An automatic recognition method for characteristic frequency of rolling bearings

In order to solve problems, such as, overdependence on single fault characteristic frequency, and the ambiguity and inefficiency of diagnosis results caused by subjective factors in fault diagnosis of rolling bearings, a method using the envelope spectrum and resonance demodulation analysis of signals was proposed to identify the fault characteristic frequency, and its multiplications and modulation frequency components of rolling bearings. Firstly, the original signal was analyzed with the envelope spectrum or resonance demodulation analysis. Then, a specific algorithm was used to identify faulty frequency components and rotating frequency in the spectrum. At last, the fault diagnosis of rolling bearings was conducted with the corresponding proportions of the identified frequency components. The fault diagnosis simulations and the life acceleration tests of rolling bearings demonstrated the validity of the proposed method.