Distributed detection and acoustic emission (AE) waveform retrieval of partial discharge (PD) in power cable joints are significant but challenging tasks. In this paper, we proposed a cable joint PD detection method based on optical fiber distributed acoustic sensing (DAS). Ordinary single-mode optical fiber (SMF) was wound on the outer semiconductive layer to obtain sufficient sensitivity. Coherent optical time domain reflection (COTDR) principle was utilized to detect the phase variation and locate the PDs. The AE waveforms were retrieved by Fourier integration and difference operation. The experimental results indicated that the sensing system had the ability of multi-point monitoring. The PD AE waveforms caused by conductor spike and insulation air gap defects were accurately retrieved.
21 February 2023 in IEEE Transactions on Instrumentation and Measurement
Optical fiber distributed acoustic sensing (DAS) is a potential approach to realize on-line monitoring and rapid localization of power cable partial discharges. Distributed detection of discharges requires a DAS system with a wide frequency band and low noise level. For this purpose, an improved coherent optical time-domain reflectometry (COTDR) method based on dual-frequency probe pulses was developed. The method doubled the length-bandwidth product (LBP) of traditional DAS systems without significantly increasing the complexity of signal demodulation. A discharge signal extraction algorithm was also proposed based on the detection results of the dual-frequency pulses to exclude the effects of fading noises and vibrations. The improved COTDR system demonstrated the frequency band extension by detecting and recovering the 50 kHz sinusoidal vibration. The sensing distance was 1.1 km and the spatial resolution was 7.15 m. Afterwards, the multi-point weak spark discharge acoustic signals were correctly measured and localized, and the frequency components up to 71.36 kHz were detected. The proposed COTDR method provides a prospective solution of distributed partial discharge detection for power cables.
1 September 2023 in IEEE Transactions on Instrumentation and Measurement
To realize the highly sensitive multiple-point detection of partial discharge (PD) acoustic emission (AE) in power transformers, an ultrasonic sensing system based on a distributed feedback fiber laser (DFB-FL) array was proposed. Four DFB-FLs were connected on a single optical fiber as sensing elements to synchronously detect PD AE signals. The heterodyne coherent detection structure was formed by introducing an acousto-optic modulator (AOM) into the unbalanced Mach-Zehnder interferometer (MZI) to realize the rapid and highly sensitive demodulation of ultrasonic signals. The influence of noises and arm length difference on the lower limit of detection (LLOD) were theoretically discussed and experimentally verified. By adopting a 70-meter-long delay fiber, the proposed system performed a 9.54 dB lower average detection limit than that of PZT. Finally, based on the proposed system, a partial discharge source was accurately located in a 1000 mm×1000 mm×1100 mm oil tank, and the localization error was 24.6 mm. The proposed method exhibited promising application prospects in sensitivity, installation convenience, anti-interference and reliability.
26 June 2024 in IEEE Transactions on Instrumentation and Measurement
Distributed detection of partial discharge (PD) in joints of long power cables is both a necessity and a challenge. In order to improve the distance and sensitivity of PD detection, an improved coherent optical time domain reflectometer (COTDR) system is proposed in this paper. A two-stage optical amplification scheme was established, and the pump currents of erbium-doped fiber amplifiers (EDFAs) were optimized to improve the signal-to-noise ratio (SNR) of the beat signal. The dual-frequency probe pulses were introduced to suppress fading noises in the distance-domain and prevent false positives in PD detection. The test results based on real cable joints indicated that the system enabled distributed detection and accurate waveform retrieval of PD events. Based on the above improvements, the proposed system enhanced the PD sensing performance from the existing 1 km sensing distance and nC-level detection limit to 20 km and 127 pC, respectively. This performance indicated that the COTDR system possessed practicality in the distributed detection of cable joints PDs.
15 November 2024 in IEEE Transactions on Dielectrics and Electrical Insulation
Metal particle pollution was one of the significant factors that endangerd the reliability of the gas insulated switchgears (GIS). The detection of metal particles in GIS was crucial for ensuring the safe operation of GIS. This paper established a calculation model for the cut-off frequency of acoustic emission (AE) generated by the metal particle collision with the GIS enclosure. The effects of metal particle materials, the Gaussian equivalent radius, and the bounce heights on the cut-off frequency of the acoustic emission signal were obtained. Through the acoustic emission measurement experiments of metal particle colliding with an aluminum plate, the accuracy of the cut-off frequency calculation model was verified. Besides, the experiment confirmed that the propagation mode of the acoustic emission signal on GIS enclosure was primarily the S0 mode Lamb wave. The attenuation of the AE signal was analyzed. The geometric attenuation and the energy loss attenuation were determined as the primary attenuation processes. A calculation model for the attenuation of acoustic emission signal was established. Based on this, the variation of the cut-off frequency after propagation on the GIS enclosure was clarified. The acoustic emission and propagation characteristic proposed in this paper provided guidance for optimizing the frequency band of ultrasonic sensors and setting up the measurement points.
Partial discharge (PD) ultrasonic detection is an early sign of the insolation defects of power transformers. The early diagnosis of PD requires the high sensitivity and reliability of ultrasonic sensing systems. For this purpose, a reformative PD ultrasonic sensing system based on phase-shifted FBG (PS-FBG) was demonstrated. By using PS-FBG as the ultrasonic sensing unit, the ultrasonic sensing system improved the response to the ultrasonic signal and overcame the electromagnetic noise. To compensate for the influence of temperature change on the ultrasonic sensing system, an automatic wavelength scanning demodulating method was carried out. The wavelength spanning strategy was optimized based on the principle of cross-correlation, in order to quicken the spanning. A PD detection test in the transformer oil was conducted, and the result shows that PS-FBG was 17.5 times more sensitive than PZT. Because of the better ultrasonic response, the proposed system was able to achieve the early diagnosis of insolation faults in a power transformer.