Adaptive Chirp Mode Decomposition Holo Spectral Analysis for Rolling Bearings Fault Diagnosis

Fault feature extraction from bearings is crucial for fault diagnosis. Time–frequency analysis is effective for extracting time and frequency features from vibration signals but struggles to capture intersignal modulation. In this study, a novel fault feature extraction method, adaptive chirp mode decomposition-Holo spectrum analysis (ACMD-HSA), is proposed. The method uses a two-layer signal decomposition process to capture the coupling between amplitude modulation (AM) and frequency modulation (FM) in nonlinear and nonstationary signals. In the first layer, the signal is decomposed by ACMD to extract the carrier frequency (CF), and the upper envelope is constructed using cubic spline interpolation. In the second layer, the upper envelope is further decomposed to obtain modulation frequency (MF) and energy. The ACMD-HSA spectrum integrates these components, enhancing fault characteristic separation. Simulation and experimental results demonstrate its superior ability to recognize modulation relationships and identify bearing faults.