Enhanced Stability in Partial Discharge Detection: A Differential Matching Approach for Fabry-Perot Interferometric Sensors

This paper proposes a white-light path-matched differential interferometry (WL-PMDI) system combined with phase-generated carrier (PGC) demodulation for ultrasonic detection of partial discharges using extrinsic Fabry-Perot interferometric (EFPI) sensors. By cascading the EFPI sensing unit with a Mach-Zehnder interferometer (MZI) as a reference and leveraging the low-coherence characteristics of broadband light sources, the system transforms traditional intensity demodulation (which is susceptible to operating point drift) into stable phase demodulation. This approach fundamentally addresses the reliability challenges of EFPI sensors in long-term monitoring within complex environments, while incorporating PGC technology to achieve high-sensitivity dynamic signal extraction. Experimental results demonstrate that the system exhibits excellent linearity and maintains output signal fluctuations below 10% under temperature variations from 30°C to 90°C and static pressure interference, showcasing superior stability. In simulated transformer oil partial discharge detection, the system clearly resolves microsecond-scale acoustic pulses with a signal-to-noise ratio exceeding 20 dB, and maintains stable performance (with only 0.5% SNR degradation) during a 180 minute continuous operation test—significantly outperforming conventional intensity demodulation schemes that fail due to operating point drift. This study provides an effective technical solution for reliable, long-term online monitoring of high-voltage equipment insulation status in strong electromagnetic interference environments.