Modeling of electrical arc on polluted ice surfaces

Ice accretion on power transmission lines can decrease the electrical insulation along the insulators. Insulator flashover as a result of ice accretion is one of the causes of power outages during ice melt.

For the first time, a research program on the modeling of electrical arc on ice surfaces has been undertaken in the High Voltage and Atmospheric Icing Laboratory and the NSERC/Hydro-Quebec/UQAC Chair on Atmospheric Icing of Power Network Equipment of the University of Quebec in Chicoutimi (UQAC). As a part of this program, the present thesis aims at the modeling and characterisation of the arc on wet-grown ice surfaces and the analysis of the flashover conditions on ice-covered insulating surfaces.

Some fundamental concepts of surface discharge, such as corona discharge, electrical arc, arc development, and mathematical model of arc on polluted surfaces, are first introduced. Then the flashover problem of ice-covered insulators and the related research activities are summarised in the literature review.

A cylindrical glass rod covered with wet-grown ice, the most dangerous type of ice associated to insulator flashover, was used in most of the experiments of the present thesis. Wet-grown ice samples were formed in the cold room of the High Voltage and Atmospheric Icing Laboratory of UQAC. By means of a high speed camera, a data acquisition system and high voltage test facilities, a series of do and ac flashover arcs on ice surfaces were recorded and analyzed. Arc speeds in various periods of arc development, such as the arc-starting period, the arc-developing period, and the finial flashover period, were measured. The relationships between arc current and arc foot radius were obtained using regression on the test results.

The arc voltage-current characteristics under do and ac voltages for cylindrical ice samples were determined. The arc electrode voltage drops were also obtained from the regression on the observations. Arc reignition conditions must be satisfied under ac voltage in order to complete the flashover in the same manner as polluted surfaces. The relative constants for these reignition conditions were obtained using the regression method on the test results, using an approach that was adapted from polluted-surface flashover.

The ice surface conductivity during flashover was also measured and an equivalent surface conductivity, γ_e, for both ac and do voltage applications was defined.

Several factors influencing the flashover on ice surfaces, such as freezing water conductivity, voltage type, arcing distance, and ice surface uniformity were analyzed. The temperatures of ice surfaces and ice body were measured during the melting period. The influence of bulk ice current on the total ice current measured was estimated. It was observed that the leakage current passed principally through the ice surfaces in the cylindrical geometry.

On the basis of the above studies and following the concept of Obenaus for pollution flashover, mathematical models were established for flashovers on ice surfaces under do and ac voltages. Using these mathematical models, the flashover voltages, critical currents, and critical arc lengths were calculated. The mathematical model was then applied to both short and long ice-covered facility insulator strings for different icing conditions. The calculated results were in good agreement with the experimental results on short insulator strings.