[Trade Journal]
Publication: General Electric Review
Schenectady, NY, United States
vol. 33, no. 7, p. 384-387, col. 1-2
Recent Improvements in High-voltage Insulator Design
Diagnosis of Insulator Faults—Complex Conditions Necessitate Extreme Care in Determining Effect of a Variable Factor—Contact Surface with Hardware a Vulnerable Point— Reglazing of Roughened Surface Increases Mechanical Strength and Decreases Electrical Failures
By D. H. ROWLAND
Research Engineer, Locke Insulator Corporation
IT can be said without reservation that trouble with high-tension insulators has decreased greatly in the last few years. There is little doubt that many of the theories advanced to explain the failures of porcelain insulators in the past were decidedly erroneous. No part of the insulator and no feature of design and assembly but were at some time or other, by some one or other, credited with the failures. Unfortunately a general acceptance of some of the explanations retarded, rather than aided, the correction of the fault. Specifications were frequently drawn up to lay such stress on one particular characteristic that other characteristics had to be neglected to produce insulators capable of meeting the stressed requirement.
It is now conceded that in the design of an insulator a compromise must be made between several factors if satisfactory service life is to be obtained. It is also definitely recognized that many of the reasons previously advanced to explain porcelain failures were only partially correct, and that the improvement in service came from the eradication of certain design or assembly features, from improved body composition, and the development of more exact methods of factory control. Only a few years ago it was impossible to run unassembled porcelain disks through more than two or three cycles from boiling water to ice water without considerable breakage. Today such failures are practically unknown.
The improvements which are evident in modern high-tension insulators were not the result of hap-hazard experimenting or the following of some fanciful theory. A great deal of painstaking research work in ceramics was essential and, owing to the highly specialized field of electrical porcelain, little assistance could be obtained from the ordinary ceramic laboratory. Practically all the work which has had any bearing on insulator improvement has been done in the research laboratories of insulator manufacturers.
The research laboratory of the Locke Insulator Corporation has been constantly employed for many years in establishing the characteristics of porcelain bodies containing different proportions of clay, feldspar, and flint. While this work has resulted in decided improvements of the actual porcelain body, probably one of the most outstanding developments from this research has been the discovery of the unexpected importance of a certain factor in the strength of porcelain.
In order to determine the strengths of different porcelains, rods having a diameter of 11.s' in. were made up and broken transversely between supports 5 2 in. apart. In addition to revealing which body had the greatest strength, this series of tests also showed that a wide variation in strength with the same body might be expected to follow variations in the character of the glaze, despite the fact that the thickness of this coating was only about 0.006 in.
Some of the results of this work were published in the March 1929 issue of this magazine under the title "The Influence of Glaze on Insulator Strength." Table I, which is here reprinted from the former article, gives an idea of the variation introduced by changing the glaze. The strengths shown in the different columns are for rods of identical porcelain coated with glazes of various chemical compositions.
The chief difficulty experienced in making tests on ceramic materials is that of eliminating all extraneous factors so that the results obtained may be known to result from the variable under consideration. For instance, in comparing changes in the strength of suspension insulators with variations in some point of design, the greatest care must be exercised to prevent differences in drying conditions, fluctuation in firing, or any other variation, however slight, in the handling of the insulators from raw material to time of test. When testing porcelain rods, for example, all test specimens were carried through the plant at the same time and fired intermingled in the same saggers. With insulators, the problem becomes more complex. Not only must the ware be carefully carried through the making and firing processes, but the cement and selection of the hardware parts must be controlled with equal care.
In the tests on suspension insulators which are described in this article, the clay was taken from the same press, run through the same pug mill, and the molds filled in strict rotation in an identical manner. As the molds containing the plastic clay passed through the machine, they were given consecutive numbers which they carried through all the subsequent processes. The ware was placed in the drying carts and later in the kilns so that the odd and even numbered insulators were equally distributed. The variable under study was limited to the odd-numbered units. The other group, comprising the even numbers, was tested to check against these results, it being assumed that no matter what extraneous factors entered into the making of the total number of insulators their influence on each lot would be approximately the same.
The factor under consideration was the treatment of the porcelain surface which comes in contact with the cement used for attaching the hardware. It has been standard practice for some years to obtain this cement grip by the use of fired porcelain ground so that the particles are about 1/32nd of an inch in diameter. This porcelain sand, or grog, is stuck to the green ware by a mixture of glue and glaze. As the ware goes through the kiln, the glue burns out and the glaze melts and attaches the sand to the surface of the ware. The glaze used in the tests described was almost as strong as that used on the rods tabulated in Column 5 of Table I.
A preliminary strength test was made on porcelain test rods having surfaces which were sanded in the usual manner. It was immediately found that the presence of the sand particles so decreased the strength of the rods that they were practically no stronger than unglazed porcelain. Following this, similar rods were prepared with the difference that after sanding the rods were again dipped in glaze. As the result of a large number of tests, it was found that this treatment raised the strength of the rods approximately 20 per cent above those sanded in the usual manner.
In an effort to determine what bearing this had on the strength of suspension insulators, the group of units that have been described was made up. Half of these insulators had the regular uncoated sanded surfaces and the remainder had the sanded surface reglazed. For the purposes of these tests a longer loading schedule than that laid down in the A.I.E.E. specifications was adopted, as it was thought that the results would be a better criterion of what could be expected in service. The units were connected together in strings, the insulators to be compared being alternated. The strings were then hung in a frame arranged with levers and weights by which a tension load could be applied. Fig. 1 shows a string on the frame under load. The initial load in this case was 10,000 lb., and each morning an additional 1,000 lb. was added and each separate disk immediately flashed over at 60 cycles.
The results of the tests were decidedly in favor of the units having the reglazed sand surfaces; but in older to check this result a second trial was made in exactly the same manner. Again the data showed conclusively that the treatment of the sand in the manner described was an essential factor in the performance of the insulator. The importance of running the test units through the plant in one group is made apparent by the results obtained in these two trials. While the difference between the treated and the untreated insulators is quite evident in each case, it is very plainly seen that it would have been difficult to have drawn absolute conclusions as to the merit of the treatment had the untreated insulators of one group been tested against the treated insulators of the other group. Fig 2(a) shows the results of the first trial and Fig. 2(b) the results of the second. The tension loads on the string are shown by the ordinates and the time of application by the abscissas. Along the curves are noted certain numbers and letters such as 2S and 3R. The former symbol indicates the electrical failure of two standard units while the latter represents the failure of three reglazed units.
The areas under the curves represent the number of days multiplied by the average load carried during the period by each insulator. By adding up these time-load areas for each group of insulators and dividing by the number of units under test, a figure is obtained which is an indication of the "average duty" performed by the insulator in question. The rectangles shown above the curves were obtained in this way, the minimum values at electrical failure being given just above the rectangles.
Regardless of all other factors of design or manufacture, this one ceramic detail goes a long way toward improving the performance and lengthening the life of the suspension unit. Figs. 3 and 4 show sketches of cross-sections of the ordinary type and the reglazed type of sanded surfaces. Examination of Fig. 3(a) shows that where the sand particles touch, a definite angle is formed. When it is considered that a piece of porcelain with glaze and sand applied is really a solid piece of glass-like material, as shown in Fig. 3(b), it can readily be seen that each of these angles represents an incipient crack which when stress is applied in a certain direction will tend to creep through the entire mass.
This is actually the case and accounts for the low values obtained in breaking sanded rods. When a coating of glaze is applied over the surface of the sand, a minute fillet is introduced between the sand particles and the depressions are therefore wavy (Fig. 4b) instead of angular (Fig. 3b). In addition to this, the re-glazing considerably strengthens the bond between the individual particles and the surface. Taken in conjunction, these two effects not only result in increased strength but in greater uniformity and consequently lengthened life.
Conclusions The electrical failure of a high-tension porcelain insulator as a general rule starts as a minute mechanical crack in the surface of the dielectric. Owing to the non-ductility of porcelain, such a crack will slowly but surely work through the material because of changes in external loading or temperature variations with consequent expansion and contraction of the porcelain. All cracks must start at the surface and therefore if the surface is strengthened in a manner which prevents the starting of cracks a great deal has been done to prevent a progressive failure. What in the past has been referred to as deterioration of porcelain has no doubt in many cases been due to a progressive crack caused by weak surface conditions.
In order to obtain the proper surface strengths there are two primary factors which must be taken care. of:
(1). The glaze used must fit the body in order to develop the maximum surface strength.
(2.) The physical shape of any roughening that is applied to the surface of the porcelain must be such that it does not tend to form angles which in turn will develop into cracks.
The tests that have been made up to the present time have been conducted on suspension insulators only. It is evident however that the condition of the surface inside the tops of pin-type insulators, or under the caps of pedestal switch insulators, is of equal importance. This is especially true where the units are subject to vibration.
Research on this subject is by no mean% completed and it is confidently expected that developments in the near future will still further improve the high-voltage insulator.
