Estimation of discharge magnitude of composite insulator surface corona discharge based on ultraviolet imaging method

Ultraviolet (UV) imaging is a non-contact discharge detection method, in this study, a method, which considers the two factors of observation distance and imager´s gain, of estimating the apparent discharge magnitude on composite insulator was proposed. According to the discharge UV image feature, the discharge emitted radiation regions in UV images were extracted with digital image processing algorithm, and then the emitted radiation area parameter was defined. In laboratory, the relationship of emitted radiation area to apparent discharge magnitude, imager´s gain and observation distance were studied respectively. Research shows that emitted radiation area increases with the enhancement of discharge intensity, but it has non-linear characteristic; in the gain range of 50 to 80%, emitted radiation area to gain has approximate exponential function relation, when the gain is 50, 60, 70 and 80% respectively, the emitted radiation area to the observation distance has approximate power function relation. Based on the above research, an adaptive neuro-fuzzy inference system (ANFIS) model was established and realized the estimation of apparent discharge magnitude, test shows that it has high prediction accuracy.