Study On Fault Feature Extraction Of Vibration Signals For Gear

As a kind of general part of machinery and equipment, gear is one of the parts which is the most prone to go wrong. Thus, the gear works normally plays a very crucial role in the normal operation of machinery and equipment. So, gear fault diagnosis is of great significance.First of all, this paper in view of the problem of gear fault vibration signal contains strong background noise, which is mainly studied in the signal processing method on inhibiting background noise and highlighting fault characteristics. The research based on the vibration mechanism of gear and the reason of the fault shows the signal characteristic of gear signal which is normal and gear owns many kinds of faults.Secondly, energy index is innovatively used in researching the vibration signal of gear, The gear crack fault signal data is took through simulation and the other signal is adding in random noise, which are analyzed by energy index method. The result shows this method is effective to extract gear crack fault signal feature and for fear fault diagnosis.At last, according to the shortage of EMD existing in reducing noise, the difference of general signal and noise signal in autocorrelation function, and correlation coefficient, put forward a kind of noise reduction method based on EMD combines with correlation analysis. This method solved problem how to determine the demarcation point of noise components and useful components, false components and useful components. It is testified by analyzing simulation signal and fault signal measured from gear system, the result shows this method achieves the purpose of noise reduction, which provides a new way for noise reduction.