Angular dependence in ultraviolet imaging of HVDC partial discharges: Mechanism investigation and correction approach

To address the observation angle effects on the diagnosis accuracy of ultraviolet (UV) imaging for partial discharge (PD) monitoring in high-voltage direct current (HVDC) equipment, this study investigates the underlying mechanism and develops a corresponding correction method. A dual-channel, angularly adjustable synchronous UV imaging platform was established to acquire synchronized datasets under various viewing angles, including both photon counting and UV spot data. A probabilistic statistical analysis of the UV spots was performed to characterize their spatiotemporal distribution, resulting in the UV spot distribution maps. The results indicate a sinusoidal relationship between the viewing angle difference and the photon counting ratio from the two channels, with a peak occurring near 45°. This angular dependence is attributed to changes in the cross section and the spatial overlap of the discharge regions in the UV spot distribution maps. A regression model based on the sine function was established to correct the photon counting results for variations in observation angle. The mechanism behind the angle-dependent photon counting variation in UV detection is elucidated, thereby providing a foundation for improving the accuracy and reliability of UV-based PD diagnostics in HVDC equipment.