Weak Fault Feature Extraction of Planetary Bearings Based on CPO-VMD Optimized by Maximum Correlated Kurtosis

To address the difficulty in weak features extraction from planetary bearing faults in strong noise environments, and improve the adaptability of parameters definition for variational mode decomposition (VMD), this paper proposes a fault diagnosis method for planetary bearings based on adaptive variational mode decomposition. Firstly, maximum correlation Kurtosis (MCK) is adopted as the fitness for evaluating the intrinsic mode function (IMF) when VMD is applied to bearing vibration signals. The crested porcupine optimizer (CPO) algorithm is employed to adaptively select the mode number K and penalty factor α. The relevant IMFs are then reconstructed based on MCK and demodulated via envelope analysis to extract fault features. Verification using both simulated signals and experimental data from planetary bearings with outer and inner race faults demonstrates that the proposed method effectively suppresses strong noise interference, accurately identifies the optimal modal component, and clearly extracts the fault characteristic frequency. Compared with methods using envelope entropy optimization and ensemble empirical mode decomposition (EEMD), the proposed approach shows clear advantages in extracting planetary bearing fault features.