Remote leakage current detector for identification of insulators discharges

Performances of high voltage insulators can be evaluated by analyses of leakage current (LC) patterns. When insulators are polluted, the development of the LCs may lead to discharges causing insulator flashovers. This paper proposes a simple method of identifying the state of discharge activities through remotely sensed LCs. A remote detector, which is capable of detecting LC by capturing resultant electromagnetic field, was first developed. Preliminary tests were conducted at the laboratory considering typical near-by interference effects on polluted insulators such as effects due to corona and line current in order to test the robustness of the proposed system under the said interference effects. The tests were extended by obtaining different LC patterns by conducting salt-fog spray tests on five different types of clean insulators (made of porcelain, glass and silicone rubber), at two salinity levels (10 g/L and 40 g/L) and six different voltage levels (50%–300% of rated voltage). LCs were also measured on five different field polluted insulators through clean fog tests. In order to verify the proposed method, LCs were measured directly for all the test cases. The directly measured LCs were classified in to three classes: sinusoidal, non-sinusoidal and discharge. The remotely sensed LCs were also classified to these three classes using a kurtosis analysis. The statistical measure of kurtosis was observed as a viable and robust parameter to distinguish these three classes based on the remote measurement. The sinusoidal class had kurtosis values less than 5; discharge group had kurtosis greater than 12; the non-sinusoidal class had kurtosis values between these two levels. Hence, this paper suggests the possibility of using kurtosis values of remotely sensed LCs to identify the state of discharge activity prior to the insulator flashover remotely while proving to be robust under typical interference effects.