测控技术2025，Vol.44Issue(2)：32-38,7.DOI:10.19708/j.ckjs.2024.06.223
基于TSA-WT-SVD的汽车转向助力泵转子裂纹声发射检测研究
Research on Acoustic Emission Detection of Cracks in Automotive Power Steering Pump Rotors Based on TSA-WT-SVD
摘要
Abstract
Aiming at the problems of strong subjectivity and high missed detection rate in the current magnetic particle inspection of cracks in the rotor of automotive power steering pump,a rotor crack defect online non-de-structive testing method based on acoustic emission(AE)technology is proposed.Firstly,two types of rotors with crack defects and no defects are assembled in the stator of the same model of double acting vane pump,and a full load test is conducted on a hydraulic test bench with loading parameters of 1 000r/min and 14 MPa.AE sensors are placed at four different positions on the pump body,and the signal sampling rate is set to 3 MHz.Then,for the AE signals collected under two types of rotors,20 segments are randomly selected from the origi-nal signals of each channel for a duration of 100 ms.For each AE segment,the time domain synchronous aver-age(TSA)algorithm is first used to improve the signal-to-noise ratio.Then,a 5-layer wavelet decomposition is performed to obtain the characteristic waveforms of the AE signal in different frequency bands.The matrix com-posed of 5 detail signals is then subjected to singular value decomposition,with the first 5 singular values used as standard features.Finally,for the rotor with defects and no defects and flawless rotors,the Mahalanobis dis-tance between the singular value vector of each AE segment and the standard feature matrix is calculated,and the minimum distance is used as the classification basis.The research results indicate that when the AE sensor is placed near the oil outlet,the accuracy of the TSA-WT-SVD method proposed in identifying rotor cracks is o-ver 95％.关键词
声发射/裂纹缺陷/域同步平均技术/小波分解/奇异值分解/马氏判别Key words
acoustic emission/crack defect/time domain synchronous averaging/wavelet decomposition/singu-lar value decomposition/Mahalanobis discrimination分类
计算机与自动化引用本文复制引用
金丹,裘杭锋,刘冬,方赛银,李明..基于TSA-WT-SVD的汽车转向助力泵转子裂纹声发射检测研究[J].测控技术,2025,44(2):32-38,7.基金项目
国家自然科学基金(32160345) （32160345）
