Rolling Bearing Fault Diagnosis Method Based On Support Vector Machine (svm)

Bearings are the core and wearing parts of the rotating machinery. Rotating machinery for general industrial make the most use of rolling bearings which plays a very important role in rotating machinery. It has been a subject of great concern for the academia and industry to monitor it and carry out fault diagnosis. Rolling bearing failure may lead to equipment noise and abnormal vibration.Serious failure may lead that the entire device as well as the entire production line can not operate normally. Therefore, the research about rolling bearing fault diagnosis has important theoretical significance and application value.Combined with the research contents from Natural Science Foundation of Jiangxi Province, this paper focuses on the fault diagnosis methods of rolling bearing which based on Wavelet Packet Analysis and Support Vector Machines. Test rolling bearing simulation fault with the designed rolling bearing bench and data acquisition systems. On this basis, it introduces wavelet analysis theory and use wavelet packet analysis to complete the noise reduction and fault feature extraction of original vibration signal data collected by the data acquisition system. Finally, associated with the obvious advantage of less sample classification from Support Vector Machine theory, this thesis proposes diagnosis methods based on combination of Wavelet Packet analysis and Support Vector Machines.The major work by this paper as the following:1. First study the vibration mechanism, failure modes and detection technology of the rolling bearing, analyze common detection methods strengths and weaknesses, use measurement data of rolling bearing vibration signal to diagnose bearing fault,and then study the rolling bearing vibration mechanism and characteristic frequency.2. Analyze the basic idea of statistical learning theory and Support Vector Machine, introduce Support Vector Machine into the roller bearing fault diagnosis, and give the basic steps and methods of Support Vector Machines applied for rolling bearing fault diagnosis.3. For the object of QPZ-I fault simulation test platform, design test bench vibration data acquisition system, conduct fault simulation of the rolling bearing inner ring, outer ring and rolling element,and provide diagnostic data for rolling bearing fault feature extraction.4. For the diagnosis object of Simulative rolling bearing fault data, The test the methods of signal de-noise based on wavelet analysis and use wavelet packet decomposition to extract the fault feature, combined with the advantages of Matlab platform, study the extraction methods of rolling bearing signal feature based on wavelet packet decomposition and instantiated.5. Lack of targeted that standard support vector machine can not be directly used to solve the fault diagnosis as this typical multi-class classification problem. elaborate Support vector machine classification method of multi-valued, employ binary tree multi-class classification algorithm to construct a classification model combined with vibration fault feature vectors extracted by the wavelet packet decomposition, diagnose of rolling bearings fault, and explore the parameter optimization of the nuclear parameters related of SVM classification accuracy.The simulation results demonstrate the correctness and effectiveness of the support vector used for rolling bearing fault diagnosis.Above all. it is feasible that rolling bearing fault diagnosis method based on Support Vector Machine by this thesis study.Analysis and diagnostic results consist with the actual.It can meet the requirements of the rolling bearing fault diagnosis and has a guiding role in the fault diagnosis of rolling bearings...