The present paper gives the results of an investigation carried out by means of the restnking-voltage indicatorf (r.v.i.), on the parameters determining rates of rise of voltage for switches in a large 33-kV substation fed through 16 J miles of 66-kV cables and three 50 000-kVA 66/33-kV transformers. The measurements were made on one set of transformers, and the results are used to deduce the transients of restriking voltage under the worst service conditions and are applicable in general to transformer substations fed by a long transmission system.The highest rate of rise of voltage which appears to exist is 970 volts/microsec. (say 1 000 volts/microsec). This will only occur when the rupturing capacity on the busbars is relatively small, the fault being supplied through a single transformer. The highest rate of rise of voltage encountered when the heaviest fault currents are involved (supply through three transformers) is 625 volts/microsec. The circuit oscillation frequencies encountered are quite low (8-14 kc./sec), and it is concluded that for frequencies of this order it is quite safe to calculate transients of restriking voltage on the assumption that the main system parameters concerned (L v C^ are approximately equal to their values measured at supply frequency. There is, however, some evidence that even at these. relatively low frequencies the effective leakage inductance of the transformers is about 7 % less than the effective value at power frequency: and it is accordingly suggested that, until further information becomes available, caution should be observed in using 50-c /sec. values of transformer leakage inductance to calculate rates of rise of voltage when the natural frequency of the system concerned is much above 20 kc /sec. Methods are described by means of which transients of restriking voltage may be calculated in appropriate cases. CONTENTS
