Discrete unified gas kinetic scheme for all Knudsen number flows: II. Compressible case

TL;DR

This paper introduces a discrete unified gas-kinetic scheme (DUGKS) for simulating multiscale compressible flows with heat transfer and shocks, maintaining asymptotic preserving properties and second-order accuracy.

Contribution

It develops an explicit finite-volume DUGKS for compressible flows that couples transport and collision processes, extending previous work to include heat transfer and shock discontinuities.

Findings

DUGKS accurately captures shock structures and non-equilibrium effects.

The scheme demonstrates good agreement with DSMC and benchmark data.

It is efficient for multiscale compressible flow simulations.

Abstract

This paper is a continuation of our earlier work [Z.L. Guo {\it et al.}, Phys. Rev. E {\bf 88}, 033305 (2013)] where a multiscale numerical scheme based on kinetic model was developed for low speed isothermal flows with arbitrary Knudsen numbers. In this work, a discrete unified gas-kinetic scheme (DUGKS) for compressible flows with the consideration of heat transfer and shock discontinuity is developed based on the Shakhov model with an adjustable Prandtl number. The method is an explicit finite-volume scheme where the transport and collision processes are coupled in the evaluation of the fluxes at cell interfaces, so that the nice asymptotic preserving (AP) property is retained, such that the time step is limited only by the CFL number, the distribution function at cell interface recovers to the Chapman-Enskog one in the continuum limit while reduces to that of free-transport for free-molecular flow, and the time and spatial accuracy is of second-order accuracy in smooth region. These features make the DUGKS an ideal method for multiscale compressible flow simulations. A number of numerical tests, including the shock structure problem, the Sod tube problem with different degree of non-equilibrium, and the two-dimensional Riemann problem in continuum and rarefied regimes, are performed to validate the scheme. The comparisons with the results of DSMC and other benchmark data demonstrate that the DUGKS is a reliable and efficient method for multiscale compressible flow computation.