航空发动机压气机管道声模态无偏稀疏重构方法  

An unbiased sparse reconstruction method for duct acoustic mode on aero-engine compressor

作　　者：文璧 李泽芃 杜军 王亚南 刘元是 乔百杰 陈雪峰[1] WEN Bi;LI Zepeng;DU Jun;WANG Yanan;LIU Yuanshi;QIAO Baijie;CHEN Xuefeng(School of Mechanical Engineering,Xi’an Jiaotong University,Xi’an 710049,China;Taihang National Laboratory,Chengdu 610213,China;AECC Sichuan Gas Turbine Establishment,Mianyang 621022,China)

机构地区：[1]西安交通大学机械工程学院,陕西西安710049 [2]太行国家实验室,四川成都610213 [3]中国航发四川燃气涡轮研究院,四川绵阳621022

出　　处：《推进技术》2025年第4期279-287,共9页Journal of Propulsion Technology

基　　金：国家自然科学基金(52305127,52475130)。

摘　　要：针对应用声模态分解技术解析压气机管道内截通声模态特征时,传统均匀环形阵列要求传感器数量多、传统稀疏估计方法精度低的问题,本文提出了一种无偏稀疏声模态重构方法,通过L_(1)范数正则化方法实现声模态向量支撑集的求解,再通过最小二乘实现模态幅值的无偏估计,最后分别通过仿真分析和实验研究验证了所提方法的优越性。结果表明:本文提出的无偏稀疏重构方法相对于经典稀疏重构方法显著提高了声模态重构精度以及辨识鲁棒性,相对于L_(1)范数正则化方法在三种不同传感器布局下,主导声模态幅值重构精度分别提升1.74 d B,2.36 d B和0.78 d B;相对于L_(1/2)范数正则化方法具有更好的阶次辨识鲁棒性。To overcome the difficulty of a large number of sensors required for a uniform circular array and low accuracy of traditional sparse estimation methods when applying acoustic mode decomposition technology to analyze the characteristics of intercepted acoustic modes within compressor ducts,this paper proposes an unbi-ased sparse acoustic mode reconstruction method which solves for the support set of the acoustic mode solution vector using L_(1)-norm regularization and achieves unbiased estimation of modal amplitudes through least squares.The superiority of the proposed method is validated through both simulation analysis and experimental research.The results indicate that the unbiased sparse reconstruction method proposed in this paper significantly improves the accuracy of acoustic mode reconstruction and demonstrates better robustness compared to classic sparse re-construction method.Compared to the L_(1) norm regularization method,the reconstruction accuracy of dominant acoustic mode amplitude is improved by 1.74 dB,2.36 dB and 0.78 dB under three different sensor layouts,re-spectively.Compared to the L_(1/2) norm regularization method,it has better robustness in order identification.

关 键 词：压气机 气动声学 管道声模态 声阵列 稀疏重构 正则化 

分 类 号：V231.92[航空宇航科学与技术—航空宇航推进理论与工程]