Band Relevance Factor (BRF): A novel automatic frequency band selection method based on vibration analysis for rotating machinery

Abstract

Monitoring rotating machinery has become a fundamental activity in the industry given the high criticality in production processes. Extracting useful information from signals is a key factor for effective monitoring. Several studies in the areas of Informative Frequency Selection bands (IFB) and Feature Extraction/Selection have demonstrated the need to identify the bands of interest in vibration signals. However, typical methods in these areas focus on identifying bands where impulsive excitations are present or analyzing the relevance of features after signal extraction. Therefore, they do not focus on other regions that may be related to changes in the dynamic behavior of machines or faults. Furthermore, the methods generally require parameter adjustments and are not automatic. To overcome these problems, the present study proposes a new approach called Band Relevance Factor (BRF). BRF aims to perform an automatic selection of all relevant frequency bands for a vibration analysis of a rotating machine based on spectral entropy. In other words, automatically identify all frequency bands that are related to changes in the behavior of the machines or faults. The results are presented through a relevance ranking and can be visually analyzed through a heatmap. The effectiveness of the approach is validated on a synthetically dataset and on two real datasets, showing that BRF is capable of automatically identifying bands that present relevant information for the analysis of rotating machinery.