Abstract :
From the above discussion it may be concluded that: Insulator trouble have been due mainly to cracking caused by expansion of metal parts, cement, etc. and to porosity. Certain designs with loose-fitting parts have been free from deterioration. Other designs should be so modified as to relieve them as far as possible from expansion troubles. Tough non-porous porcelain is desirable. The old method of basing everything on electrical tests should be abandoned. Severe electrical tests are often harmful. The electrical strength is often secondary to other characteristics. An electrical, mechanical and porosity uniformity test should be established, in which a small percentage of the product is tested to destruction from day to day to ascertain if it is running brittle or porous or is weakened by firing strains. For the very high voltages that are at present being considered, greater reliability may in many respects be anticipated than for the lower voltage lines. The lightning arc-over voltage and dielectric strength will be relatively higher and induced lightning voltages, sufficient to cause arc-over, will be less than on low-voltage lines. Increasing the number of units in series decreases the probability of complete string failure. Uneven distribution of voltage on the string becomes more serious at the higher voltages because of the high stress on the unit near the line. For strings of more than four or five units the stress is practically a constant percentage of the operating voltage independent of the string length.
