Skip to content

Release 4.0.0

Latest

Choose a tag to compare

@BoltzplatzMirza BoltzplatzMirza released this 25 Sep 11:33

Breaking/Parameter Changes

  • Major restructuring has been required to accommodate the variable polynomial degree (p-adaption) feature

Documentation/Tools/Regression Testing

  • Python script for the simulation of 1D shock tubes (see tools/ShocktubeScript folder) [5f009b9]
  • superB: bugfix in output of time-dependent fields, output of linear conductors as a single file, removal of unused TimeStampLength [0da3e85]
  • Upgrade extract_userblock.py from Python 2 -> 3 [60afdfc]

Features

  • Several models for electron-based secondary electron emission (SEE) (User guide)
    • Square/exponential fit, and after Seiler 1993 [8accb21]
    • Vaughan formula (Villemant 2019) [7743b18]
    • Possibility to always emit a single secondary when using variable particle weights
  • Simple back-scattering in centre-of-mass frame using cross-section data in MCC (User guide) [bd9d715]
  • Variable polynomial degree (p-adaption) (User guide) [562c7dd]
  • Surface super-sampling for the triangle-based tracking routine (ie. most DSMC use cases) (User guide) [8799992]
  • Granular flow / solid particles can be simulated using a background gas distribution (User guide) [1d69a9b]

Improvements

  • Improved PETSc solver options and documentation in user guide. Currently, available options include a direct Cholesky solver (requires MUMPS), an iterative Conjugate Gradient solver (and a pipelined version for multi-node applications) with a block Jacobi preconditioner, and a flexible GMRES solver with either BoomerAMG (requires Hypre) or block Jacobi preconditioning. [8accb21]
  • Support of multiple dielectric regions with different parameters for the Poisson solver [cc5ed0e]
  • Background gas regions: allowing overlap by overwriting the previous region, warning during initialization in that case [0da3e85]
  • Read-in of magnetic field: compare against relative machine precision to detect not equidistant fields [0da3e85]

Fixes

  • Bugfix in CalcVarWeightMPF for DoCellLocalWeighting [2b91783]
  • Fix conservation in case of a recombination with weighting in the chemistry module [cfd81c1]
  • Bugfix in sampling for adaptive surface flux: SamplingIteration must not be utilized in case of a relaxation factor [82ae9e1]
  • Bugfix in background gas distribution: do not skip the calculation of the mean vibrational energy for chemical reactions [483c0ed]
  • Bugfix in CalcSurfaceValues for the combination of inner BCs + CalcSurfaceImpact and PorousBC [cc5ed0e]
  • Bugfix in CalcBoundaryParticleOutput with inner BCs: loop over nSides to include inner BCs in area calculation [cc5ed0e]
  • Bugfix in pre-commit hook: initialize MPICOMMWORLDERR variable with 0 [cc5ed0e]
  • Bugfix in surface flux parallel to the rotational axis [bf93d86]
  • Bugfix in 2D particle tracking in combination with clone insertion [bf93d86]
  • Fix boundary sampling allocation for varying wall temperature [7f2500c]
  • Fix reset of the surface coverage by the LoadBalance routine [9b7fd01]
  • Bugfix in raytracing simulations: distanceMin has not been updated in CalcWallAbsoprtion [7af0afb]
  • Bugfix in macroscopic restart in combination with variable time step for an axisymmetric simulation [2f468e5]
  • Bugfix in SamplePressTensHeatflux calculation [28e7474]
  • Bugfixes for restart & macroscopic restart with constant massflow surface flux (type = 4) [a97e08e]
  • Fixed a bug in the BGK collision operator occurring with variable time step or variable weighting factor [fd7a513]