Research On Fault Diagnosis And Remaining Useful Life Prediction Method Of Rolling Bearing Based On Accuracy Diagram

Rolling bearings are one of the most widely used components in mechanical equipment.Its overall state directly affects the safe operation of mechanical equipment.In recent years,rolling bearing fault diagnosis has become a key research field of rotating machinery,but general fault diagnosis can only be evaluated from the presence or absence of faults,which is obviously not enough.When a fault occurs,the degree of fault needs to be evaluated,and the remaining service life is the most direct method of assessing the condition of the bearing.Therefore,this paper combines fault diagnosis and life prediction methods to comprehensively analyze bearing faults.On the basis of the previous fault diagnosis methods,an accuracy graph algorithm based on frequency band selection is proposed,and for the problems of noise sensitivity and interference frequency in the algorithm,an improved method is proposed to improve the fault diagnosis ability of rolling bearings.In the life prediction,in order to extract more effective features,this paper combines the improved precision map algorithm to extract the high-dimensional features of the bearing.At the same time,in order to eliminate redundant features,this paper proposes a method based on envelope polygon and computational geometry.Finally,the improved characteristics and the constructed health index are used to train the prediction model to obtain the prediction result of the remaining life of the bearing.The main contents of this article are as follows:1.If the frequency bands that are intrinsically related to mechanical faults can be found,the fault diagnosis capability can be significantly improved.Therefore,a precision graph algorithm based on frequency band selection is proposed in this paper.This algorithm uses unbiased autocorrelation combined with energy entropy and energy spectrum entropy to extract signal shock and circulatory stationary characteristics,and uses classification method to distinguish health and fault signals and select fault sensitive frequency bands,which lays a foundation for fault diagnosis of bearing.2.Aiming at the noise sensitivity problem of the accuracy graph algorithm.The band selection reliability of the accuracy graph algorithm is reduced,and the spectrum still contains many extraneous frequencies.Based on the in-depth study of the accuracy graph algorithm,the minimum entropy deconvolution algorithm is used to highlight the fault impact component of the signal and reduce noise interference.A multi-point optimally adjusted minimum entropy deconvolution algorithm is used to accurately extract the fault characteristics of sensitive frequency bands,thereby improving the ability of rolling bearing fault diagnosis.3.Bearing remaining life prediction needs to extract many features,but the bearing vibration signal often contains a lot of noise,and directly extracting features will cause a large error.Therefore,this paper uses the minimum entropy deconvolution and precision graph algorithm to preprocess the signal to highlight the degradation trend information hidden in the features.At the same time,too many features will contain information that has nothing to do with the test status.In order to extract more effective features,this paper proposes a feature selecting method based on envelope polygons and computational geometry algorithms.Calculate geometric algorithm construction parameters to quantify the similarity between features to select features and provide high-quality features for the prediction of the remaining life of the bearing.4.At present,most life prediction methods still use a single model,and the training results are not very satisfactory.In this paper,an adaptive integration algorithm is used to optimize the algorithm of the extreme learning machine.This algorithm further improves the prediction performance of the extreme learning machine and avoids model training accidental error.Compared with the general method of considering only a single health indicator,this paper constructs local and global health indicators,which can be compatible with the local optimal problem while considering the global,further improve the accuracy of life prediction and avoid large errors.