Power line corona noise prediction from small cage measurement

[摘要] ENGLISH ABSTRACT:The radio noise (RN) performance of an AC high voltage transmission line is a determiningfactor when accessing the reliability of the design. The conducted RN level across the terminalsof a PLC receiver, used in the teleprotection system of the line, is of particular concern. In thisdissertation, existing empirical and semi-analytical RN prediction methods are evaluated bycomparing the conducted RN levels of four South African lines to the predicted levels.Existing RN measurement protocols are consolidated and a new RN measurement protocol isproposed for the comparison. A prediction methodology, based on the EdF, EPRI and IREQsemi-analytical methods, is formulated. The semi-analytical methods are not able to accuratelypredict the variation of the conducted RN with frequency in the PLC band, nor are they able topredict the RN performance of short lines. A wideband excitation function, which considers thestatistical characteristics of the measured RN pulse train, is therefore proposed in thisdissertation. Having investigated the extraction of this excitation function from ESKOM'sMegawatt Park corona cage, it is concluded that a small corona cage is required. The design ofa small cage is presented and the extraction of the excitation function is described. Attention isgiven to assessing the effect of changes in space charge distributions. The effect of space chargecollision with an insulated cage on the extracted pulse parameters is linear, provided space chargediffusion is negligible. The onset streamer, positive streamer and negative streamer corona modeswere identified as the only modes with measurable RN in the frequency regime 150 kHz to 30MHz. Formulas are derived for the prediction of the inception gradients for these modes. The RNperformance of an ACSR Hare conductor is extracted from the small cage. The RN performanceof the Acacia-Koeberg line is also approximated in the PLC band.The dissertation concludes that the proposed wideband excitation function method is capable ofaccurately predicting the variation of RN levels with frequency. It can also be used to predict theRN performance of short line sections, as well as the relationship between RMS, QP and averageRN levels from a transmission line.

[发布机构] Stellenbosch University