Intelligent framework for degradation monitoring, defect identification and estimation of remaining useful life (RUL) of bearing

Abstract

The proposed intelligent framework seamlessly integrates degradation monitoring, defect identification, and remaining useful life (RUL) estimation for a comprehensive and holistic solution to bearing health assessment. It leverages a novel directed divergence (DD) measure to compare probability distributions, generating a Health Indicator (HI) that captures the health status of the bearings. This HI serves as the cornerstone for the framework's cohesive approach. An alert system, utilizing the modified Pauta criterion and iterative window scanning, enhances anomaly detection, while a double-check mechanism minimizes false alarms. For defect identification, dynamic analysis-assisted wavelet filtering harnesses a customized wavelet filter generated from simulated interactions between defective components and rolling elements. In parallel, the framework employs a graph convolution network (GCN) with a 1D health indicator, derived from the proposed directed divergence measure between the reference data and running conditions, to estimate RUL. By unifying these components, proposed framework offers an efficient, integrated, and proactive solution for monitoring, defect identification, and RUL estimation in bearing health assessment, ultimately enhancing system reliability and maintenance strategies.