The Fault Diagnosis Of Rolling Bearing Based On Empirical Mode Decomposition And Multi-Scale Entropy

In common mechanical equipment, rolling bearing as an important part of the whole equipment and its performance in the operation of performance is good or bad, directly related to the device of high and low production efficiency, thus fully accurate and timely recognition in the process of rolling bearing in operation state, is particularly important.Rolling bearing fault occurs because equipment operation of the internal components and the process continues with load speed, and other reasons, led to some inner and outer ring rolling bearing and roller failure, and the failure occurs, at the same time of the vibration signal is also associated with sexual showed severe nonlinear and non-stationary characteristics.In this paper, the improved empirical mode decomposition with multi-scale entropy, is proposed based on multiscale entropy of intrinsic mode function of a recognition method of fault diagnosis, is applied to the vibration signal of rolling bearing, combining with the experiment data, make a complete analysis of theory and method of application, standing in a new perspective, to detect fault diagnosis, according to the results by the improved empirical mode decomposition and the combination of multi-scale entropy, the entropy value, intrinsic mode function using LIBSVM 3.1 package, can be the realization of the accurate identification of fault diagnosis.This study reflects the actual value and significance.This paper mainly focuses on the following content:First of all, the research significance and development of fault diagnosis as a whole article guide language, followed by the introduction of the theory and method, based on wavelet function of empirical mode decomposition algorithm and multi-scale entropy algorithm and the fault pattern of rolling bearing and formation reasons explained, the improved empirical mode decomposition algorithm and multi-scale entropy respectively,the fault data to do a feature extraction unilaterally interpreted fault recognition effect, at last, using the LIBSVM 3.1 software package, and multi-scale entropy and the improved empirical mode decomposition method are combined, realize the fault diagnosis and identification.