In the investigation described in this paper, measurements have been made and statistical data obtained on the multiple character and wave shape of the currents in direct strokes and the currents discharged by lightning arresters. These data, in conjunction with other data on the field performance of arresters, have revealed definite differences between the nature of direct strokes and the currents they produce in arresters. 1. Arrester discharge currents have fewer components than found in direct strokes. Only about 30 per cent were found to be multiple and not over 8 components were recorded, while from 50 to 70 per cent of direct strokes were multiple with as many as 32 components recorded. 2. Components of arrester discharges are of lower crest magnitude. Only 30 per cent of those measured exceeded 1,000 amperes, while for direct strokes 50 per cent exceeded 5,000 amperes. 3. The initial high current portions of the components of arrester discharges do not vary greatly in wave shape and are similar to those of direct strokes, having for both types of discharges times to half value which, in general, lie between 25 and 100 microseconds and average about 50 microseconds. 4. The long duration low magnitude portions of direct strokes vary over wide limits, both in magnitude and duration. However, for arrester discharges they are seldom present to any appreciable degree. This portion of the discharge may have current magnitudes as high as 1,000 amperes and measurable durations from 50 to 20,000 microseconds in direct strokes, while lightning discharges through arresters seldom last more than 300 microseconds. 5. There is definite evidence that system neutral ground conditions influence the currents discharged by arresters on systems, and that the most severe duty is imposed on arresters on ungrounded systems, while the least severe service conditions appear on four-wire grounded neutral circuits. This is shown by the fact that no long duration surge components have been recorded on four-wire circuits, while about 50 per cent of the records obtained on ungrounded circuits had components of long duration. It is also shown by the performance record of the new Autovalve lightning arrester. With over 40,000 arrester years service this arrester has experienced only 18 lightning failures or a lightning failure rate of 0.045 of one per cent. Of 27,000 arrester years experience on four-wire systems there have been no failures, while of 4,500 arrester years of service on ungrounded circuits there have been 11 lightning failures. 6. There is indication that one of the factors influencing the nature of a lightning stroke is earth resistivity or geological structure. Observations lead to the inference that lightning discharges of long duration are more prevalent in regions of high earth resistivity and ancient geological structure than elsewhere. Thirteen of the 14 long duration arrester current records and 16 of the 18 arrester failures due to lightning occurred on four systems in two regions of high earth resistivity for which there was only 1,078 arrester years experience. 7. These data indicate that it is only with a combination of high earth resistivity and type of system for which the transformers cannot absorb the long duration portion, that appreciable long duration arrester discharges and accompanying severe duty on arresters is obtained. The data on which this is based are not yet conclusive, but the indications are sufficiently definite that this probability should not be overlooked.
Read full abstract