Numerical analysis of the unsteady behavior of cloud cavitation around a hydrofoil based on an improved filter-based model

• Published in: Journal of hydrodynamics. Series B Vol. 27; no. 5; pp. 795 - 808
• Main Author: 张德胜 王海宇 施卫东 张光建 Van ESCH B.P.M.（Bart）
• Format: Journal Article
• Language: English
• Published: Singapore Elsevier Ltd 01.10.2015; Springer Singapore; Research Center of Fluid Machinery Engineering and Technology, Jiangsu University, Zhenjiang 212013, China%Department of Mechanical Engineering, Eindhoven University of Technology, Eindhoven 5600MB, The Netherlands
• Subjects: cloud cavitation; density correction method; Engineering; Engineering Fluid Dynamics; filter-based model (FBM); hydrofoil; Hydrology/Water Resources; Numerical and Computational Physics; Simulation; unsteady behavior; 基础; 数值分析; 校正方法; 模型; 水翼; 滤波器库; 空化数; 非定常特性; cloud cavitation; hydrofoil; density correction method; filter-based model (FBM); unsteady behavior
• DOI: 10.1016/S1001-6058(15)60541-8
• ISSN: 1001-6058

The unsteady cavitation evolution around the Clark-Y hydrofoil is investigated in this paper, by using an improved filter-base model（FBM） with the density correction method（DCM）. To improve the prediction accuracy, the filter scale is adjusted based on the grid size. The numerical results show that a small filter scale is crucial for the unsteady simulations of the cavity shedding flow. The hybrid method that combines the FBM and the DCM could help to limit the overprediction of the turbulent viscosity in the cavitation region on the wall of the hydrofoil and in the wake. The large value of the maximum density ratio ρ1 /ρv, clip promotes the mass transfer rate between the liquid phase and the vapor phase, which results in a large sheet cavity length and the vapor fraction rise inside the cavity. The cavity patterns predicted by the improved method are verified by the experimental visualizations. The time-average lift, the drag coefficient and the primary oscillating frequency St for the cavitation number σ= 0.8, the angle of attack, α= 8°, at a Reynolds number Re= 7×10^5 are 0.735, 0.115 and 0.183, respectively, and the predicted errors are 3.29%, 3.36% and 8.93%. The typical three stages in one revolution are well-captured, including the initiation of the sheet/attached cavity, the growth toward the trailing edge（TE） with the development of the re-entrant jet flow, and the large scale cloud cavity shedding. It is observed that the cloud cavity shedding flow induces the vortex pairs of the TE vortices in the wake and the shedding vortices. The positive vorticity vortex of the re-entrant jet and the TE vortices interacts and merges with the negative vorticity vortex of the leading edge（LE） cavity to produce the shedding flow.