用连续小波变换识别内燃机噪声源  

Using continuous wavelet transform to identify internal combustion engine sound source

作　　者：杨金才[1] 郝志勇[1] 贾维新[1] 

机构地区：[1]浙江大学机械与能源工程学院,浙江杭州310027

出　　处：《浙江大学学报（工学版）》2006年第3期404-407,428,共5页Journal of Zhejiang University：Engineering Science

摘　　要：为了直接利用连续小波变换进行内燃机噪声源识别,提出了一种连续小波变换的改进算法.研究了信号组成成分的幅值和频率变化对小波系数的影响,根据连续小波变换的基本性质,对连续小波变换后产生的与信号频率有关的衰减系数进行修正,通过对比分析从Matlab6.5的小波工具箱中找出了适合振动噪声信号分析的小波函数.工程信号的分析结果表明,经过对连续小波变换后小波系数的修正,使小波系数的大小能正确反映信号各组成成分的幅值大小,从而可以利用小波系数形成的云图直接进行内燃机噪声源识别.To identify the sound sources of internal combustion engine by directly using continuous wavelet transform （CWT）, an improved algorithm of CWT was proposed. The influence to wavelet coefficients caused by the change of signal component amplitude and frequency was investigated. According to the base properties of CWT, the wavelet coefficients reduced by the coefficient relating to the signal frequency were corrected. The suitable wavelet functions for vibration noise signal processing were selected out from the wavelet toolbox of Matlab 6.5 by comparing various wavelet functions. The analysis results on engineering signal indicate that the wavelet coefficients corrected after continuous wavelet transform can properly show the amplitude magnitude of various signal components. The map of corrected wavelet coefficients can directly be used to identify the sound sources of internal combustion engine.

关 键 词：连续小波变换 内燃机 噪声源识别 

分 类 号：TK401[动力工程及工程热物理—动力机械及工程]