Abstract
In diesel engine fault diagnosis, non-stationary vibration signal is easily disturbed by strong noise. In view of the shortcomings of empirical mode decomposition (EMD) and wavelet transform in de-noising, a de-noising method is proposed, which is Empirical Wavelet Transform (EWT) autocorrelation analysis. Taking advantages of EMD and wavelet transform, the Fourier spectrum is adaptively divided by EWT, and the intrinsic mode components of different frequency are extracted through constructed wavelet filter, and the method can effectively eliminate modal aliasing and solve adaptive problems in wavelet de-noising. At the same time, autocorrelation analysis can make the random noise decay to zero, and the modal components with high frequency random noise are dealt by autocorrelation analysis. The method is used to de-noising and compared the de-noising effect of EWT and EMD. Results show that the method can effectively decompose the intrinsic mode component with less number, and there is no false mode, and the de-noising effect is better than EMD de-noising. The method is feasible and effective in de-noising by vibration signals of diesel engine.
Highlights
The complex motion of diesel engine often brings a lot of noise, which makes the useful information of vibration signal disappear in the background noise
In the area B and C, a vibration sensor is installed on top of the 6# cylinder head and the other one is installed on top of the 1# cylinder head respectively to collect vibration signals of the diesel engine
The method of vibration signal denoising based on Empirical Wavelet Transform (EWT) autocorrelation analysis is proposed
Summary
The complex motion of diesel engine often brings a lot of noise, which makes the useful information of vibration signal disappear in the background noise. Wavelet transform and Empirical mode decomposition have some applications in the de noising of vibration signals [1,2,3], there are some shortcomings in the wavelet denoising method, such as the effect of wavelet denoising and the selection of wavelet base, which makes the wavelet denoising not adaptive. Mourad Kedadouche [9] et al put forward a new method of bearing fault diagnosis based on OMA and empirical wavelet transform. Jinglong Chen [10] et al put forward the application of empirical wavelet transform based on measured vibration signals in fault diagnosis of wind turbine generator bearings. A vibration signal denoising method based on empirical wavelet transform autocorrelation analysis is proposed. The effectiveness of denoising is verified by simulated and measured signals
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
