Intelligent fault diagnosis for home appliances using acoustic-vibration feature fusion and sparse representation

This paper proposes an intelligent fault diagnosis method for home appliances by adaptively fusing acoustic and vibration features. Sound and vibration signals are first decomposed via multi-scale wavelet analysis (5 layers for sound, 7 for vibration) to capture transient characteristics. Multidimensional features including energy distribution, statistics, and entropy are then extracted. An adaptive fusion mechanism dynamically weights these features based on their discriminative power and operational conditions. Sparse representation is subsequently applied to the fused features to enhance robustness by filtering noise and redundancy. The refined features are matched against a pre-established fault dictionary to identify the fault type and location with a confidence score. Experimental results demonstrate that the proposed method achieves high recognition accuracy and strong adaptability, providing a reliable solution for intelligent maintenance.