Improved Local Oscillatory-characteristic Decomposition Method And Its Application In Fault Diagnosis Of Rotating Machinery

Nowadays,with the increasing complexity of large rotating machinery equipment,the real-time online monitoring of its running state has important practical significance.It is the key of fault diagnosis to extract the characteristic information which can represent the running state of machinery from the fault vibration signal.Signal analysis and processing technology is a common method to extract rotating machinery fault characteristic information.Because of most mechanical fault vibration signals with the characteristics of non-linear,non-stationary and low signal-to-noise ratio,the time-frequency analysis method can provide the time-domain and frequency-domain information of non-stationary signals at the same time,and has strong local description ability,so it is very suitable for processing the fault vibration signals of rotating machinery equipment.However,common time-frequency analysis methods,such as Windowed Fourier transform(WFT),Wigner distribution(WD),Wavelet transform(WT),Hilbert-Huang transform(HHT)and Local mean decomposition(LMD),etc.,all have their own limitations.Therefore,it is necessary to explore new techniques for analyzing and processing vibration signals of mechanical faults.Recently,after a new adaptive time-frequency analysis method Local oscillatory-characteristic decomposition method is proposed,it has been successfully applied in the fault diagnosis of rotating machinery due to its high efficiency of operation speed Based on this method,the paper proposes a new adaptive time-frequency analysis method-Rational spline-Local oscillatory-characteristic decomposition(RS-LOD)The main research contents are as follows:(1)The rational spline-Local oscillatory-characteristic decomposition method is proposed,which can decompose multi-component non-linear,non-stationary signals into smooth mono-component signals,so as to solve the problem of lack of smooth distortion of mono-oscillation component obtained from the original LOD decomposition.And the RS-LOD method and Hilbert demodulation method are combined to extract the fault characteristic information of rotating machinery effectively(2)Aiming at the problem of over envelope and under envelope in the empirical envelope method(EE)of mono-component signals demodulation,an improved empirical envelope method(IEE)is proposed,and compared with Hilbert demodulation method and Teager energy operator demodulation method.At the same time,combined RS-LOD and IEE demodulation method,the fault feature information of rotating machinery is extracted effectively.(3)In view of the problem that it is difficult to extract the early weak fault of rolling bearing due to the interference of background noise and its installation and manufacturing error,a time-frequency method based on RS-LOD and weighted derivative dynamic time warp method(WDDTW)is proposed.And the validity of the method is verified by analyzing the vibration fault signal of rolling bearing in the bearing data center of the University of Western Reserve in the United States.