The Application Of Singular Value Decomposition Denoising And Permutation Entropy Method In Gear Fault Diagnosis

Gear is a universal transmission part among machinery equipments,which is widely used in modern machinery.The gear fault occurs easily because of its complex structure and poor working conditions,and the gear is a consumable part.80% of the faults are related to the gear faults among transmission machinery including gear equipments.In all parts of the gearbox,the fault proportion of the gear itself is maximum,and the failure rate is more than 60%.Once the gear fault occurs in the machinery equipments,it is possible to make production interrupted,lead to huge economic losses of the enterprise and catastrophic consequences which are harmful to life.So the further in-depth study of mechanical fault diagnosis method for gear is of great practical significance.In order to reduce the influence of noise on gear fault diagnosis,SVD de-noising method is adopted to improve the accuracy of fault diagnosis in this paper.First of all,this paper describes the theoretical basis of SVD de-noising method based on Hankel matrix.Common three kinds of singular value threshold processing methods are used,and the experimental test is carried out on the vibration fault test platform of rotating machinery,then the gear fault vibration signals can be obtained.The effect of de-noising for Singular Value mid-value method is found to be more superior than the other two methods by the comparison of Signal-to-Noise Ratio and Root Mean Square Error of three methods and time and frequency domain analysis,and the effectiveness and feasibility of SVD de-noising method in gear fault diagnosis is verified.Secondly,the permutation entropy method is used to extract the characteristic information of gear fault,whose calculation is simple and fast.It means that permutation entropy is introduced in fault diagnosis of gear.The permutation entropy algorithm and its process are described in detail,the characteristic of permutation entropy algorithm is analyzed,phase space reconstruction of the experimental signal is carried out by MATLAB software,and the calculated permutation entropy value is used as the fault eigenvector of the gear,whose anti-noise performance and mutation detection effects are good.The results prove that permutation entropy could represent the change of state of gears.Finally,this paper uses the support vector machine as the intelligent diagnosis method,studies the multi-class SVM classifier.The multi-class SVM classifier is built by adopting common several kernel functions.Permutation entropy eigenvector of gear data in different status respectively combines with the multi-class SVM classifiers to train,typical fault modes of the gear are obtained,thus fault diagnosis and classification of gear can be carried out.The classification effect of themulti-class SVM classifier which is built by adopting radial basi Kernel function is better than the other two kernel functions by the above comparative analysis.At the same time,the combination of permutation entropy and neural network and the combination of permutation entropy and support vector machine(SVM)are used to diagnose the gear faults,and it is verified that the method of the combination of permutation entropy and support vector machine(SVM)is more accurate for the fault diagnosis of gears from the results of the comparative analysis.