Advanced possibilities of synchronous conventional and UHF PD measurements for effective noise suppression

On-site PD measurements are usually intensively interfered by external disturbances such as corona from nearby systems and switching activity from the grid, respectively. Therefore, very often no clear assignment of PD sources to the tested object is possible. However, a substantiated risk assessment requires such assignment and, even more, an identification and location of the PD fault. As the conventional PD measurement methods according to IEC 60270 are sensitive to external interferences, UHF detection methods become more and more popular because of their high level of noise suppression. This is based on the fact that most noise signals do not contain frequency contents of several hundreds of MHz. Thus a special UHF sensor mounted close or even inside the object to be tested permits a sensitive focusing on partial discharges as external noise and disturbances are highly attenuated. Hence, UHF signal detection ensures a high level of sensitivity and selectivity, respectively. As a major disadvantage, UHF measurements cannot be calibrated in terms of PD charge magnitudes. With international standards implying limit values (pass / fail criteria) specified in pico-Coulombs, PD measurements need a valid calibration to deliver comparable results. The classical way to combine the UHF sensitivity and the calibrated IEC compliant measurement consists of two consecutive PD measurements using different frequency spectra. With a highly disturbed IEC-range, this may not be possible at all. With a modern synchronous PD measuring system it is possible to perform these two measurements simultaneously. Additionally, and as a main benefit, it is possible to use the UHF PD data as trigger source for the conventional electrical PD measurement. This allows detecting PD with a high sensitivity and selectivity in the UHF-range, but allocating the appropriate PD level in terms of pico-Coulombs with the calibrated IEC-compliant PD detection. Examples will be presented in the following.