跨声速压气机转子通用k-@湍流模型的机器学习辅助优化

打开文本图片集

关键词：跨声速轴流压气机转子；湍流模型；机器学习；气动性能中图分类号：TKO5文献标志码：ADOI：10.7652/xjtuxb202601018 文章编号： 0253-987X（2026）01-0180-11

Machine Learning-Assisted Optimization of Generalized k-@ Turbulence Model for Transonic Compressor Rotors

FAN Zhixiang，ZHANG Xiawen，LI Zhen，JU Yaping，ZHANG Chuhua （School of Energy and Power Engineering，Xi'an Jiaotong University，Xi'an 7loo49，China）

Abstract： To address the common issue of insuficient computational accuracy in aerodynamic characteristics and complex flows under of-design conditions of compressors using existing eddy-viscosity turbulence models in Reynolds-averaged Navier-Stokes （RANS） methods，the generalized kω （GEKO） turbulence model introduces six free parameters that can be calibrated for different types of flows，improves the computational accuracy of internal flows in compressors. This study focuses on machine learning and multi-objective optimization of the free parameters of the GEKO turbulence model for a transonic axial compressor rotor（NASA Rotor 67） under three operating conditions： near-choke，design point，and near-stall. The computational accuracy of the optimized GEKO turbulence model is compared with that of the default GEKO turbulence model and the standard shear stress transport （SST） turbulence model. The results show that compared to the default GEKO turbulence model，the optimized GEKO turbulence model significantly improves computational accuracy in terms of aerodynamic performance curves， spanwise parameter distributions at the rotor outlet plane，and flow field details. Compared to the SST turbulence model，the optimized GEKO turbulence model provides more accurate performance curve calculations， with a 5.4% improvement in stall margin calculation accuracy. Additionally， the optimized GEKO turbulence model achieves higher accuracy in calculating shock wave locations and intensities. The GEKO turbulence model demonstrates significant effectiveness in enhancing the numerical simulation accuracy of transonic axial compressor rotors， providing a reference for the development of high-precision RANS turbulence models.

Keywords： transonic axial compressor rotor; turbulence model; machine learning；aerodynamic performance

压气机作为航空发动机的核心部件之一，承担着为发动机提供高压工质的重要任务，了解和认识压气机内部的复杂流动，对于高性能压气机设计至关重要。（剩余13931字）