025, DOI:10.1088/1361-6587/ae1c6c

    work page doi:10.1088/1361-6587/ae1c6c 2025

  5. [5]

    E. M. Hollmann, C. Marini, D. L. Rudakov, E. Martinez- Loran, et al. Measurements of post-disruption runaway electron kinetic energy and pitch angle during final loss instability in DIII-D.Plasma Phys Control. Fusion, 67:035020, 2025

    2025

  6. [6]

    E. M. Hollmann, D. L. Rudakov, C. Marini, E. Martinez- Loran, et al. Characterization of runaway electron impact on instrumented sacrificial limiters on DIII-D. Nucl. Fusion, 66:056035, 2026

    2026

  7. [7]

    Testing tungsten plasma facing components in WEST and AUG tokamaks: lessons for ITER.Nucl

    Y Corre, J Gaspar, K Krieger, S Ratynskaia, et al. Testing tungsten plasma facing components in WEST and AUG tokamaks: lessons for ITER.Nucl. Fusion (accepted), 2026

    2026

  8. [8]

    Ratynskaia, P

    S. Ratynskaia, P. Tolias, T. Rizzi, K. Paschalidis, et al. Modelling the brittle failure of graphite induced by the controlled impact of runaway electrons in DIII-D.Nucl. Fusion, (2):024002, 2025

    2025

Show all 32 references
  1. [9]

    Rizzi, S

    T. Rizzi, S. Ratynskaia, P. Tolias, Y. Corre, et al. Modeling of runaway electron induced damage on boron-nitride tiles in WEST.Nucl. Mater. Energy, 46:102097, 2026

    2026

  2. [10]

    Kaselouris, V

    E. Kaselouris, V. Dimitriou, I. Fitilis, A. Skoulakis, et al. The influence of the solid-to-plasma phase transition on the generation of plasma instabilities.Nat. Commun., 8:1713, 2017

    2017

  3. [11]

    Kaselouris, G

    E. Kaselouris, G. Tamiolakis, I. Fitilis, A. Skoulakis, et al. Instability growth mitigation study of a dielectric coated metallic wire in a low current Z-pinch configuration. Plasma Phys. Control. Fusion, 63:085010, 2021

    2021

  4. [12]

    J. O. Hallquist.LS-DYNA Theory Manual. Livermore Software Technology Corporation, Livermore, CA, 2006

    2006

  5. [13]

    del Castillo-Negrete, L

    D. del Castillo-Negrete, L. Carbajal, D. Spong, and V. Izzo. Numerical simulation of runaway electrons: 3-D effects on synchrotron radiation and impurity-based runaway current dissipation.Phys. Plasmas, 25:056104, 03 2018

    2018

  6. [14]

    Hell´ e, R

    A. Hell´ e, R. Hugon, F. Brochard, S. Chouchene, et al. Dynamics of the material ejection in a dipolar arc in continuous regime.J. Phys. D: Appl. Phys., 57:425201, 2024

    2024

  7. [15]

    Brochard, A

    F. Brochard, A. Shalpegin, S. Bardin, T. Lunt, et al. Video analysis of dust events in full-tungsten ASDEX Upgrade. Nucl. Fusion, 57:036002, 2017

    2017

  8. [16]

    Carbajal, D

    L. Carbajal, D. del Castillo-Negrete, D. Spong, S. Seal, and L. Baylor. Space dependent, full orbit effects on runaway electron dynamics in tokamak plasmas.Phys. Plasmas, 24:042512, 2017

    2017

  9. [17]

    M. T. Beidler, D. del Castillo-Negrete, L. R. Baylor, D. Shiraki, and D. A. Spong. Spatially dependent modeling and simulation of runaway electron mitigation in DIII-D.Phys. Plasmas, 27:112507, 2020

    2020

  10. [18]

    D. A. Humphreys and A. G. Kellman. Analytic modeling of axisymmetric disruption halo currents.Phys. Plasmas, 6:274256, 1999

    1999

  11. [19]

    Smoothed particle hydrodynamics.Rep

    J J Monaghan. Smoothed particle hydrodynamics.Rep. Prog. Phys., 68:1703, 2005

    2005

  12. [20]

    G. R. Johnson and T. J. Holmquist. An improved computational constitutive model for brittle materials. AIP Conf. Proc., 309:981, 1994

    1994

  13. [21]

    Properties of atj graphite

    V Kachur. Properties of atj graphite. Technical report, Westinghouse Electric Corp., Pittsburgh, Pa. (USA). Astronuclear Lab., 11 1964

    1964

  14. [22]

    B. D. Huddleston, T. A Mason, C. J. Gibson, C. Z. Angell, and N. Rasmussen. Response of graphite to dynamic loading and hypervelocity jet impacts. Idaho National Laboratory (INL), Idaho Falls, ID (USA), 11 2023

    2023

  15. [23]

    Jaulin, D

    V. Jaulin, D. H´ ebert, B. Aubert, J.-L. Rullier, et al. Laser- induced cratering of a 3DCC material at mesoscale: Ex- periments and simulations.EPJ Web Conf., 183:01028, 2018

    2018

  16. [24]

    De Angeli, P

    M. De Angeli, P. Tolias, S. Ratynskaia, D. Ripamonti, et al. Evidence for high-velocity solid dust generation induced by runaway electron impact in FTU.Nucl. Fusion, 63:014001, 2023

    2023

  17. [25]

    Levy and J.F

    S. Levy and J.F. Molinari. Dynamic fragmentation of ceramics, signature of defects and scaling of fragment sizes.J. Mech. Phys. Solids, 58:12, 2010

    2010

  18. [26]

    Shenoy and K.-S

    V.B. Shenoy and K.-S. Kim. Disorder effects in dynamic fragmentation of brittle materials.J. Mech. Phys. Solids, 51:2023, 2003

    2023

  19. [27]

    B. N. Breizman, P. Aleynikov, E. M. Hollmann, and Runaway electron induced explosions of graphite; modeling versus controlled DIII-D experiments9 M. Lehnen. Physics of runaway electrons in tokamaks. Nucl. Fusion, 59:083001, 2019

    2019

  20. [28]

    Ratynskaia, E

    S. Ratynskaia, E. Thor´ en, P. Tolias, R. A. Pitts, et al. Resolidification-controlled melt dynamics under fast transient tokamak plasma loads.Nucl. Fusion, 60:104001, 2020

    2020

  21. [29]

    Ratynskaia, A

    S. Ratynskaia, A. Bortolon, and S. I. Krasheninnikov. Dust and powder in fusion plasmas: recent developments in theory, modeling, and experiments.Rev. Mod. Plasma Phys., 6:20, 2022

    2022

  22. [30]

    Tolias, M

    P. Tolias, M. De Angeli, D. Ripamonti, S. Ratynskaia, et al. Wall cratering upon high velocity normal dust impact. Fus. Eng. Des., 195:113938, 2023

    2023

  23. [31]

    De Angeli, P

    M. De Angeli, P. Tolias, F. Suzuki-Vidal, D. Ripamonti, et al. Normal high velocity solid dust impacts on tiles of tokamak-relevant temperature.Nucl. Mater. Energy, 41:101735, 2024

    2024

  24. [32]

    Tolias, M

    P. Tolias, M. De Angeli, D. Ripamonti, S. Ratynskaia, et al. Wall damage due to oblique high velocity dust impacts. Fus. Eng. Des., 225:115675, 2026

    2026

Pith tools

Reviewed July 11, 2026 · model on record in the stance chip above.