Rolling Bearing Fault Diagnosis Using Adaptive Variational Mode Extraction Based on Harmonic Coherence Factor Under Strong Noise

Abstract

To address the challenge where early fault signals of rolling bearings are easily submerged by strong noise, and fault features are difficult to extract under harsh working conditions, this paper proposes an adaptive fault diagnosis method termed HCF-IAPO-VME. Firstly, a novel harmonic coherence factor (HCF) is constructed to simultaneously evaluate the intensity of periodic impulse features and the consistency of harmonic structures without relying on prior fault frequency information. Secondly, the Improved Arctic Puffin Optimization (IAPO) algorithm is adopted, and HCF is taken as the fitness function to adaptively determine the optimal parameters of variational mode extraction (VME). Different from conventional VME and optimization-based methods that depend on prior fault frequencies and single-feature metrics, the proposed method achieves adaptive parameter optimization and reliable weak fault extraction under heavy-noise conditions. Subsequently, the optimized VME is used to extract fault-related modes, and envelope demodulation is applied to identify fault characteristic frequencies. Simulation signals and two public experimental datasets are used to verify the performance of the proposed approach. Quantitative and qualitative comparisons with VMD, FK and AVME are carried out using the fault feature coefficient (FFC), kurtosis, signal-to-noise ratio (SNR) and envelope entropy (EE). The results show that HCF-IAPO-VME can effectively suppress strong noise interference and accurately extract weak early fault features in rolling bearings.