Achlys Tritium Diffusion

The Achlys application aims to model macroscopic tritium transport processes through fusion materials using the MOOSE Finite Element Framework Permann et al. (2020), Gaston et al. (2015), Lindsay et al. (2021).

The Achlys source code is hosted on Github, a dockerfile is available to easily obtain the software.

Governing Equations

Achlys implements a formulation of the Foster-McNabb equations presented byHodille et al. (2015) and Delaporte-Mathurin et al. (2019) as shown in Eq. (1), Eq. (2), and Eq. (3).

(1)

(2)

(3)

is the diffusivity of the species through some material in units of . This is modelled by an Arhhenius relation as given by Eq. (4). The D0 and energy value parameters are determined experimentally for any pair of solute atom and bulk material lattice. (4)

, in units of , is the reaction rate for the trapping process and is modelled by Eq. (5) where is the lattice constant in and is the number density of solute sites in the material. (5)

, in units of , is the reaction rate for the de-trapping process from the -th trapping site. This is modelled by the Arrhenius type equation as given by Equation Eq. (6) where E is the energy barrier a trapped atom must overcome to leave the site and is referred to as the attempt frequency. (6)

Kernel selection

The kernels selected for each of the governing equations given in Eq. (1), Eq. (2), and Eq. (3) are shown below. Note that all kernels are provided by the MOOSE framework without modification except in a single case: the ADTrappingEquilibriumEquation is a custom kernel implemented within Achlys.

Note also that Automatic Differentiation Lindsay et al. (2021) has been employed within Achlys so AD kernels, BCs, and objects should be used wherever available. More information on this system is available through the MOOSE documentation.

Mobile Concentration

Trapped Concentration, site

Temperature Evolution

Symbols, units, and meanings

Here, symbols denote concentrations of the hydrogen isotope either in the trapped or mobile phase as indicated by the relevant subscript. is the concentration of a given trap type and other symbols are defined in Table 1. Mathematically, the units of concentration can either be expressed in terms of or atomic fraction; the Achlys application uses atomic fraction in its calculations.

Table 1: Nomenclature and symbol meanings

Parameter NameSymbol(s)Unit
Species Concentration
Trap Concentration
time
External sources
Trapping reaction rate
Lattice constant
Lattice sites
De-trapping reaction rate
De-trapping pre-exponential factor
Energy barriereV
Boltzman Constant
Diffusion Coefficient
Diffusion pre-exponential factor
Lattice Density
Temperature
Thermal Conductivity
Specific Heat Capacity
Mass Density

A Note on Units and Scaling

Achlys solves for the concentrations of the mobile and each trapped phase in the units of atomic fraction, this is the concentration in divided by the number density of solute atoms. Most physical processes are unaffected by this conversion and the SI units can be retrieved by applying a simple scale factor to the outputs, an exception being recombination boundary conditions which evaluate the square of the mobile concentration. In this case the scale factor must be re-applied within the boundary condition (see ADSurfaceRecombination for example).

The concentrations are scaled down in this way as a bug was experienced in early development where diffusion kernels would cease to function once the magnitude of the scalar field surpassed some threshold. This issue is yet to be further investigated.

Energies are expressed in eV (though only appear in non-dimensional ratios - so transforms could be possible) and all other units are SI.

source

Complete source definitions for all objects implemented within Achlys are detailed below.

Achlys kernels

Achlys BCs

  • ADSurfaceRecombinationImposes the integrated boundary condition , where is a variable and the constant, k, is calculated from an Arhhenious expression based on provided (constant) values of E and T
  • ADSurfaceRecombinationCoupledTImposes the integrated boundary condition , where is a variable and the constant, k, is calculated from an Arhhenious expression based on an energy barrier, E, and a coupled Temperature.
  • DirichletEquivalentPfcImplantBCImposes the essential boundary condition , where is a constant calculated assuming the recombination fluxquickly becomes equal to the incident flux. The recombination term is calculated using the simulation temperature and the pre-exponent and Energy values provided
  • DirichletPfcFunctionFluxRampBCImposes the essential boundary condition , where is a constant calculated assuming the recombination fluxquickly becomes equal to the incident flux. The recombination term is calculated using the simulation temperature and the pre-exponent and Energy values provided
  • PfcFunctionFluxRampBCImposes the essential boundary condition , where is a constant calculated assuming the recombination fluxquickly becomes equal to the incident flux. The recombination term is calculated using the simulation temperature and the pre-exponent and Energy values provided

Achlys Materials

Achlys Postprocessors

  • SideDesorptionFluxComputes the integral of the flux over the specified boundary
  • SideMaxValueFind a specific target value along a sampling line. The variable values along the line should change monotonically. The target value is searched using a bisection algorithm.
  • SideMinValueFind a specific target value along a sampling line. The variable values along the line should change monotonically. The target value is searched using a bisection algorithm.
  • ValueAtPointFind a specific target value along a sampling line. The variable values along the line should change monotonically. The target value is searched using a bisection algorithm.
  • VariableIntegralComputes a volume integral of the specified variable

References

  1. Rémi Delaporte-Mathurin, Etienne A. Hodille, Jonathan Mougenot, Yann Charles, and Christian Grisolia. Finite element analysis of hydrogen retention in iter plasma facing components using festim. Nuclear Materials and Energy, 21:100709, 2019. URL: https://www.sciencedirect.com/science/article/pii/S2352179119300547, doi:https://doi.org/10.1016/j.nme.2019.100709.[BibTeX]
  2. Derek R. Gaston, Cody J. Permann, John W. Peterson, Andrew E. Slaughter, David Andrš, Yaqi Wang, Michael P. Short, Danielle M. Perez, Michael R. Tonks, Javier Ortensi, Ling Zou, and Richard C. Martineau. Physics-based multiscale coupling for full core nuclear reactor simulation. Annals of Nuclear Energy, 84:45–54, 2015.[BibTeX]
  3. E.A. Hodille, X. Bonnin, R. Bisson, T. Angot, C.S. Becquart, J.M. Layet, and C. Grisolia. Macroscopic rate equation modeling of trapping/detrapping of hydrogen isotopes in tungsten materials. Journal of Nuclear Materials, 467:424–431, 2015. URL: https://www.sciencedirect.com/science/article/pii/S0022311515300660, doi:https://doi.org/10.1016/j.jnucmat.2015.06.041.[BibTeX]
  4. Alexander Lindsay, Roy Stogner, Derek Gaston, Daniel Schwen, Christopher Matthews, Wen Jiang, Larry K Aagesen, Robert Carlsen, Fande Kong, Andrew Slaughter, and others. Automatic differentiation in metaphysicl and its applications in moose. Nuclear Technology, pages 1–18, 2021.[BibTeX]
  5. Cody J. Permann, Derek R. Gaston, David Andrš, Robert W. Carlsen, Fande Kong, Alexander D. Lindsay, Jason M. Miller, John W. Peterson, Andrew E. Slaughter, Roy H. Stogner, and Richard C. Martineau. MOOSE: enabling massively parallel multiphysics simulation. SoftwareX, 11:100430, 2020. URL: http://www.sciencedirect.com/science/article/pii/S2352711019302973, doi:https://doi.org/10.1016/j.softx.2020.100430.[BibTeX]