A Multiscale Two Grid Strategy for Kinetic Modeling of Burning Plasmas

In laboratory thermonuclear burning plasma experiments—such as those in inertial and magnetic confinement fusion (I/MCF) devices—deuterium-tritium (DT) fuel at ~10 keV produces nearly monoenergetic 3.5 MeV alpha particles, which drive internal energy generation. Vlasov-Fokker-Planck (VFP) kinetic modeling of these plasmas faces multiscale challenges. The 3.5 MeV alphas slow down to form a highly localized 10 eV–10 keV ash population, introducing a velocity-space separation of factors from 30 to 600. This scale disparity makes traditional grid-based VFP solvers computationally unfeasible.

We present a consistent, conservative multiscale two-grid method for modeling fusion-born alphas, building on Peigney’s approach [1]. Our method uses separate velocity-space grids for alpha and thermal ash populations. A physics-based sink term removes particles transitioning to ash speeds, preventing unresolved singularities in the alpha grid. This removed population is reintroduced as a source term in the ash VFP equation, maintaining detailed balance and conserving mass, momentum, and energy. We implemented this method in the iFP VFP spherical implosion code [2], demonstrating its effectiveness on multiscale thermonuclear burning ICF problems [3].

[1] B. E. Peigney et al., J. Comp. Phys., 278 (2014) 416-444.

[2] W. Taitano et al., Comp. Phys. Comm., 263 (2021) 107861.

[3] B.L. Reichelt et al., arXiv:2506.06672v1 [physics.plasm-ph].