Abstract：The contact electrical contact fault of gas insulated switchgear（GIS）will affect its electrical

performance. A multi-physics coupling model of GIS contact vibration was first established to address the difficulty

in diagnosing contact faults. An experimental platform for high-current detection of GIS was built using AC/DC test

equipment，simulating different operating conditions and collecting vibration signals to demonstrate the feasibility

of vibration detection and experimentation. The wavelet packet energy entropy diagnostic method was proposed

and compared with time-domain and frequency-domain analysis methods，demonstrating its good anti-interference properties. Finally，online experiments were conducted at the substation site，analyzing signals indicating that under the same current，the equipment exhibited signal frequency domain amplitudes of 100 Hz and 800 Hz respectively for normal and abnormal states，with significant differences in time-domain signal acceleration as well. Analyzing wavelet packet energy entropy，it was found that during normal operation，the energy entropy in the first and eighth frequency bands was 0.601 and 0.023 respectively，while during abnormal operation，the entropy in the first and eighth frequency bands was 0.041 and 0.142 respectively，with a significant decrease in entropy in the first frequency band and a significant increase in the eighth frequency band. The conclusion indicates that this method can effectively diagnose the contact status of GIS contacts.

      Key words：electrical contact；gas insulated switchgear（GIS）；multi-physical field coupling；vibration signal；wavelet packet；diagnose faults