Application Of The Improved EMD Method In The Fault Diagnosis Of Rolling Bearings

Mechanical fault diagnosis is the key to manufacturing equipment for effective basic maintenance,improved transformation of the mechanical equipment for the protection of the safe and stable operation of great significance and prolong equipment life,improve the efficiency of inspection and preparedness.The important entry point for fault diagnosis of mechanical equipment and the key point is that the vibration signal monitoring machinery and equipment,by monitoring and analysis of vibration signals of mechanical failure of the device can effectively diagnose the current operating status of the device.Therefore,in today's and future transformation and upgrading of modern manufacturing industry,effective mechanical fault diagnosis is still the crucial factor in its realization.To diagnose equipment faults through the vibration signal of machinery and equipment,it is inevitable to obtain a clearer signal tomorrow.The premise is more necessary to overcome all aspects of the mechanical equipment operating environment interference,such as a variety of vibration signals staggered,from time to time there will be more obvious aspects of his noise.In this way,the extraction of the characteristic signal becomes the key to the implementation of the equipment fault diagnosis,otherwise the signal with the obvious characteristic can not be extracted,and accurate equipment fault diagnosis can not be performed.In fault signal analysis,with the increasingly comprehensive analysis technology,such as correlation analysis,time-frequency analysis,time-domain analysis,analysis of frequency-domain analysis has been well-known and widely used.Status quo mechanical fault vibration signal extraction analysis above and over the years many scholars and researchers of the signal characteristics of mechanical failure,we found EMD mechanical failure in the response to these signals,although there will be a noise problem situation,endpoint issues,but to a certain extent On the mechanical failure signal to adapt to low signal to noise ratio,the frequency component is complex,not smooth enough for these characteristics of failure analysis and application.Therefore,in order to improve the above EMD stability,this method ismore comprehensive,mature and accurate for fault diagnosis.This paper makes a statement and explanation for the improvement.The main research contents include the following aspects:(1)In order to solve the noise problem caused by the decomposition of EMD greater impact of this situation,we must first deal with the noise,noise noise reduction is the most viable,so we mentioned methods MRSVD,the first use of multi-resolution singular value decomposition(Multi-resolution singular value value decomposition,MRSVD)to noise reduction and then EMD decomposition,the feasibility of the simulation signal and the rolling bearing fault feature extraction experiments made a valid proof.(2)In order to solve the problem of finite-length signal analysis and processing in the signal analysis,LS-SVM model is used in this paper to improve it.The research shows that the predictive parameters of LS-SVM model are sensitive to the state changes.Therefore,it is very effective to analyze the state changes of the system by using the LS-SVM model as the feature parameter.