摘要
分离涡模拟DES是压气机流动模拟中常用的高保真湍流模式。为了使DES准确解析湍流,数值耗散必须限制在合理范围内。然而,当前的压气机流动DES类研究工作中仍然普遍采用高耗散的迎风格式。本文首先基于DES类方法计算的各向同性衰减湍流结果,定量比较了多种不同数值格式的耗散,证实了高耗散迎风格式严重低估中高波数湍流能量。高阶重构格式可以一定程度上改善该问题,但能量耗散仍然过高。本文在高阶重构的基础上,进一步引入自适应耗散函数修改Riemann求解器,构造了自适应耗散格式。该格式在全波数范围都能准确地预测湍流能谱。将该格式配合DES类方法模拟跨声速离心压气机流动,其预测的压比相比于三阶迎风格式,更加接近实验结果。此外,自适应耗散格式显著提高了中小尺度流动结构的分辨率。分析表明,在使用DES类方法模拟压气机流动时,有必要采用数值耗散较低的离散格式,以准确预测压气机总体性能和流动结构。本文构造的自适应耗散格式是一种良好选择。
Detached Eddy Simulation(DES)is a commonly used high-fidelity turbulence model for analyzing compressor flows.Achieving high-resolution turbulence simulations with DES requires limiting numerical dissipation.However,upwind schemes with significant dissipation remain prevalent in DES for compressor flows.This paper examines dissipation of various numerical schemes using a decaying isotropic turbulence test case.The study reveals that the 3rd-order upwind scheme exhibits excessive dissipation,particularly evident in the elimination of turbulent kinetic energy at high wavenumbers in the turbulent energy spectrum.While higher-order reconstruction schemes mitigate this issue,they still exhibit dissipative behavior at higher wavenumbers.To address this,an adaptive-dissipation scheme is introduced,incorporating a modified Riemann solver with an adaptive function.This scheme accurately predicts the turbulent energy spectrum across all wavenumbers.Compared with the outcomes of the adaptive-dissipation scheme and the 3rd-order upwind scheme in DES for a transonic centrifugal compressor,the former demonstrates superior performance.Furthermore,the adaptive-dissipation scheme excels in resolving small coherent structures within the compressor flow.In summary,this investigation emphasizes the importance of selecting numerical schemes with reduced dissipation for DES in compressor flows.Such an approach enhances the accuracy of compressor performance prediction and allows for higher-resolution depiction of flow structures.The presented adaptive-dissipation scheme emerges as a favorable choice for these purposes.
作者
田铖
江思雅
符松
TIAN Cheng;JIANG Si-ya;FU Song(School of Aerospace,Tsinghua University,Beijing 100084,China;CAEP Software Center for High Performance Numerical Simulation,Beijing 100088,China;Institute of Applied Physics and Computational Mathematics,Beijing 100088,China)
出处
《计算力学学报》
CAS
CSCD
北大核心
2024年第1期145-152,共8页
Chinese Journal of Computational Mechanics
基金
国家科技重大专项(2017-II-0004-0016)
国家重点研发计划(2019YFA0405200)
国家重大项目(GJXM92579)资助.