The Research On Fault Diagnosis Technology Of Rolling Bearing Based On Android

The research on the fault diagnosis of large and critical equipments in metallurgical enterprises has important theoretical significance and engineering application value.As fault diagnosis instrument used by the current metallurgical enterprises has the disadvantages of inconvenient carrying,relatively unitary test method,poor ability of algorithm analysis,failure to achieve remote data transmission and remote fault diagnosis in time.This paper has carried out the research on the extraction of bearing fault characteristics,the intelligent diagnosis method and the development of Android technology.A data acquisition and analysis instrument is made,and a fault diagnosis system based on Android mobile phone is designed.The data acquisition analyzer uses a dual channel signal preprocessing circuit,which can better extract the micro impulse signal.The control part of the fault diagnosis instrument adopts the DSP(TMS320F28335)combined with Wi Fi(CC3200)to acquire the vibration signals of large equipment,and send them to the Android mobile phone terminal for fault diagnosis on-site,or remote mobile clients for fault diagnosis online.The Android mobile terminal has the functions of waveform display,feature extraction and comprehensive fault diagnosis of vibration data.The main work includes:(1)Research on the method of bearing fault feature extraction.In addition to traditional method of feature extraction,a kind of adaptive resonance demodulation feature extraction method based on genetic simulated annealing algorithm is proposed in this paper.With the help of the excellent global optimization characteristics of the genetic simulated annealing algorithm,better result of extracting the weak fault impact signal from complex interference noise is obtained.(2)Research on intelligent fault diagnosis method.In the fault diagnosis system,a BP neural network fault diagnosis model based on genetic algorithm is established.Using the noise reduction characteristics of resonance demodulation and the global optimization characteristics of genetic algorithm,the training and learning model parameters achieved by our method have better performance in comparison with the traditional BP neural network.Then,the model is applied to the intelligent fault diagnosis system of Android mobile phone,making the diagnosis more intelligent and accurate.Various functions of the whole system are verified on the platform of fault diagnosis.It shows that the diagnostic instrument is easy to carry,convenient to use in industrial field,and accurate to diagnose various faults of bearings.