Shunt Capacitors in Large Transmission Networks [includes discussion]

Abstract

Mr. Wieseman indicated that the d-c test does not stress the insulation in the same manner as the a-c test. We are in agreeinent with this, but believe that for maintenance testing, where the purpose is only to break down dangerously weakened portions of the winding, this difference between a-c and d-c stress distribution is not a serious limitation.In answer to the question by Mr. Johnson and Mr. Moses regarding the location of the 17-kv a-c failure following a satisfactory 40-kv d-c test, this failure occurred in the slot about 1 inch in from the end of the core.We believe that Mr. Crellin has provided a very complete and correct explanation for the "hump" in the curve for phase A, Fig. 3, where the direct volts versus megohms are plotted. Mr. Hill has also commented on the fact that the current points for the lower voltage tests are not as low as they would have been if more time had been allowed for the winding to become fully charged. However, examination of the curves of the paper show that the small difference in the currents at low d-c test voltages do not affect the final extrapolated curves. In the cases reported we were not using the lower voltage portion of the curves. Mr. Hill comments that our test points were obtained by reducing the d-c applied voltage to zero and then applying another increment of higher voltage. Reference to the second paragraph under "Test Procedure" in our paper will show that we generally followed the procedure that Mr. Hill recommends.J. T. Madill asked about winding temperature and type of insulation. All of our tests were made with the windings at room temperature, about 26 to 28 degrees centigrade. The 60,000-kva capacitor had class-B insulation consisting of asphaltimpregnated mica tape. The coils were finished with glass tape, with corona preventive treatment.We believe that the application of d-c testing to machine insulations will be developed into a very positive and effective tool for determining certain types of insulation damage, and for determining the necessity to rewind machines.As now used, neither a-c or d-c highpotential tests show the condition of turnto-turn insulation. We would like to see some good method developed for stressing this insulation on multiturn stator coils of large machines.IT IS well known that the most efficient operation of a power distribution network is obtained when the power factor of the load and throughout the ne-twork is approximately 100 per cent. Various steps, including the use of overexcited synchronous motors and shunt capacitors, are taken to obtain high operating power factors. Transmission lines and networks, however, when operated at the load and voltage levels required to give best economy, require large amounts of kilovars and operate at power factors substantially less than unity. At the supply or generating end of a given transmission system, the required kilovars can usually be obtained most economically from the generators by providing the necessary overexcitation capacity. At the load or receiving end of the system...