不连续性分类
变量(数学)
系列(地层学)
应用数学
计算
流量(数学)
算法
计算机科学
非线性系统
湍流
休克(循环)
数学优化
数学
数学分析
物理
机械
几何学
医学
古生物学
量子力学
内科学
生物
作者
Xuan Liu,Meiyuan Zhen,Jinsheng Cai,Fei Liao
摘要
The WENO family schemes have been widely used in various compressible turbulence flow simulations due to its excellent shock-capturing capability and high resolution. However, due to the massive calculation is needed by WENO scheme, the variable-based flux is usually used in engineering numerical calculation, rather than characteristic-based reconstruction that require matrix operation. For problems containing strong shocks, variable-based reconstruction may produces numerical oscillations, while characteristic-based reconstruction rarely do. In this paper, a novel scale separation mechanism without free parameter is proposed to construct an efficient hybrid WENO scheme, i.e. WENO-H, by which the characteristic-based or variable-based reconstruction can be accurately selected for different region of the flow, so as to obtain a higher resolution and more stable shock capturing scheme while improving the efficiency. According to the novel scale separation mechanism the new scheme performs characteristic-based reconstruction near discontinuities and switches to variable-based reconstruction for smooth region. Linear variable-based fluxes with less computation are used directly in the smooth region, while nonlinear WENO-Z and more robust characteristic-base fluxes are used in the discontinuous region. Several one dimensional and two dimensional numerical tests are performed to validate and evaluate the scheme. Numerical results shows that WENO-H series schemes maintain essentially non-oscillatory flow filed near discontinuities. Besides, compared to the WENO-Z, the WENO-H series schemes are 10% faster in 1D problem and 30% faster in 2D problem, and saved more than 1.6 times computational cost compared with TENO.
科研通智能强力驱动
Strongly Powered by AbleSci AI