Research On Fault Diagnosis Method Of Roller Bearings In Annealing Kiln Main Drive System

The roller table of the main transmission system of the annealing furnace is a key component for carrying and pulling glass products.Once a bearing failure occurs,the roller table will lose the smooth running,which will have a significant impact on the quality of glass products,and even cause the entire production line to stop production,which will cause the enterprise Cause huge economic losses.In order to ensure the smooth operation of the roller table of the annealing kiln and reduce the defective rate of glass products,it is of great engineering significance to study the fault diagnosis method of the roller table bearing.Vibration signal analysis method is one of the most widely used and most effective mechanical fault diagnosis methods,it has the advantages of online and non-destructive detection,which is very suitable for the diagnosis needs of don’t stop running disassembly and detection of roller bearings.This paper focuses on the key technologies of roller bearing fault vibration signal noise reduction,feature extraction and evaluation,fault pattern recognition,etc.,and proposes an intelligent diagnosis method for roller bearing faults based on grey relational entropy analysis and sensitive feature evaluation.Designed and developed an annealing kiln roller bearing fault detection and diagnosis system.The main research contents of the thesis are as follows:(1)Aiming at the strong mechanical background noise interference of the roller bearing fault signal of the annealing kiln of the glass production line,a signal noise reduction method based on gray correlation entropy analysis is proposed.This method uses Empirical Mode Decomposition(EMD)to decompose the original fault signal into multiple intrinsic mode function(IMF),uses grey relational entropy analysis to select the IMF components and performs wavelet threshold denoising,and then reconstruct the noise-reduced component and the remaining components to obtain an effective fault vibration signal.The noise reduction results of rolling bearing data and the actual measurement experiments of roller bearings show that the signal-to-noise ratio and root mean square value of bearing fault signals have been significantly improved.The noise reduction method proposed in this paper is simple and effective.(2)Aiming at the problems of low accuracy and poor efficiency in fault diagnosis of roller bearings,an intelligent fault diagnosis method based on sensitive feature evaluation and neural network is proposed.This method uses wavelet packet transform to extract time domain and frequency domain features,defines sensitive feature evaluation factors based on the distance between fault feature classes and within classes,selects the sensitive feature set,and then uses the radial basis function(RBF)neural network to identify the fault feature set.Using rolling bearing data and annealing kiln roller bearing fault test data to verify the distance-based sensitive feature evaluation method,the results show that the sensitive features evaluated and selected effectively reduce the dimensionality of the full feature set and improve the accuracy of fault diagnosis.(3)Aiming at the non-damage detection and fault diagnosis of roller table bearing in annealing kiln,a fault detection and diagnosis system for roller table bearing in annealing kiln is designed according to the fault signal noise reduction method and sensitive feature selection method proposed in this paper.Use the measured signal to demonstrate the function,verify the simplicity and effectiveness of the system fault diagnosis function,and provide technical support for the efficient diagnosis of the roller bearing fault of the main drive system of the annealing furnace of the glass production line.