In our actual condition monitoring, we often only need to judge whether the rolling bearing is good or bad and how long it can be used. However, it is often not practical to diagnose the fault of a certain part of the bearing in the precision analysis and diagnosis. Practical diagnostic accuracy due to factors such as working conditions, often can not find the corresponding characteristic frequency of rolling bearings。 Although the wavelet analysis developed in recent years and the rapid resonance vibration demodulation analysis technology are more accurate, but the required equipment investment is larger, more analysis is still needed, on-site fault diagnosis staff generally less application. In the practical diagnosis, we take the combination of the dimensionless and the dimensionless parameters to diagnose the bearing rapid fault, that is, the frequency vibration velocity in the spectrum analysis is combined with the bearing kurtosis to make a comprehensive diagnosis. When both conditions exceed the standard, we judge that the bearing is faulty.
This method of judgment after three years of practice, proved that the fault diagnosis of rolling bearings is very practical. Judge fast, accurate, accurate rate of more than 90%. Diagnosed bearings are basically in the late stage of failure, with very good economic benefits.
In addition, when it is detected that the low-frequency vibration of the rolling bearing is very large, excluding the unit misalignment, unbalance, loose structure and basic structural factors of resonance, even if there is no characteristic frequency of the rolling bearing, the trouble of the rolling bearing should be decisively judged for overhaul。