用A计权连续小波变换识别内燃机噪声源

为了适应内燃机噪声声源识别的需要,当用连续小波变换进行信号分析时,常对小波变换算法作适当的改进.依据连续小波变换的叠加性和尺度转换性质,提出了变换后小波系数的频率修正方法,即在修正小波变换时对信号中不同频率成分小波系数的不同衰减,使变换后小波系数的大小能够准确反映信号中不同频率成分的幅值特性,在此基础上提出了A计权连续小波变换算法,对频率修正后的小波系数进行A计权修正,使之更加适合实际工程中声源识别的需要.利用该方法对某发动机进行了噪声源识别,取得了良好的识别效果.

Using A-weighting CWT to identify internal combustion engine sound source

To meet the need of internal combustion engine(IC engine) sound source identification,the algorithm of continuous wavelet transform(CWT) should be improved when it was used to process the noise signals.A frequency modification method of wavelet coefficients was proposed based on the additivity property and scale transform property of CWT.The different attenuation to different frequency components of the signal caused in CWT was corrected.The magnitude of the wavelet coefficients modified by this method can properly show the amplitude property of the signal components with different frequency.Further,an A-weighting CWT algorithm was put forward.The wavelet coefficients after frequency modification were modified once more with A-weighting.This algorithm was more suitable for the sound identification in actual engineering application.An engineering case was studied and good identification results were obtained.