Research On Fault Diagnosis Method Of Gear Box Of Rotor System Based On Starting Current

Gearbox,as a common rotating machinery equipment for adjusting speed and transmitting torque,has been widely used in industrial production.The normal operation of gearbox is a prerequisite for the safe and stable production of rotating machinery.Therefore,it is of great significance to diagnose the fault of gears as early as possible.In actual production,the traditional vibration signal analysis method and noise analysis method are difficult to implement because of inconvenient installation of sensors and susceptibility to background noise interference.The motor current signature analysis method is a gear fault diagnosis method with the advantages of high signal-to-noise ratio,high information integration and convenient information acquisition.During the operation of many large rotating units,the rotating speed often increases or decreases because of the great changes of the working state.Therefore,it is necessary to extract and analyze the fault characteristics of motor current under the changed speed circumstance.The motor starting current signal is taken as an example,to extract the fault features of the equal angle resampling signal to identify typical gearbox faults.The main contents and conclusions are as follows:Firstly,the fault diagnosis method of rotor system based on motor current information is introduced.Then,the variation law of stator current caused by gear fault of rotor system is explored.The mathematical model of load and motor current of rotor system and the mathematical model of different gear fault excitation are established.Next,based on the motor mathematical foundation,the integrative coupling model of speed motor and gear torque output is established in Matlab/Simulink module.Then,motor current signal is collected to identify gear faults.Secondly,taking the starting current as an example,the feature extraction method of current signal under variable speed is studied.Starting current as a typical time-domain non-stationary signal,the traditional time-domain and frequency-domain analysis methods cannot effectively extract its fault characteristics.Therefore,a time-frequency domain analysis method based on equal angle resampling is proposed,which first transforms the non-stationary signal in time domain into the stationary signal in angle domain,then extracts and analyses the features in time and frequency domain.Considering the inconvenience of installation of tachometer and photoelectric encoder,an angle resampling method without tachometer based on instantaneous frequency estimation is chosen here.Based on this method,the angle domain resampling of starting current signal is finally finished.Then the order analysis and time domain characteristic parameter analysis of resampling signal are carried out,and the gear fault diagnosis is preliminarily completed.Finally,the method of pattern recognition for gear fault is studied.Sensitive time-domain feature parameters of resampled current signal are selected by feature evaluation method.SVM classification method based on statistical learning theory and BPNN classification method based on knowledge are used for fault pattern recognition,and the gear fault recognition rate is analyzed and compared.The results show that GA-BPNN has better adaptability and higher accuracy when identifying gear fault with characteristic parameters of motor starting current signal.The comprehensive fault recognition rate is 90.1%,of which 96.32% is for broken fault,81.51% is for wear fault and 92.47% is for fault-free gear.