An Efficient Difference Scheme for Gas Dynamics Equations
Jianhu Feng,Li Cai,Wenxian Xie
DOI: https://doi.org/10.3969/j.issn.1000-2758.2005.02.017
2005-01-01
Xibei Gongye Daxue Xuebao/Journal of Northwestern Polytechnical University
Abstract:Existing difference schemes for gas dynamics equations such as WENO (weighted essentially non-oscillatory) scheme and CWENO (central WENO) scheme, in our opinion, still leave much room for further improvement in efficiency. We present a difference scheme that is much more efficient than WENO and CWENO schemes at the cost of a little lowering in resolution (any difference scheme of end-order or higher is usually considered to belong to the high-resolution class; our proposed second-order scheme is a little inferior to the 4th-order WENO and CWENO schemes in high resolution). We present a second-order semi-discrete central-upwind scheme for multidimensional inviscid compressible Euler equations. Taking into consideration the local speed of nonlinear wave propagation in Riemann fans, we calculate the widths of the local Riemann fans more accurately, obviate the necessity of staggering between two sets of grids, and obtain a scheme with much smaller numerical viscosity. The high-resolution method is just a semi-discrete central-upwind scheme connecting upwind scheme and central scheme. At the same time, the piecewise linear polynomial is reconstructed by a convex combination of the dissipative MinMod limiter by Tadmor et al and the compressive UNO limiter by Harten et al. We checked the feasibility of our method with three numerical simulation examples. Two numerical examples (four shock-wave Riemann problem and single-shock, single-rarefaction, two-contact-discontinuity problem) were the same as those in Ref. 12 by Kurganov and Tadmor, which gives high-resolution results with 3rd-order difference scheme. For the four shock-wave Riemann problem, results obtained with our efficient 2nd-order difference scheme agree with those in Ref. 12 and even the precision obtained is about the same as that in Ref. 12. For the single-shock, single-rarefaction, two-contact-discontinuity problem, our numerical results also agree with those in Ref. 12.