Abstract:

To address the difficulties in fault feature extraction and insufficient recognition accuracy of electro-hydraulic actuator systems under complex operating conditions, a fault diagnosis method based on multi-scale time-frequency interactive fusion network is proposed. The method comprehensively captures fault information by constructing a multi-scale feature extraction module that integrates time domain, frequency domain, time-frequency domain and statistical features, effectively characterizes the latent coupling relationships among sensors using signal feature library, temporal convolutional structure and graph modeling mechanism, and introduces spatial and channel dual attention mechanism along with fault-specific enhancement paths to strengthen the expression of weak fault features, thereby achieving deep modeling and classification discrimination of multi-source signals. Experimental validation on the constructed electro-hydraulic actuator simulation platform shows that the proposed method achieves a fault recognition accuracy of 99.03% on the test set, with ablation experiments further verifying the synergistic gain effect among various functional modules. The results demonstrate that the proposed method exhibits excellent fault diagnosis accuracy and robust system adaptability, effectively meeting the intelligent monitoring requirements of typical electro-hydraulic actuation systems.

Key words: electro-hydrostatic actuator（EHA）, fault diagnosis, multi-scale feature extraction, attention mechanism, deep learning