Weibull distribution parameters for fault feature extraction of rolling bearing

A novel approach to fault feature extraction using Weibull distribution parameters is proposed. After the original signal of bearing vibration is modeled as the Weibull distribution, its scale parameter is extracted as a new feature vector for the bearing running state. The tests results of fault diagnosis of the rolling bearing verify that this new feature can catch the regularity of changes in the information of bearing vibration more sensitively and accurately, and have higher separability suitable for pattern recognition by support vector machine classifier.