Multi-Scale Temporally Enhanced Bearing Fault Diagnosis under Variable-Load Conditions

Abstract

This paper proposes a multi-scale temporally enhanced approach for bearing fault diagnosis under variable-load conditions. The method integrates multi-scale feature extraction with temporal enhancement mechanisms to capture both fine-grained and long-range dependencies in vibration signals, enabling robust fault classification across diverse operating conditions.