跨音速
噪音(视频)
计算机科学
大涡模拟
计算流体力学
航空航天工程
机械
物理
空气动力学
工程类
人工智能
湍流
图像(数学)
作者
Guillaume A. Brès,Kan Wang,Christopher Ivey,Sanjeeb Bose,Gaku Okubo,Toshihiro Kamatsuchi,Hisato Tanaka
摘要
GPU-accelerated wall-modeled large-eddy simulations of the NASA Source Diagnostic Test (SDT) fan benchmark are performed at transonic conditions for three different outlet guide vane (OGV) configurations, i.e., baseline, low-count and low-noise. The full-wheel, whole rotor/stator stage and the nacelle are simulated at cut-back and take-off conditions where the fan operates at 87.5% and 100% of the design rotational speed. Fan and stage performance predictions are compared to available experimental measurements and show good agreement in terms of total pressure and temperature ratio, as well as adiabatic efficiency. Inlet, exhaust and total sound power levels are also predicted. The comparisons with microphone data show that both broadband and tonal noise are well captured, with the exception of the low-frequency Multiple Pure Tone (MPT) or "buzz-saw" noise associated with small blade-to-blade geometrical variations for fan with supersonic rotor tip condition. An additional simulation was performed at the high rotational speed for a slightly modified fan geometry with randomly-staggered blades similar to the variations in the installed SDT fan: with this modification, the "buzz-saw" noise is now present in the LES results and leads to closer agreements with the acoustic measurements.
科研通智能强力驱动
Strongly Powered by AbleSci AI