A Feature Extraction Method for Fault Classification of Rolling Bearing based on PCA

This paper discusses the fault feature selection using principal component analysis (PCA) for bearing faults classification. Multiple features selected from the time-frequency domain parameters of vibration signals are analyzed. First, calculate the time domain statistical features, such as root mean square and kurtosis; meanwhile, by Fourier transformation and Hilbert transformation, the frequency statistical features are extracted from the frequency spectrum. Then the PCA is used to reduce the dimension of feature vectors drawn from raw vibration signals, which can improve real time performance and accuracy of the fault diagnosis. Finally, a fuzzy C-means (FCM) model is established to implement the diagnosis of rolling bearing faults. Practical rolling bearing experiment data is used to verify the effectiveness of the proposed method.