Partial discharge identification is of great significance in monitoring power equipment. Although substantial work has been done in partial discharge pattern recognition, most studies have been performed in noise‐free environments. To improve the accuracy of partial discharge pattern recognition, this paper introduces a partial discharge image feature extraction method based on upright speeded‐up robust features. Each group of phase‐resolved pulse sequence data is converted into grayscale images to construct a sample library of partial discharge defects. Upright speeded‐up robust features is used to extract features in grayscale images. The improved support vector machine is used to achieve the best classification of partial discharge sources. The sensitivity of local features to different phase resolutions and different image resolutions of partial discharge phase‐resolved pulse sequence patterns was investigated, and the optimal phase resolution and image resolution were used to construct phase‐resolved pulse sequence patterns. In addition, the recognition effect of this feature extraction method under different noise conditions was also analysed. The results prove that upright speeded‐up robust features can effectively deal with noisy data. The results of this research can provide references for the storage setting and recognition of three‐dimensional images of on‐site partial discharge data.

Sensitive signal perception and accurate fault diagnosis are key to partial discharge (PD) optical detection. To improve the effectiveness of the PD optical detection and diagnosis, this paper uses optical sensing technology based on a light guide rod (LGR) to measure the typical PDs in SF ₆ . The characteristics of optical PD signals are analyzed and a fault diagnosis method for the PD signals detected by the LGR is proposed. An optical and electrical PD experimental platform is built, four different discharge defects are designed, the time-domain signal characteristics of PD optical signals are analyzed, and phase-resolved pulse sequence (PRPS) patterns are plotted. On this basis, a PD optical signal fault diagnosis model based on the pyramid histogram of oriented gradient (PHOG) and optimized support vector machine (SVM) algorithm is established. The results show that the accuracy of this model reaches 96.7%, which is about 10% higher than other algorithms, verifying the effectiveness of detecting PD with an LGR and the reliability of the optical signal diagnostic method based on PHOG features. The results of this study can provide references for PD optical detection and diagnosis in gas insulated switchgear (GIS) with an LGR.

In recent years, the environmentally friendly gas, C₄F₇N, has been recognized as a potential SF₆ alternative gas due to its excellent insulation. To study the streamer discharge characteristics of different electrode shapes in C₄F₇N, a fluid model was established by COMSOL, and simulations were performed. Streamer discharge as a function of varying electrode shapes was analyzed, and the variation law of electric field intensity and photon flux was obtained, which serves as a theoretical basis for optical detection and reliable diagnosis of partial discharge. Simulation results demonstrate that under the same discharge conditions, the electric field inhomogeneity of the needle–plane model, rod–plane model, and ball–plane model decreased successively and the breakdown difficulty increased. With the increase in the applied voltage, the peak electric field intensity of the streamer head was found to increase, whereas an increase in the electrode spacing was found to decrease this property. The change in photon flux was found to be similar to the change in the electric field intensity.

The optical method and UHF method are important techniques for detecting partial discharge (PD) in gas-insulated switchgear (GIS). However, optical signals and ultra-high frequency (UHF) signals may suffer from different degrees of signal loss or interference for different PD types, which leads to incomplete feature information in the optical or UHF patterns and reduces the accuracy of pattern recognition. In this paper, a PD detection device in GIS based on the UHF and light guide rod (LGR) technologies is designed. An experimental platform for electrical and optical PD detection in GIS is set up, and measurements of typical PD are carried out. Optical and UHF time-domain signals of corona discharge, floating discharge, and particle discharge are obtained. Then, an image fusion algorithm based on guided filtering fusion (GFF) is proposed to fuse the optical and UHF phase-resolved pulse sequence (PRPS) patterns Subsequently, a feature extraction method based on speeded-up robust features (SURF) for PD images is proposed. Finally, the recognition effects of multiple classifiers are compared. The results show that the image fusion and feature extraction method for UHF and optical PD proposed in this paper can improve the accuracy of fault diagnosis up to 97.1%.

More electric aircraft (MEA) is an important direction for future aircraft development. The printed circuit boards (PCB) of power electronics equipment in MEA operates in a complex environment of low air pressure, square wave voltage, compact layout and strong electromagnetic interference, which makes the PCB more prone to partial discharge (PD). However, the effective detection method and related discharge characteristics for PCB PD under square wave voltage are still unclear. Therefore, this paper proposes a PCB PD detection method based on fluorescent optical fiber, which has good anti-electromagnetic interference ability. The PD characteristics of the PCB with three typical structures under three air pressures are studied, including PD non-breakdown characteristics, changes in optical pulse amplitude and pulse repetition rate during PD aging, PD breakdown path analysis and PCB surface electric field simulation. The analysis shows that air pressure, space charge and PCB surface flatness all have important influences on the PD of PCB under square wave voltage. It provides important theoretical support for the insulation protection design and fault diagnosis of PCB for MEA in the future.