Pulse Sequence Analysis - a diagnostic tool based on the physics behind partial discharges

Abstract

The analysis of partial discharge signals is a well-known tool to monitor the degradation of electric equipment. In contrast to the common practice of the analysis of data sets (phase angles of occurrence and discharge magnitudes) accumulated during a predefined time interval, the Pulse Sequence Analysis (PSA) is based on the evaluation of data sets in which also the individual sequence or partial discharge events are registered.The classical analyses are mainly based on the frequency distributions of the pulse heights of discharge events and the phase angles of the applied voltage at which they occur. Detailed analyses of partial discharge sequences reveal that - if liquid or solid dielectrics are involved - the phase angle of the externally applied voltage is not a meaningful parameter in the partial discharge process. The ignition of the partial discharge is determined by the local electric field at the defect site and not by the absolute value or the phase angle of the externally applied voltage.Experiments with different model defects as well as measurements of commercial apparatuses show that there are characteristic sequencesof partial discharge events. In many cases a systematic shift of the phase angles of occurrence occurs, which produces an apparently `statistic scatter' of the phase angles unless the correlation between consecutive discharges is taken into account. Sequence correlated parameters such as the voltage differences of the applied voltage or time differences between consecutive discharges are far more decisive parameters in the process. By means of the pulse sequence analysis, local defects can be characterized more precisely and single and multiple discharge sites can be clearly distinguished.A physically based model to describe discharge phenomena will be discussed and experimental results will be compared with results of numerical simulations of partial discharge processes.