Time domain bearing diagnosis using kurtosis based filtering and fuzzy Logic

摘要

This paper introduces a bearing defects diagnosis technique based on time domain signal processing, first step being a kurtosis based filtering. The filtering detects the shock pulses generated by bearing defects on the vibration signal. The diagnosis is done using a fuzzy classifier that provides defect alerts for all kinds of bearing structural defects. The detected defects periods are the inputs for the classifier.