Abstract
The established model of the flashover of an insulator, whose surface is covered with a moist, conducting layer of pollution, envisages that it develops from the formation and bridging of a dry band by a partial-arc pre-discharge. This pre-discharge is assumed to possess arc characteristics, so that the electric field across the dry band will fall as the partial arc current increases. Observations of insulators which are only lightly polluted, however, show that pre-discharges are of low luminosity and rich in ultraviolet, and resemble the spark leader inception and extension sequence in air gaps. This is because the current limitation by low values of surface-layer conductance can inhibit the transition to an arc at the pre-discharge stage. An alternative model using simplified voltage-current equations is developed to represent these spark properties rather than an arc characteristic. The predictions of the partial-arc model and this new approach are both tested here, using published experimental data from light-pollution fog tests in the authors' laboratory. For this purpose, the partial-arc theory is further developed in order to reveal its implied prediction for the increase of pre-discharge length with applied voltage before flashover. A corresponding relationship is obtained for the increase of spark leader length using the new model. It is found that for light pollution, the dry-band spark model better represents the test data than the partial-arc model. Consequent predictions are made for the variation of flashover voltage with pollution severity in this regime, which may account for anomalous insulation failures.
Original language | British English |
---|---|
Article number | 5448096 |
Pages (from-to) | 417-424 |
Number of pages | 8 |
Journal | IEEE Transactions on Dielectrics and Electrical Insulation |
Volume | 17 |
Issue number | 2 |
DOIs | |
State | Published - Apr 2010 |
Keywords
- Arc
- Corona
- Flashover
- High voltage
- Insulation contamination
- Mathematical model
- Outdoor insulation
- Spark