Study On State Detection And Fault Diagnosis Of Radial Bearing For Flying Wheel Energy Storage

With the rapid development of energy storage,flywheel energy storage(flywheel energy storage system)has gradually entered people's view.Because of its environmental pollution-free,short-time high power,high efficiency and long life,flywheel energy storage is widely used in enterprises UPS?wind power plants,emergency power supply support,aerospace and other fields,but also because of its high speed rotation characteristics,rolling bearings as the main components become extremely vulnerable components.Therefore,it is of great significance to study the bearing fault diagnosis of flywheel energy storage to ensure the safety,stability and stability of the equipment.Firstly,the significance and importance of bearing fault diagnosis in rotating machinery industry such as flywheel energy storage are introduced,the present situation of flywheel energy storage and fault diagnosis is expounded,the finite element analysis of flywheel rotor is introduced,the gap between stator and rotor,resonance frequency and harmonic response at fixed frequency are optimized,and the reference of mechanical characteristics is provided for flywheel energy storage fault diagnosis.Secondly,a set of motion monitoring platform including hardware and software is built,and the selection of acceleration sensor,eddy current sensor and other components and the construction of the platform are introduced in detail.The finite element analysis results of flywheel rotor are verified by various detection signals,which also provide data set and state basis for fault diagnosis.Thirdly,the advantages and disadvantages of traditional research methods in the field of fault diagnosis are analyzed,and a research method based on wavelet packet energy spectrum feature extraction and convolution neural network fault diagnosis is proposed.Using the open fault data,the data set is decomposed by wavelet packet to analyze the characteristics and spectrum range of the fault,and the wavelet packet energy spectrum is used to extract the feature signal as the energy vector as the data set to train the convolutional neural network model,and the partial data set is used as the test to verify the feasibility of the fault diagnosis of the method.Finally,according to the fault diagnosis data of flywheel energy storage system,based on the application and optimization of wavelet packet energy spectrum feature extraction and convolution neural network fault diagnosis,the two fault effects are detected,and after the fault analysis results are obtained,the external factors are judged and the main causes of the failure are analyzed.In this paper,on the premise of the feasibility of using open fault data verification method,the radial bearing fault diagnosis of flywheel energy storage is verified,and the fault study step is added.Compared with the traditional fault diagnosis method,the fault misjudgment rate is reduced,which has great progress and practical application value.