RAPIT (Rigorous Advanced Plasma Integration Testbed): A Parallel Scientific Computational Platform

Jong-Shinn Wu

Department of Mechanical Engineering

National Chiao Tung University


    Many important and challenging science and engineering problems require modeling of complex plasma and flow physics applying hybridization of different continuum- and/or particle-based solvers. Examples may include plume analysis of reaction control thrusters on upper-stage rocket and satellite in orbit, rocket plume analysis at high altitude, aerodynamic analysis of atmospheric-pressure dielectric barrier discharge (DBD) actuator, radical distribution of atmospheric-pressure plasma jet, ion thruster plume analysis, and plasma distribution in etching and thin-film deposition chambers at low pressure, to name a few. These studies often utilize independent solvers developed previously and integrate them in a non-self-consistent approach, which makes their applications and future extension highly inflexible. Thus, a highly flexible simulation platform, which allows straightforward addition and integration of different solvers with a self-consistent approach while maintaining efficient computations, is strongly needed to tackle some challenging problems with complex plasma/flow physics. In this talk, I will report our recent development of a new C++ object-oriented multi-physics simulation platform named Rigorous Advanced Plasma Integration Testbed ( RAPIT) using unstructured-grid finite-volume method with parallel computing through MPI (message passing interface) on distributed-memory PC clusters. The proposed RAPIT with both embedded PDE and particle solver related objects can easily accommodate continuum- and/or particle-based solvers with some proper hybridization algorithm in a self-consistent way. For the former, it may include, but not limited to, the Navier-Stokes (NS) equation solver for general gas flow modeling and the plasma fluid modeling code for general low-temperature plasma modeling. For the latter, it may include the particle-in-cell Monte Carlo collision (PIC-MCC) solver for very low-pressure gas discharge simulation and the direct simulation Monte Carlo (DSMC) solver for rarefied neutral gas flow modeling. Many distinct features of RAPIT include single or multiple mesh(es) for different solvers or species with automatic interpolation relation, essentially the same source code for 2D and 3D problems due to nearly operator-like programming style, and embedded parallel implementation, among others. Some preliminary results of DSMC, PIC-MCC and NS equation and fluid modeling solvers in many practical engineering problems will be presented in this talk. In addition, a byproduct of RAPIT, ultraMPP (ultra-fast Massively Parallel Processing), which is a parallel computing platform for PDE related solvers, will also be briefly introduced. It is designed to greatly reduce the development time of parallel 2D/3D codes from years to weeks while maintaining a highly manageable and consistent source coding framework for researchers. Some major findings along with outlook are summarized at the end of this presentation.
Remark: Co-author
Y.~M.Lee and M.~H. Hu
Plasma Taiwan Innovative Corp.
Juh-bei City, Hsunchu County, Taiwan