Gearbox Fault Diagnosis Method Based On Sequential Testing Theory Research

The function of the gear is mainly reflected in the process of gear box transmission power,the fault diagnosed of gear box is diagnosing the fault of rolling bearing,gear,rotor and other components.Vibration and noise diagnosis is the most common and most effective method of gear fault diagnosis,the characteristics of gear box failure often hidden in the vibration and noise signal,through the analysis of vibration and noise signal and get the gear box fault reasons.The purpose of this paper is to study around the diagnosed of the vibration and noise signals in the gear box.Firstly, A comprehensive analyzed of the particle filtering, one of the filtering method, is done.It is pointed out that the method has the defects of poor real-time this inherent and puts forward the solving methods:The radial basis function network(RBFN) learning method is introduced into particle filter method which means clustering all the particles of the focusing particle in the sampling time,timely updating the states of particles,improving the estimation accuracy,eliminating the errors caused by process noise to achieve the goal of improving the performance of the particle filter.Secondly,the idea of putting the use of sequential probability ratio test algorithm in fault diagnosis is discussed,and it is verified.After checking,it is proved that selecting gear crack on diagnostic analysis is not only feasible but also more efficient:using the sequential probability ratio test algorithm to test the vibration signal managed by particle filter and combing the root mean square error algorithm not only will distinguish different carck of gear but also will classify the fault.Thirdly, the gear’s crack fault diagnosis method based on parameter sequential probability ratio test is to compare each hypothesis test data with the set value.With no predetermined observation sample population is its advantage.So,put forward a new method that combined particle filter optimized by RBF network with sequential probability ratio test:firstly,using particle filter method optimized by RBF network to denoise the signals to obtain stable signal,then regarding kurtosis value which is highly sensitive to the impact vibration as the characteristic parameter and using time domain analysis method to deal with signals to denoise processing numerical characteristics,finally using the sequential probability ratio test algorithm to analysis the fault of gear crack.The experimental research shows that the new method presented in this paper is effective and reliable for the fault diagnosis of gear crack.