Comparison of four treatment methods for modeling streamer–to–spark discharge problems

Abstract A spark is a discharge phenomenon that occurs when streamers contact each other to form discharge channels and is characterized by rapidly increasing plasma densities and amounts of energy deposited. High–density plasma generally exists throughout the discharge channel, making numerical simulations challenging, especially under conditions requiring rapid evaluation. The conventional explicit plasma–fluid model exhibits extremely low computational speeds when modeling spark discharges, whereas the widely used linear field and ambipolar diffusion models require algorithm switching. Systematic evaluations and comparative analyses of streamer–to–spark processes have not yet been reported, and issues of computational efficiency and accuracy have not yet been adequately addressed. Therefore, in this study, we (1) compared, analyzed, and summarized the limitations of the conventional explicit plasma–fluid, linear field, and ambipolar diffusion models for spark–discharge simulations and (2) integrated structured adaptive mesh refinement and semi–implicit methods into the 2D parallel plasma solver PASSKEy to achieve comprehensive streamer–to–spark simulations without model switching. An optimal time–step range for our code in spark modeling is proposed, and the corresponding errors are estimated.