Fault diagnosis for rolling bearing early fault based on standardization transformation stochastic resonance

It is very hard to detect the early fault of rolling bearing with classical methods when the fault signal energy is too low and the noise is too strong. Stochastic resonance (SR) theory is a method to enhance the weak signal submerged in strong noise. But classic SR is hard applied to practice for large parameters problem. The existing large parameter stochastic resonance models (LPSR) need either high sampling frequency or long sampling data length. A novel method named standardization transformation stochastic resonance (STSR) is proposed in this paper to solve the large parameter problem with low sampling frequency and short sampling data length. The proposed STSR is compared with other two LPSR models by simulation. A novel fault diagnosis for rolling bearing early fault based on STSR is also proposed in this paper. It is applied to detecting the early fault of a deep groove ball rolling bearing successfully. The practical application and the contrast between the other two LPSR methods verify the effectiveness of fault diagnosis for rolling bearing early fault based on STSR.