Apparent charge magnitude in on-line PD diagnostics on medium-voltage power cables

Estimating PD charge magnitude by comparing the response on a real PD and the response on an injected calibration signal with known magnitude in terms of apparent charge is usually not feasible for on-line monitored power cables. The injection cannot take place at a representative position in the cable but only at the cable end, whereas the PD can arise at any position within the cable. Additionally, the PD calibrator should be connected to an energised cable. PD magnitude is viable information if behaviour of similar defects in different cable circuits is to be compared. The present paper discusses a charge estimate method for on-line monitored cables. The method is based on modelling the complete chain of events, i.e. from PD signal excitation at the defect location via its propagation along the cable connection up to detection of the PD signal. By adding a pulse signal injection circuit to the sensors at both cable ends, relevant parameters for modelling the PD signal can be extracted. Monitoring long cable connections is economic attractive but signal loss limits the sensitivity in terms of minimum detectable apparent charge. This issue is discussed on basis of a monitored cable connection over 5.5 km long.