Gas-insulated switchgear (GIS) is a critical piece of equipment

and the assessment of its insulation condition relies heavily on the early detection of partial discharge (PD). Existing detection methods

such as the ultra-high frequency (UHF) technique

still face limitations in practical applications

including insufficient sensitivity

weak anti-interference capability

and difficulties in feature extraction. To overcome the bottlenecks of conventional detection technologies and address the frequency band limitation of traditional UHF detection (0.3~1.5 GHz)

this study employs a cesium atomic vapor cell with a center frequency of 3.27 GHz for the first time. Targeting the frequency range beyond the reach of UHF methods

a full-scale GIS partial discharge quantum detection platform was constructed

and the Rydberg atom quantum sensing technology is applied to investigate protrusion and floating defects in a 126 kV full-scale GIS. Experimental results demonstrate that the quantum detection system achieves a signal detection limit of 10 mV

with the measured PD amplitude being 0.6 times that of the UHF signal

exhibiting a linear correlation between the two. Statistical analysis of discharge amplitude versus discharge repetition rate reveals that the repetition rate of protrusion defects is significantly higher than that of floating defects. Furthermore

maps of discharge amplitude-pulse count and discharge amplitude-pulse interval time were statistically compiled

revealing the intrinsic differences between the two defect types. These findings validate the feasibility of the Rydberg atom technology for GIS partial discharge detection and lay a foundation for insulation defect identification using quantum sensing.