SoPlasmaFoam: an OpenFOAM-based solver for streamer and dielectric barrier discharges with adaptive mesh refinement

arXiv:2607.05137 · physics.plasm-ph, cs.CE, physics.comp-ph · Submitted 2026-07-06 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Physics. The day in general physics: instruments, methods and results across the field.

Kai: We're looking at the paper: "SoPlasmaFoam: an OpenFOAM-based solver for streamer and dielectric barrier discharges with adaptive mesh refinement".

Title and authors: Kai: So, focusing on the title and authors of "SoPlasmaFoam: an OpenFOAM-based solver for streamer and dielectric barrier discharges with adaptive mesh refinement," what's the main takeaway there? It sounds like a very specific tool for a very specific physical problem.

Mira: It’s about taking plasma physics, specifically streamer and dielectric barrier discharges, and building it into an open-source solver that uses OpenFOAM as its base. The authors are Rention Pasolaria and Konstantinos Kourtzanidis from the ARTEMIS Laboratory at the University of Western Macedonia.

Lev: For us in error correction research, what I look for in a tool like this is how it handles the complexity of the underlying equations. Does it have robust ways to manage those coupling terms between drift, diffusion, and reaction?

Kai: Well, they are tackling that directly by solving the transport equations and the Poisson equation self-consistently. They aren't just slapping pieces together; they are trying to solve them all at once either explicitly or semi-implicitly.

Mira: And the big deal is that they’ve chosen a specific numerical method for particle transport, calling it ROUNDF, which they say outperforms all the standard TVD limiters in terms of accuracy and robustness when dealing with those strong coupling effects between drift and diffusion processes.

The paper's summary: Kai: So, let’s talk about what the authors actually summarized in this paper. They basically introduced SoPlasmaFoam as an open-source, multi-region plasma dielectric solver that uses OpenFOAM combined with PETSc and blastAMR.

Mira: The core idea they are presenting is solving the drift-diffusion-reaction transport equations coupled self-consistently with the Poisson equation. They explicitly mention that this coupling can be done either through a fully explicit treatment or a semi-implicit formulation of the Poisson equation.

Lev: That choice between explicit and semi-implicit for the Poisson equation is crucial because it directly impacts stability, which is where my world comes in. Explicit treatments demand very small time steps due to stability constraints from things like dielectric relaxation times or electron drift velocities.

Kai: And they mention that the time step size is restricted by factors like those relaxation times and chemical reaction speeds. They have to use a second-order backward scheme for time integration throughout this work, which is standard but important context.

Mira: The authors also highlight that the choice of flux discretization for particle transport was particularly important because of how strongly drift and diffusion are coupled together in these systems. That's where they picked ROUNDF as the best option over other schemes like TVD limiters.

The paper's improvements: Kai: Now, let’s look at what improvements the authors suggest or have developed to make this solver better. They focus on several things, starting with how they handle the mesh itself.

Mira: One key improvement is using blastAMR to support adaptive mesh refinement that works on hexahedral and polyhedral meshes. This means it can automatically increase resolution locally in regions of interest, like a streamer head, without needing complex pre-defined geometries.

Lev: That’s useful because traditional methods often require you to define the mesh upfront; having something that handles arbitrary polyhedral meshes is a big win for simulating complex interfaces.

Kai: They also addressed the challenge of solving the Poisson equation across multiple regions. The paper notes that solving it in a partitioned way can be slow, so they opted for a monolithic solution of both domains, even though that results in larger and more complex matrices.

Mira: And they tackled temporal constraints by using semi-implicit Poisson solvers or implicit transport schemes to help overcome the stability barriers imposed by those rapid physical changes. They are also working on developing a drift-robust wall boundary condition which is supposed to be accurate even in the drift-dominated limit, unlike conventional mixed boundary formulations that fail there.

Conclusion: Kai: So, wrapping up what we’ve heard on "SoPlasmaFoam: an OpenFOAM-based solver for streamer and dielectric barrier discharges with adaptive mesh refinement," the main thing is that they’ve created a framework that handles multi-region coupling, uses specific numerical schemes like ROUNDF, and has built in adaptive mesh refinement.

Mira: The implication is that this tool should be much better equipped to simulate complex discharge phenomena accurately because it’s designed for the non-equilibrium physics found in non-thermal plasmas. It moves beyond just a simple flow model to include those chemical and electric interactions.

Lev: From my side, what I see is that while the framework is solid, running this on real hardware will still be memory bound, especially with adaptive mesh refinement enabled. The paper confirms that with adaptive mesh refinement on the positive streamer case, it's competitive with other fast plasma codes up to six to twelve cores on a single node.

Kai: So they’re saying this tool is competitive in terms of speed for resolving the positive streamer case and shows strong parallel scaling. It really gives researchers a practical path forward for studying these kinds of discharges.

Mira: And they are looking ahead, mentioning that future work involves coupling it with Cantera to incorporate chemistry, which aims to create a unified plasma-fluid solver that studies flow patterns and energy exchange between the plasma and the bulk fluid.

Lev: That coupling with chemical kinetics is where things get really interesting for me; modeling those reactions in environments that can't handle high temperatures or external energy input is a major goal.

Kai: So, to summarize, SoPlasmaFoam provides a validated solver that's fast on the streamer benchmark and sets up the path for more complex multiphysics simulations down the line. We’ll be looking at other papers next.

Rention Pasolaria, Konstantinos Kourtzanidis

Advanced Renewable Technologies for Energy & Materials Integrated Systems (ARTEMIS) Laboratory · Department of Mechanical Engineering, University of Western Macedonia

physics.plasm-ph, cs.CE, physics.comp-ph

Submitted: 2026-07-06

Updated: 2026-07-06

Journal ref: Computer Physics Communications, Volume 331, 2027, Article 110505

DOI: 10.1016/j.cpc.2026.110505

Code: https://github.com/STFS-TUDa/blastAMR

Project page: http://www.mcs.anl.gov/petsc

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 85/100

The gist: SoPlasmaFoam presents an open-source, multi-region plasma–dielectric solver built on OpenFOAM and integrated with PETSc and blastAMR to solve drift-diffusion-reaction transport equations coupled…

Key concepts

Adaptive Mesh Refinement (AMR)
AMR allows the simulation grid to automatically increase resolution in specific areas where needed, such as the tip of a streamer. This focuses computational power where gradients are steep, leading to highly accurate results without needing a fixed, overly dense mesh everywhere.
ROUNDF Scheme
This is a specific numerical method used for particle transport that was found to be superior to standard schemes. It accurately captures sharp changes in particle movement, which is crucial for modeling the rapid ionization fronts seen in streamers, ensuring high accuracy and robustness.
Monolithic Multi-domain Coupling
The solver can handle plasma and dielectric regions simultaneously within the same computational framework. This allows for accurate modeling of complex geometries where the plasma interacts with solid materials, handling arbitrary curved interfaces effectively.
Cantera Integration
Coupling SoPlasmaFoam with Cantera allows researchers to include chemical reactions in the simulation. This is important for studying environments where chemical kinetics are important, enabling the modeling of reactions alongside plasma transport.

Terminology

Summary

SoPlasmaFoam presents an open-source, multi-region plasma–dielectric solver built on OpenFOAM and integrated with PETSc and blastAMR to solve drift-diffusion-reaction transport equations coupled self-consistently with the Poisson equation The solver is competitive with the fastest reported plasma codes on the streamer benchmark when adaptive mesh refinement is used

Solver Architecture and Components

The solver, SoPlasmaFoam, is built on OpenFOAM and integrates several state-of-the-art third-party libraries to make it suitable for large-scale high-performance computing Matrix systems can be solved either with the native OpenFOAM linear solvers or through the petsc4Foam framework, which incorporates the PETSc library into OpenFOAM, providing access to a broad range of linear and ODE solvers with both CPU and GPU support Adaptive mesh refinement is provided through the blastAMR library, which supports both hexahedral and polyhedral meshes and operates natively in 1D, 2D, 2D-axisymmetric and 3D configurations The solver also supports dielectric regions of arbitrary number and shape, and the Poisson equation can be solved monolithically across all regions simultaneously

Numerical Methodology

The numerical methodology involves solving the governing equations using a segregated timemarching strategy within an outer PIMPLE loop Two coupling strategies are supported: a fully explicit treatment and a semi-implicit treatment of the Poisson equation For time integration, a second-order backward scheme is mainly employed throughout this work The choice of flux discretization for particle transport is particularly important due to the strong coupling between drift and diffusion processes The ROUNDF scheme emerged as the clear best choice, outperforming all standard TVD limiters in both accuracy and robustness

Validation Results

The solver was validated against canonical test cases, including a low-pressure DC glow discharge and the positive streamer benchmark of [19] The results show good agreement with established fluid-model plasma solvers from the literature The solver's performance in the positive streamer case is competitive with other codes, and it exhibits strong parallel scaling up to 6-12 cores on a single node

Performance Analysis

The performance analysis confirms the expected memory-bound behavior of a finite volume code, with good single-node scaling The dominant cost is the electromagnetic block, which accounts for approximately two-thirds of the total runtime AMR enabled results show that the solver is very fast at resolving the positive streamer case, placing it among the fastest available

Future Directions

The general framework shows great promise and lays the path for multidimensional, multiphysics plasma simulations Future work will focus on implementing the Local mean Energy Approximation (LMEA/LEA) through the solution of the electron energy equation, photoionization, field emission boundary conditions, and circuit models for predicting discharge voltage and current Another major development chapter is the incorporation of chemistry by coupling OpenFOAM with Cantera The ultimate goal is a unified plasma-fluid solver capable of studying flow patterns in plasma, electrohydrodynamic (EHD) forces acting on the fluid, and energy exchange processes between the plasma and the bulk fluid

Conclusion

The solver was validated against canonical benchmark cases, including a low-pressure DC glow discharge and positive streamers in dry air at different pre-ionization background levels The drift-robust wall boundary condition was also developed and validated, overcoming the Péclet-dependent failure of the conventional mixed-boundary formulation that is commonly employed in plasma-fluid codes The solver exhibited strong parallel scaling up to 6-12 cores on a single node, beyond which memory bandwidth saturation reduced further speedup, confirming the memory-bound nature of the finite volume algorithm The AMR enabled results show that SoPlasmaFoam is competitive with the fastest reported plasma codes on the streamer benchmark and the methodological findings provide practical guidance for the broader plasma simulation community

Author Contributions

Rention Pasolari: Data curation, Formal analysis, Investigation, Methodology, Software, Validation, Visualization, Writing - original draft Konstantinos Kourtzanidis: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation

Acknowledgements

This work was supported by the European Union Project CAIPIRINH3A, under the GA Number 101191768 The research project is implemented in the framework of H.F.R.I call “4th Call for H.F.R.I.’s Research Projects to Support Postdoctoral Researchers” (H. p. 2

Where the paper says it

Recently, Vidyut3d [25], an AMReX-based solver that introduces adaptive mesh refinement and hybrid CPU/GPU support, has been published. The KAUST Clean Energy Research Platform has independently introduced reactPlasFOAM [26], another OpenFOAM-based unified plasma-combustion solver that couples the ChemPlasKin gas-plasma kinetics library [9] to OpenFOAM, providing AMR, dynamic load balancing, and on-the-fly electron Boltzmann solves.

F.R.I Project Number: 28223) Open access publishing was supported by HEAL-Link (Hellenic Academic Libraries Link)

References

[1] J. E. Thomas, K. Stapelmann, Plasma control: A review of developments and applications of plasma medicine control mechanisms, Plasma 7 (2) (2024) 386–426

[19] B. Bagheri, J. Teunissen, U. Ebert, M. M. Becker, S. Chen, O. Ducasse, O. Eichwald, D. Loffhagen, A. Luque, D. Mihailova, J. M.-Plewa

[24] A K Verma, A Venkattraman, Somafoam: An openfoam based solver for continuum simulations of lowtemperature plasmas

[43] S Bnà, I Spisso, M Olesen, G Rossi, Partnership for advanced computing in europe PETSc4FOAM: A library to plug-in PETSc into the OpenFOAM framework

[44] blastAMR: Load-balanced adaptive mesh refinement libraries from blastfoam ported to ESI OpenFOAM

[45] X Deng, A unified framework for non-linear reconstruction schemes in a compact stencil. part 1: Beyond second order

[46] X Deng, A new open-source library based on novel highresolution structure-preserving convection schemes

[47] D L Scharfetter, H K Gummel, Large-signal analysis of a silicon Read diode oscillator

[48] G Chourdakis, D Schneider, B Uekermann, OpenFOAM-preCICE: Coupling OpenFOAM with external solvers for multi-physics simulations

[49] A Villa, L Barbieri, M Gondola, R Malgesini, An asymptotic preserving scheme for the streamer simulation

[52] S Balay, S Abhyankar, M F Adams, J Brown, P Brune

[53] A Derzsi, P Hartmann, I Korolov, J Karácsony

[54] A V Phelps, Z L Petrovic, Cold-cathode discharges and breakdown in argon: surface and gas phase production of secondary electrons

[58] K Kourtzanidis, L L Raja, Three-electrode sliding nanosecond dielectric barrier discharge actuator: modeling and physics

[61] D G Goodwin, H K Moffat, I Schoegl, R L Speth, B W Weber

[63] R Mao, M Lin, Y Zhang, T Zhang

[64] A G Özbay, A Hamzehloo, S Laizet, P Tzirakis

[62] D Zhou, H Zhang, S Yang, A robust reacting flow solver with computational diagnostics based on openfoam and cantera

[41] COMSOL AB, COMSOL Multiphysics®, stockholm, Sweden

[42] D Levko, R Upadhyay, K Suzuki, C Shukla, L Raja

[43] S Bnà, I Spisso, M Olesen, G Rossi

[48] G Chourdakis, D Schneider, B Uekermann

[51] S Ferraris, Implicit treatment of coupled boundary conditions

[59] K Kourtzanidis, Full cycle, self-consistent, two-dimensional analysis of a packed bed dbd reactor for plasma-assisted co 2 splitting: spatiotemporal inhomogeneous, glow to streamer to surface discharge transitions

[60] R Morrow, N Sato, The discharge current induced by the motion of charged particles in time-dependent electric fields; sato’s equation extended

Table 2

Group / Code AMR ∆xmin Ncells,max ∆t Corr. Cores Processor Wall time (µm)

|---|---|---|---|---|---|---|---

CWI∗ (Afivo) Yes 3.05/1.53/0.76 — dyn. — 4 Xeon E3-1271 v3, 3.6 GHz 5 min

CWI∗ (Afivo) No 3.0 dyn. — - - -

ES∗ Yes 3.9 20 × 10 6 dyn., max 1 ps — 1 Core i7-6700K, 4. p. 4

Where the paper says it

The code is built in a modular fashion, making extensive use of the object-oriented features of C++ and the OpenFOAM run-time selection mechanism. Base classes are defined for the core physical components, like electromagnetics, plasma species, and plasma transport, from which specific models are derived through inheritance and polymorphism.

0 GHz 20 h

FR∗ No 3.0 1.1 × 10 6 dyn., — - - -

FR∗ No 1.5 — dyn. — - - -

CN∗ Yes 2.0 6.5 × 10 5 dyn., max 10-11 s — 4 Xeon E3-1271 v3, 3. p. 5

Where the paper says it

These conditions are summarised as follows. For positive and negative ions: Γ± = Γth,± + max 0, Γdrift,± (14) For neutral species: Γn = Γth,n (15)

6 GHz 8 h

TUE∗ No — - - -

DE∗ No 4.2 × 10 6 dyn., max 5 ps — 6 Xeon E5-2690 15 h

reactPlasFOAM [26] Yes 3.125 1.1 × 10 5 — - - - — - — — — 20 min

Vidyut3D [25] Yes 1.5 ∼ 2 × 10 6 dyn., ∼ 2.5 ps 264 EPYC-Genoa + 4×H100 (GPU) 30 min

SoPlasmaFoam No 1.5 3.5 × 10 6 dyn., max 1 ps — - - - — - — --- ---

Figure 24

Strong-scaling analysis of the positive streamer Case A for the coarse (449,413 cells) and dense (3,472,738 cells) meshes on the WS (Ryzen 9 5900X) and HPC (EPYC 7543P) systems: (a) wall-clock execution time over 100 time steps, (b) speedup S = T1/Tn with the ideal linear scaling for reference, and (c) parallel efficiency E = S/n with the ideal 100% line

Figure 25

MPI profile statistics on the WS (Ryzen 9 5900X) for the positive streamer Case A: (a) MPI time as a fraction of the average per-rank elapsed time, and (b) MPI load imbalance, defined as (tmax − tmin)/tmax over the cumulative MPI time of each rank

Table 6

Wall-clock time breakdown of the main solver components for the positive streamer case. Times correspond to the maximum across all MPI ranks

Figure 15

Electric field magnitude (MV/m) for the AMR case at t = 12 ns. (a) Full streamer structure, mirrored about the symmetry axis. p. 3

Where the paper says it

It is worth stating precisely how SoPlasmaFoam differs from existing OpenFOAM-based plasma solvers and from the closely related open-source adaptive codes. Relative to SOMAFOAM [24], which spans a wide pressure range but mentions no AMR capability beyond what the underlying OpenFOAM framework natively provides (3D hexahedral cells only), and has not been applied to streamer dynamics, the present solver is oriented toward streamer and surfacedischarge propagation and brings a full-dimensional AMR capability (1D through 3D and axisymmetric), PETSc back-end and ROUND limiters into a single OpenFOAM-native framework.

(b) Zoomed view of the highlighted region in (a), showing the electric field together with the adaptive mesh

Figure 14

Maximum absolute electric field as a function of streamer length for the AMR case compared against the uniform 3 µm mesh result

Figure 13

Convergence study of the outer PIMPLE iterations for the positive streamer Case A using a semi-implicit Poisson scheme: (a) maximum absolute electric field Emax and (b) streamer length L(t), both as a function of time up to 5.5 ns, and an adjustable time step with maximum dielectric relaxation ratio Cε = 2

Figure 11

Convergence study of the outer PIMPLE iterations for the positive streamer Case A: (a) maximum absolute electric field Emax and (b) streamer length L(t), both as a function of time up to 5.5 ns, each with 1, 2, 3 and 4 outer PIMPLE correctors

Figure 12

Convergence study of the outer PIMPLE iterations for the positive streamer Case A: (a) maximum absolute electric field Emax and (b) streamer length L(t), both as a function of time up to 5.5 ns, each with 1, 2 and 3 outer PIMPLE correctors

Figure 10

Validation of the high-background ionization streamer (Case A) against reference data from [19] and [27]: (a) maximum absolute electric field, (b) reduced streamer length z − v0t

Figure 16

Contour plots of the electric field magnitude (V/m) at three time instances (t = 5, 15 and 20 ns) for the positive streamer Case B, with a background ionization of 109 m−3 for both electrons and ions

Figure 17

Validation of the low-background ionization streamer (Case B) against reference data from [19] and [27]: (a) maximum absolute electric field, (b) reduced streamer length z − v0t

Figure 18

Computational domain of the nanosecond SDBD actuator. The powered electrode (50 µm thick, 1 mm long) sits on top of a dielectric layer (0. p. 4

Where the paper says it

Zero-gradient (Neumann) boundary conditions are applied to the electric potential in the far-field regions, while Dirichlet boundary conditions are imposed on the electrode surfaces. In simulations involving multiple domains, such as plasma regions coupled with dielectric materials, continuity of the electric potential and normal component of the electric displacement field is enforced through coupled (jump) boundary conditions at material interfaces.

3 mm, εr = 3.5), above a 10 mm grounded electrode

Figure 22

Electric field magnitude contours at t = 3.5, 4.5, 7.5, and 10 ns for the nanosecond SDBD actuator simulation using the Scharfetter-Gummel (SG) scheme

Table 4

Numerical schemes used in the SDBD simulation

Term Scheme Time derivative backward Divergence Gauss ROUNDF Laplacian (transport) Gauss harmonic limited corrected 1.0 Laplacian (Poisson) Gauss linear limited corrected 0. p. 3

Where the paper says it

Matrix systems can be solved either with the native OpenFOAM linear solvers or through the petsc4Foam framework [43], which incorporates the PETSc library into OpenFOAM, providing access to a broad range of linear and ODE solvers with both CPU and GPU support, and enabling efficient scaling on HPC clusters. Adaptive mesh refinement is provided through the blastAMR library [44], which supports both hexahedral and polyhedral meshes and operates natively in 1D, 2D, 2D-axisymmetric and 3D configurations, in contrast to the standard OpenFOAM AMR implementation which is limited to 3D.

5 Interp. mob. & diff. Harmonic Interp. p. 6
Where the paper says it

Spatial gradients, such as those required for the calculation of the electric field, are evaluated using a second-order least-squares scheme. Regarding interpolation, mobility and diffusivity are interpolated using harmonic interpolation, while all other quantities are interpolated linearly.

densities Linear snGrad potential limited corrected 0. p. 3
Where the paper says it

Matrix systems can be solved either with the native OpenFOAM linear solvers or through the petsc4Foam framework [43], which incorporates the PETSc library into OpenFOAM, providing access to a broad range of linear and ODE solvers with both CPU and GPU support, and enabling efficient scaling on HPC clusters. Adaptive mesh refinement is provided through the blastAMR library [44], which supports both hexahedral and polyhedral meshes and operates natively in 1D, 2D, 2D-axisymmetric and 3D configurations, in contrast to the standard OpenFOAM AMR implementation which is limited to 3D.

5 snGrad densities limited corrected 1. p. 5
Where the paper says it

The normal electric flux through face f is therefore obtained as: ϕE, f = − ∂ϕ ∂n ! f Sf (6) which in OpenFOAM is evaluated via the fvc::snGrad operator.

0

Table 5

Hardware specifications of the two systems used for the scalability analysis

| Specification | WS | HPC |

|---|---|---|

| CPU | AMD Ryzen 9 5900X AMD EPYC 7543P |

| Cores / Threads | 12 cores 32 cores |

| Max CPU frequency | 4.95 GHz 2.79 GHz |

L3 cache 64 MiB 256 MiB DDR4 @ 2133 MT/s (2-channel) DDR4 @ 8-channel RAM

Figure 23

Surface charge accumulation on the dielectric surface for the nanosecond SDBD actuator simulation at t = 6 ns and t = 10 ns, shown in logarithmic scale

Hardware specifications of the two systems used for the scalability analysis

| Specification | WS | HPC |

|---|---|---|

| CPU | AMD Ryzen 9 5900X AMD EPYC 7543P |

| Cores / Threads | 12 cores 32 cores |

| Max CPU frequency | 4.95 GHz 2.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Systematic assessment of convective flux schemes on a stiff scalar-advection problem and on the positivestreamer benchmark shows that the ROUNDF scheme outperforms all tested TVD limiters and is recommended for streamer transport. This provides a robust numerical method for accurately capturing steep gradients characteristic of ionization fronts, improving simulation fidelity compared to standard TVD schemes.

  2. A quantitative study of how tightly the Poisson and transport equations must be coupled demonstrates that a semi-implicit Poisson formulation does not by itself remove this requirement, indicating that even when Courant numbers are well below unity, the coupling must be tightened for accuracy. This informs the design of tighter coupling algorithms within solvers to prevent errors arising from segregated treatments.

  3. A drift-robust wall boundary condition acts directly on the discretized matrix coefficients and remains accurate in the drift-dominated limit, where the conventional mixed-boundary mapping fails. This allows for more reliable simulation of surface phenomena where charge accumulation is critical, overcoming limitations encountered by conventional boundary mappings.

  4. The solver can be configured with adaptive mesh refinement using blastAMR, which supports AMR on both hexahedral and polyhedral meshes and functions natively in 2D and 2D-axisymmetric configurations. This capability allows the system to automatically increase resolution locally in regions of interest, such as streamer heads, without requiring complex pre-defined geometries.

  5. The solver supports monolithic multi-domain coupling for plasma and dielectric regions, which supports an arbitrary number of curved dielectric interfaces. This enables accurate modeling of complex geometries in surface discharge simulations where the plasma and solid materials interface with varying curvatures.

  6. The solver can be coupled with Cantera to incorporate chemistry, addressing the need for stiff chemistry by exploring operator splitting or fully coupled solution strategies. This allows AI systems to model complex chemical kinetics alongside plasma transport, enabling the simulation of reactions in environments that cannot withstand high temperatures or external energy input.

  7. The system's performance analysis confirms that with adaptive mesh refinement, the solver is competitive with the fastest reported plasma codes on the streamer benchmark, demonstrating high efficiency for resolving complex discharge physics. This confirms its viability as a fast and accurate tool for emerging applications like plasma-assisted combustion and flow control.

Abstract

SoPlasmaFoam is an open-source, multi-region plasma-dielectric solver built on OpenFOAM, integrated with the PETSc linear-algebra suite (CPU and GPU back-ends), the blastAMR adaptive-mesh-refinement library (hexahedral and polyhedral meshes), and the ROUND family of high-resolution convective schemes. It solves drift-diffusion-reaction transport for charged species, coupled self-consistently to Poisson's equation explicitly or semi-implicitly, with plasma and dielectric regions joined by a monolithic multi-domain coupling for arbitrary curved interfaces. This work makes three contributions. First, a systematic assessment of convective schemes on a stiff scalar-advection problem and the positive-streamer benchmark shows that Scharfetter-Gummel is stable but excessively diffusive on coarse meshes, while ROUNDF outperforms all tested TVD limiters and is recommended for streamer transport. Second, an analysis of Poisson-transport coupling shows that fixed-point correction loops critically control accuracy, that a semi-implicit Poisson formulation does not remove this requirement, and that coupling must be tightened even when Courant and dielectric-relaxation numbers are well below unity. Third, a drift-robust wall boundary condition acting on discretized matrix coefficients is introduced, remaining accurate in the drift-dominated limit where conventional mixed-boundary mappings fail. The solver is validated against a low-pressure DC glow discharge and the positive-streamer benchmark, and its multi-region capability is demonstrated on a nanosecond surface dielectric barrier discharge. Performance analysis confirms memory-bound finite-volume scaling and shows that with AMR the solver is competitive with the fastest reported plasma codes. The framework provides a modular foundation for multiphysics simulations in plasma-assisted combustion, plasma processing, and plasma-based flow control.

Sources

Related papers