Optical partial discharge diagnostic in SF gas insulated system via multi-spectral detection.

Light emitted from partial discharges (PDs) in SF gas is investigated in the view point of insulation status diagnostics. Light intensity and integral spectrum of PD are proved to have significant correlations with PD activities and the involved dielectric interfaces. Based on synchronous light pulse detections in ultraviolet (UV), visible (VIS) and near-infrared (NIR) regions, the multispectral stochastic PD diagnosis is realized with more information beyond the conventional PD detections. The optical PD detection is also compared with ultra-high frequency (UHF) PD detection and acoustic emission (AE) PD detection in term of linearity, sensitivity and statistical property. Beyond that, a clustering analysis is performed on the multispectral photon counts of different conditions by quantifying them in a ternary diagram where some distinguishable regions emerge for PD recognition.