48 BELL SYSTEM TECHNICAL JOURNAL 



overvoltage on a sound phase at the instant of grounding than would 

 be the case in an isolated system is, of course, of importance in this 

 connection. This has been examined from a theoretical standpoint 

 in the second subdivision of this section, with the conclusion that 

 there is no material difference in this respect. 



There remains the method of grounding in which a high resistance 

 is employed in the connection from neutral to ground. By "high " here 

 may be meant the "critical"^ resistance or one of smaller magnitude, 

 but still so large that in the event of a solidly grounded phase, the 

 sound phases are brought to subtsantially full delta voltage above 

 ground. There are probably few cases where electromagnetic indu- 

 tive effects due to accidental grounds on a power system are a matter 

 of importance, in which a neutral resistance small enough to avoid 

 this rise of voltage on the sound phases would be effective as a meas- 

 ure of relief. This method of grounding would thus not avoid the 

 electrically induced voltages which arise when the reactor is used, 

 although it would presumably be effective in preventing the spread 

 of trouble to other parts of the power system if positive operation of 

 selective relays is secured. Inasmuch as it presents fewer difficulties 

 from this last point of view than the Petersen reactor, grounding 

 through a moderate resistance has a definite advantage over the latter 

 method from the standpoint of inductive effects at fundamental 

 frequency, provided sufficient resistance can be used to limit the 

 electromagnetically induced voltages to tolerable values. 



Where this is impracticable from the standpoint of power system 

 operation, the relative merits of the two systems would have to be 

 decided by balancing the effective suppression of the transient electro- 

 magnetic inductive effects by means of the reactor, plus the expecta- 

 tion of occasional disturbances continuing over the intervals neces- 

 sary for the location and disconnection of the faulty line, against the 

 imperfect suppression of the former effects by means of the resistor, 

 plus the limitation to very brief intervals of electrically induced dis- 

 turbances otherwise the same as with the reactor. It is obvious that 

 the factors controlling such a decision would vary widely in different 

 cases, and that practical experience with both methods would be of 

 great value in estimating their relative importance. It is, of course, 

 possible that future development may remove some of the disad- 

 vantage at which the Petersen coil now finds itself in respect to the 

 matter of relay protection for interconnected networks. Such de- 

 velopment would presumably be of importance also to the critical 



* Le., the resistance for which a discharge to ground passes from the oscillatory 

 to the non-oscillatory type. 



