Two statistical particle split and merge methods for Particle-in-Cell codes

Two new particle split and merge methods are presented for electromagnetic Particle-in-Cell codes. They are based on a statistical approach and are designed to conserve mass, momentum and energy. In addition, the correct spatial resolution of these values and the velocity distribution functions are reproduced. It is shown that charge and current densities before and after splitting and merging are in a good agreement. Furthermore, the reconstruction of the velocity distribution function is demonstrated. The new methods may lead to significantly reduced computational effort and memory consumptions which are exemplarily demonstrated by the 3D simulation of a streamer discharge formation.