Research on fault law of rolling bearing under different fault levels and loads with HHT method

Bearing is one of the most important components of rotating machinery. The vibration signals are generally nonlinear and nonstationary while operating. The failed rolling bearing will damage to the machine, or cause a serious loss of property. There are a lot of methods about fault diagnosis of bearing, such as shock pulse method, resonance demodulation. Especially the HHT (Hilbert-Huang Transform) method with the adaptive advantage has gradually become a very promising method to extract the characteristics of nonlinear, nonstationary signal. In this paper the variant energy method was introduced in HHT to reduce the computation of the decomposed signal, which effectively improved the computation, and then an experimental platform was designed and established. The bearing fault categories can be diagnosed correctly in dealing with the vibration signals using this method and the fault law is discovered that the trend of the vibration signal fault characteristic frequency amplitude changes with the load increasing. The bearing failure mechanism provides beneficial reference for further research of nonlinear signal analysis