Research On Fault Feature Extraction And Optimal Parameters Selection For Gear

In rotating machinery,gear is a very important component,and it is also one of the components which has a high failure rate.The faults of gear may cause a particularly serious impact on the mechanical equipment,and will bring considerable economic and security losses.Therefore,it is very important for mechanical equipment maintenance to detect and eliminate the gear faults in time.In the diagnosis of mechanical equipment,the most important issue is feature extraction.Then,the optimal parameters selection is a key problem,which affects the accuracy of fault detection and diagnosis directly.In this paper the gear are took as the research object,the main researches include as follows:1.Gear fault feature extraction method based on VMD and average energy.Compared with EMD method,VMD method can effectively solve the problem of mode mixing and has high computational efficiency.Therefore,this paper proposes a feature extraction method based on VMD.Firstly,use VMD method to decompose the vibration signal of the gear.Secondly,calculate the average energies for K modal components,and then get a character vector with K dimension.The QPZZ-II fault simulation platform which was produced by Jiangsu Qianpeng limited company is used to verify the proposed method.The results show that fault diagnosis accuracy rate based on proposed extracted features method reached 100%.2.Feature parameters selection method based on scatter degree.In order to effectively select a few optimal characteristic parameters from many characteristic parameters,according to “Maximum between class scatter,and minimum within class scatter.” a evaluation index is proposed based on the scatter.Aiming at providing the basis for parameter selection.In order to verify the effectiveness of the proposed method,the EMD,VMD,and wavelet packet are used to decompose the original vibration signals.And then the parameters of the time-domain,information entropy and frequency-domain are computed.Secondly,the indexes of each characteristic parameter extraction by using the evaluation index proposed in this paper arecalculated.Using the proposed method in this paper.After that,the optimal parameters are determined by the set threshold.The validity of the optimal feature parameters is verified by using minimum distance based on Euclidean distance.The simulation results show that the selection method of optimal feature can efficiently select the optimal parameters.