The Extrinsic Fiber Fabry–Perot Interferometer (EFPI) fiberoptic ultrasonic sensor can be used to detect partial discharge ultrasonic signals inside gas-insulated switchgear (GIS) and has various uses in pattern recognition research. Compared with traditional piezoelectric sensors, it offers both high sensitivity and strong resistance to interference. Based on this information, we construct four typical PD models (representing the tip, metal particle, suspension, and surface) in a GIS cavity filled with 0.4 MPa SF6 gas, 0.6 MPa SF6N2 gas, and 0.5 MPa C4F7NCO2 gas. We then use the EFPI sensor to detect PD ultrasonic signals, extract their waveform characteristics to form a database of characteristic parameters, and apply the Transformer algorithm. The detected signal offers outstanding pattern recognition when applied to GIS discharge samples in the laboratory, and the Transformer algorithm achieves a 100% recognition success rate, which is much higher than that of Support Vector Machine (SVM) machine learning algorithms.
Gao et al. (Tue,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: