Supersonic flow past the 24-degree two-dimensional compression ramp of Settles was simulated by employing three popular engineering turbulence models, namely Kok k-ω TNT, k-ω Baseline (BSL) and k-ω Shear Stress Transport (SST). The performance of the three turbulence models was evaluated and summarized. All turbulence models predicted larger separation zone than the experimental results, and the separation zone of Kok k-ω TNT was the largest. Optimization research was done for the k-ω BSL and k-ω SST by changing the turbulence constant σω1, the Bradshaw number α1, the Von Karmen number κ of SST and the turbulence constant σω1, σω2, Von Karmen number κ of BSL. The change of σω1 or σω2 had little influence on the size of the separation. As the accretion of α1 or κ, the simulated separation zone became smaller. Good choice of α1 or κ could improve the separation simulation accuracy. α1 = 0.4 and κ = 0.45 gave better results than the conventional values for SST and the best choice of κ for BSL was 0.44.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

An Assessment of Turbulence Models for Supersonic Compression Ramp Flow

  • Lan Wang,
  • Haixin Ding,
  • Zhouqin Fan

摘要

Supersonic flow past the 24-degree two-dimensional compression ramp of Settles was simulated by employing three popular engineering turbulence models, namely Kok k-ω TNT, k-ω Baseline (BSL) and k-ω Shear Stress Transport (SST). The performance of the three turbulence models was evaluated and summarized. All turbulence models predicted larger separation zone than the experimental results, and the separation zone of Kok k-ω TNT was the largest. Optimization research was done for the k-ω BSL and k-ω SST by changing the turbulence constant σω1, the Bradshaw number α1, the Von Karmen number κ of SST and the turbulence constant σω1, σω2, Von Karmen number κ of BSL. The change of σω1 or σω2 had little influence on the size of the separation. As the accretion of α1 or κ, the simulated separation zone became smaller. Good choice of α1 or κ could improve the separation simulation accuracy. α1 = 0.4 and κ = 0.45 gave better results than the conventional values for SST and the best choice of κ for BSL was 0.44.