Experimental investigation of PD diagnosis based on consecutive pulses data

Nowadays partial discharge (PD) measurement and analysis has known as a most effective tool in insulation monitoring of high voltage (HV) equipments. Almost all types of insulation defects is more or less accompanying with PD activities, while each has its own special patterns. To have a reliable estimation of defect type, patterns with high recognition potential should be used. Recently some studies have been made on patterns which are based on difference of excitation voltage magnitude at the instant of occurring PD pulses. The reports show good results, however, more investigations are needed to make the method applicable in practice and ensure the uniqueness of mentioned patterns for each defect type. For this purpose some tests on artificial PD defects as well as real PD activities in power transformers have been experienced and the test condition effects on patterns have been investigated. The results of these experiments are studied in this contribution.