VMD combining wavelet packet fuzzy entropy and BKA-SVM for motor bearing fault diagnosis

Objectives Aiming at the problem of low reliability of fault diagnosis due to the inconspicuous fault signals and difficult to capture fault features in rolling bearings of ship motors, the article proposes a feature extraction scheme that integrates Variational Modal Decomposition (VMD) and Wavelet Packet Fuzzy Entropy, and introduces the Blackwinged Kite Algorithm (BKA) optimized Support Vector Machine (SVM), which is the most suitable method to extract features. WPFE) feature extraction scheme, and introduces a Support Vector Machine (SVM) model optimized by Black-winged Kite Algorithm (BKA) to achieve fault diagnosis. Methods The collected motor bearing vibration signals are decomposed by VMD, and the optimal Intrinsic Mode Function (IMF) components are selected according to the principle of minimum envelope entropy. Then, the optimal IMF components selected are further refined using wavelet packet technique, and the fuzzy entropy value is calculated. Finally, the BKA-SVM model is constructed and the feature data are imported for fault diagnosis and classification. Results Through different algorithms optimization SVM simulation comparison experiments and self-constructed experimental platform validation, the diagnosis accuracy of BKA-SVM model for three different sets of samples is as high as 98.33%~100%. Conclusions Compared with the Particle Swarm Optimization (PSO) algorithm, Sparrow Search Algorithm (SSA) and Newton-Raphson-based optimizer (NRBO) optimized SVM model, BKA-SVM has higher classification effect and accuracy in the extraction and diagnosis of rolling bearing faults, which can provide a reference basis for the fault diagnosis of ship motor bearings.