Fault Diagnosis of Rolling Bearing Based on EEMD Information Entropy and Improved SVM

In order to identify the fault mode of rolling bearings rapidly, this paper presents a method of extracting bearing fault feature based on information entropy, and takes the extracted fault feature vector as the input of optimized SVM to identify bearing faults. Firstly, the original signal is decomposed by EEMD, and the information entropy of each order IMF is extracted. Then, the main information is extracted by principal component analysis (PCA), which is formed the input feature vector of SVM. Finally, the parameters of SVM are optimized by Grid Search method to classify the fault information. The experimental results show that the method can identify different bearing fault information quickly and effectively.