Research On Fault Diagnosis Of Bearing Based On Adaptive Stochastic Resonance

With the rapid development of large-scale industrial mechanical equipment,it is of great research value to study mechanical fault diagnosis,and the feature extraction of fault signal is the basis of mechanical fault diagnosis.In early fault diagnosis,compared with background noise,the fault signal is often too weak to be extracted effectively.The stochastic resonance algorithm can transfer some energy from noise to the useful weak signal and amplify the weak signal,so as to realize the fault diagnosis.Therefore,The stochastic resonance algorithm is used to solve the problem of the difficulty of extracting weak feature signal in the bearing fault.Firstly,combined with the basic structure and common faults of rolling bearing,the vibration characteristics of rolling bearing were studied,and the fault frequency of each component was obtained,which provided the basis for the diagnosis of bearing fault signal.Secondlly,the theory of stochastic resonance was discussed,including the mathematical model,basic theory and measurement index of stochastic resonance,and the influence of different system parameters on the effect of stochastic resonance was analyzed.Aiming at the problem that the classical stochastic resonance is only suitable for the detection of small parameter signals,while the bearing fault signals is a periodic large parameter signal with large noise,which does not conform to the small parameter condition,a variable scale stochastic resonance method was adopted.The validity of the variable scale stochastic resonance was verified by processing the simulation signals of large parameters.Then aiming at the problem that the parameters of stochastic resonance system determine the effect of resonance,the firefly algorithm with fast running speed and less algorithm parameters was adopted to optimize the system parameters of stochastic resonance in parallel.Namely an adaptive stochastic resonance method based on the firefly algorithm was proposed.The simulation and experimental results showed that the method can enhance the weak feature signal,improve the signal tonoise ratio,and realize the feature extraction of weak signal effectively..When the optimization effect of this method was compared with that of adaptive stochastic resonance based on traditional intelligent optimization algorithm(genetic algorithm and ant colony algorithm),it was concluded that the adaptive stochastic resonance method based on firefly algorithm has shorter running time and better optimization effect,which is more suitable for the rapid processing of big data's massive signals in engineering.Finally,the fault diagnosis system of rolling bearing based on Labview was set up.The system has the functions of data acquisition,fault frequency calculation,data processing and fault diagnosis.The data processing uses the stochastic resonance method to carry out the spectrum analysis and then extracts the frequency of the fault feature.The experimental verification of the measured fault signal of the bearing showed that the system can effectively complete the data acquisition,processing and diagnosis of the bearing fault signal,which has a broad application prospect.