Investigations on detection sensitivity of partial discharge defects in various high-voltage GIS configurations with synthetic air and SF6 insulation

Many test and environmental parameters as well as measurement equipment capabilities can have an impact on the sensitivity of PD measurements to identify different imperfections in the insulation of high-voltage GIS. Especially, in the case of defect detection based on ultra-highfrequency (UHF) measurements, some of these quantities can superimpose in a complex manner, and result in vanishing influences on a kind of overall detection sensitivity.

Therefore, dielectric investigations in different-sized HV GIS configurations can help to identify which possible parameters and their combination can have a dominating effect on the applied PD detection procedure.

The application of a two-step-procedure (CIGRE sensitivity verification) is common to characterize a PD detection ability for certain PD defects in HV GIS. However, this procedure compares measurements data (recipient) which have been initiated by fundamentally different signal sources (emitter). These can be pulse generators on the one hand when used to inject ultra-fast signals into PD sensors which are typically placed on the enclosure, and artificial PD defects on the other hand.

In the latter case, signal propagation has its origin in different locations within HV GIS, e.g.

moving particle on the enclosure, protrusions on the conductor, and particle on insulation on the surface of a solid insulation in between. These differences have already been identified in the past, but their quantification is complex, and investigations of their impact on detection sensitivities have not been conducted frequently. Some measurement results related to this parameter complex are presented.

Overall, this paper discusses PD measurements in synthetic air (Clean Air) and SF6 carried out in different GIS arrangements, from small-sized to big-sized GIS, and related investigations on the achieved results, based on signal amplitudes and frequency ranges for typical defect types.

Specific defect types have been changed in terms of their position inside the GIS arrangement.

The target of this paper is to present more comprehensive PD studies on artificial insulation defects and its results, and to contribute to a better understanding of the specificities in PD behaviour appearing in the considered insulating gases. Moreover, the impact of different geometric GIS designs and PD sensor parameters are reflected to decouple meaningful and critical partial discharges.

In addition, this paper reviews suitable PD detection techniques on HV GIS and their application as well as failure identification and determination of defined parameters for a reliable risk assessment.