Unified Flow Solver Combining Boltzmann and Continuum Models for Simulations of Gas Flows for the Entire Range of Knudsen Numbers V. V. Aristov 1 , A. A. Frolova 1 , S. A. Zabelok 1 , V. I. Kolobov 2 , and R. R. Arslanbekov 2 1 Dorodnicyn Computing Center of the Russian Academy of Sciences, Vavilova str. – 40, 119991, Moscow, Russia aristov@ccas.ru, frolova@ccas.ru, serge@ccas.ru 2 CFD Research Corporation, 215 Wynn Drive, Huntsville, Alabama 35805, USA vik@cfdrc.com, rra@cfdrc.com 1 Introduction We have developed a Unified Flow Solver (UFS) for simulations of gas flows for the entire range of Knudsen numbers from rarefied to continuum regimes [1, 2]. The UFS methodology is based on the direct numerical solution of the Boltzmann equation for rarefied flow domains [3, 4] and the kinetic schemes of gas dynamics for the continuum flow domains [5, 6]. This approach enables an easy coupling of kinetic and continuum solvers because similar numerical techniques are used for solving both the Boltzmann and continuum equations. The UFS can separate rarefied and continuum domains and use appropriate solvers to combine efficiency of continuum models with accuracy of kinetic models. Domain decomposition criteria and coupling algorithms are important part of the UFS. The computational framework of the UFS is provided by a tree-based data structure of the GERRIS Flow Solver (GFS) [7] enabling dynamically adaptive Cartesian grid with support of complex boundaries. The UFS can automatically generate Cartesian mesh around embedded boundaries defined through standard files, perform dynamic adaptation of the mesh to the solution and geometry, detect kinetic and continuum domains and select appropriate solvers based on continuum breakdown criteria. Parallelization of the UFS has been performed enabling dynamic load balancing for multi- processor systems. This paper presents solutions of several test problems for supersonic and subsonic flow regimes to illustrate current capabilities of the UFS for different applications. UFS extensions to multi-component mixtures of polyatomic gases are also discussed. H. Deconinck, E. Dick (eds.), Computational Fluid Dynamics 2006, DOI 10.1007/978-3-540-92779-2 113, c  Springer-Verlag Berlin Heidelberg 2009