Hybrid Method with Parallel-Factor Theory, a Support Vector Machine, and Particle Filter Optimization for Intelligent Machinery Failure Identification

Here, a novel hybrid method of intelligent fault identification within complex mechanical systems was proposed using parallel-factor (PARAFAC) theory and adaptive particle swarm optimization (APSO) for support vector machine (SVM). The multi-scale analysis studied to reconstruct tensor feature information based on three-dimensional matrix time, frequency, spatial vectors. A wavelet used transform the original multi-channel experimental data acquired from gearbox into multi-level structure. optimal correspondence among two-dimensional signals in frequency time domains different modes established by PARAFAC theory. An APSO algorithm developed obtain parameter structures an SVM classifier. comparison with existing time–frequency showed that PARAFAC-PSO-SVM diagnosis model effectively eliminated redundant multi-dimensional features but retained important components. PARAFAC-APSO-SVM diagnostic achieved fast, accurate, simple fault-classification results, could provide theoretical application diagnosis.