Li Shunming1,Xu Qingyu2
(1.College of Energy and Power Engineering,Nanjing University of Aeronautics and Astronautics,Nanjing 210016,China;2.School of Architectural Engineering and Mechanics,Xi'an Jiaotong University,Xi'an 710049,China)
Abstract:Some weak vibration signals are very useful for equipment inspection and fault forecast but are difficult to extract by Fourier transform or dyadic wavelet transform. In comparison with dyadic wavelet,harmonic wavelet has such excellent properties as easy to construct,simple expression,phase holding characteristic,and clear relation with Fourier transform.It is found that the weak partial signal can effectively be extracted by harmonic wavelet transform in frequency domain and can easily be reconstructed.Moreover,the weak cycle signal can be extracted from strong noise by harmonic wavelet transform in frequency domain.Computation and engineering results show that the extraction ability of weak signal and the extraction precision of harmonic wavelet transform in frequency domain are higher than that of dyadic wavelet transform and Fourier analysis method.The noise leak problem in weak signal extraction by dyadic wavelet package and Fourier transform is eliminated by harmonic wavelet transform.
Keywords:weak signal;vibration signal;harmonic wavelet;extract