Study On The Method Of Gears' Fault Diagnosis In Variable Conditions

As the main transmission part in mechanical equipments,the gears are easy to break down in the gearbox because their working environment is very complex and.The traditional method of diagnosing gears' faults is usually used in the condition that gears run in a steady state,but the gears usually run in unstable condition.The method of the gears' fault diagnosis in variable condition is studied in this paper.First of all,through detailed analyzing the characteristic of gears' time domain signals,the gears' signal acquisition system in variable condition is designed based on the order analysis,and the method of order analysis is used to transform the time domain signal of variable condition into the angle domain signal.Secondly,the EMD method of the improved mirror continuation is used to decompose the gears' angle domain signal in variable condition,then the decomposed results are flitted by the correlation coefficient method to improve signal-to-noise ratio.Thirdly,according to the results that the order spectrum is demodulated by Hilbert demodulation,the location of fault is acquired to prepare for extracting the feature of the gear fault in variable condition.Finally,the fault diagnosis system based on BP neural network is established to identify different gear state,which selects the energy values of different orders containing fault component as input feature vectors and identifies the types of gear system according to the output value.The experiments show that the improved EMD method and the correlation coefficient method in this paper can be used to improve the noise ratio for the angle domain signal.The model based on the BP neural network can effectively identify the type of gears' faults.