Influence of grid resolution in fluid-model simulation of nanosecond dielectric barrier discharge plasma actuator

Abstract

Two-dimensional numerical simulation of a surface dielectric barrier discharge (SDBD) plasma actuator, driven by a nanosecond voltage pulse, is conducted. A special focus is laid upon the influence of grid resolution on the computational result. It is found that the computational result is not very sensitive to the streamwise grid spacing, whereas the wall-normal grid spacing has a critical influence. In particular, the computed propagation velocity changes discontinuously around the wall-normal grid spacing about 2 μm due to a qualitative change of discharge structure. The present result suggests that a computational grid finer than that was used in most of previous studies is required to correctly capture the structure and dynamics of streamer: when a positive nanosecond voltage pulse is applied to the upper electrode, a streamer forms in the vicinity of upper electrode and propagates along the dielectric surface with a maximum propagation velocity of 2 × 108 cm/s, and a gap with low electron and ion density (i.e., plasma sheath) exists between the streamer and dielectric surface. Difference between the results obtained using the finer and the coarser grid is discussed in detail in terms of the electron transport at a position near the surface. When the finer grid is used, the low electron density near the surface is caused by the absence of ionization avalanche: in that region, the electrons generated by ionization is compensated by drift-diffusion flux. In contrast, when the coarser grid is used, underestimated drift-diffusion flux cannot compensate the electrons generated by ionization, and it leads to an incorrect increase of electron density.

An adaptive Cartesian embedded boundary approach for fluid simulations of two- and three-dimensional low temperature plasma filaments in complex geometries

“…We observe a height-of-flight of the positive streamer above the surface and find that the sheath region l < 30 µm from the surface contains fewer electrons and is incompletely shielded by space and surface charge effects. The topological features that we observe is also observed in planar simulations by others [31,54,9,27,55,56], also with PIC codes [57], and the plasma sheath thickness that we observe is consistent with analytical estimations [55].…”

Streamer and surface charge dynamics in non-uniform air gaps with a dielectric barrier

“…The accuracy of the velocity computation for a streamer propagating along a dielectric surface is sensitive to mesh size. An acceptable accuracy level has been observed for mesh size in the range of 1 µm [14]. Reduction of the mesh size below Figure 7 Electric field strength along streamer propagation path for geometry G2, U = 35 kV: a) t = 0 ns (background field), b) t = 0.3 ns (streamer initiated from the blade tip, propagating towards the barrier), c) t=7 ns (streamer propagation along barrier) d) t = 16 ns, e) t = 59 ns, f) t = 80 ns (streamer at the barrier edge) g) electric field after removing the space charge and restarting the simulation at 80 ns (the surface charge is shielding the blade tip, see section 4.5), The field is only plotted in regions where |E| ≥ 2.6 kV/mm.…”

“…Drift-diffusion models are widely used to simulate streamers in air [4,7,8,[11][12][13][14][15][16][17][18][19]. They are based on a set of driftdiffusion-reaction continuity equations describing the evolution of each type of charged species i:…”

A computational study of negative surface discharges: Characteristics of surface streamers and surface charges

“…However, we expect that this approximation, which was also used in e.g. [1,22,15], has no strong effect for the transient (non-equilibrium) simulations presented here. The minimum grid spacing Δx used for the adaptive mesh is about 1.2 μm.…”