[Trade Journal]
Publication: Journal of the American Institute Of Electrical Engineers
New York, NY, United States
p. 480-482, col. 1-2
A Solution of the Porcelain Insulator Problem
By E. E. F. CREIGHTON, General Electric Co., Schenectady, N. Y. and F. L. HUNT, Turner Falls Power and Electric Co.
MANY YEARS ago porcelain insulators on transmission lines began to crack in great numbers. In the earlier cases a few disks in many strings had cracked before the transmission engineers had discovered the condition. Indeed, there was at that time no reason to believe that insulators would deteriorate in any way. As a result of these many unobserved failures which simply reduced the factor of safety in each insulator one accidental failure finally introduced surges which punctured many other insulators just on the point of failure. These failures were tens of miles apart. Thereby the whole system was put out of use.
Today these general failures are avoided by a systematic test of each disk at intervals of time depending upon the conditions of installation and the judgment of the transmission engineers.
The cause of most of the failures of these insulators may possibly be reduced to one condition, namely the presence of the Portland cement. To be sure, the Portland cement at times was only indirectly the cause when it simply supplied the moisture which very slowly distilled into porous porcelain. The most usual cause of failure of the disks, however, was due to a characteristic of Portland cement itself. When Portland cement is dried out it shrinks. When it is again wet it expands not only back to its initial size, but a little more. In each cycle of drying and re-wetting the cement increases in volume until it tightly fills the entire space between the metallic hardware and the porcelain. When this condition is reached an unusually warm day will cause an unusual expansion of the metal, and the resulting strain will be transmitted directly through the Portland cement to the porcelain. A crack will result.
Progress in decreasing these failures has been made by the manufacturers of porcelain insulators. For example, instead of resting the cap directly on the porcelain of the usual suspension insulator it is now separated by a slight clearance which prevents the cap from pushing the head off the insulator. This improvement is easily made and has no detrimental factor.
For the expansion of the pin the conditions have been possibly somewhat improved by the use of a layer of soft material next to the porcelain. Unless other conditions are made to conform with this change the insulator is weakened mechanically.
Improvements have been made in the matter of open porosity by greater care in firing. Some manufacturers have also endeavored to overcome this difficulty by glazing all surfaces.
In spite, however, of all of these conditions, the older insulators on transmission lines from all manu-facturing companies still fail in sufficient percentages to require the expense of a periodic test and examina-tion. There is an amelioration without a cure. In other words, the life of the insulator has been increased but the tests must still be continued.
The writers attempted to find several solutions of this problem and there is described herewith the satisfactory results of one of these methods. Porcelain disks and hardware were purchased, unassembled, from a reliable manufacturer of insulators. The porcelains were all examined relative to porosity. They were all highly vitrified, in fact there was not the slightest trace of open porosity in a single one of them. A few faults were on the side of overvitrification which manifested itself in a detrimental way by slight checks. Punctures, where they occurred, took place through these checks. Very severe high-frequency voltages were applied by means of the well-known oscillator. The losses due to these extra severe tests were reasonable.
The solution of the main problem—the prevention of failure of the present type of cemented insulators—lies in arresting the expansion of the cement. The easiest method of accomplishing this is by removal of the moisture from the thoroughly set cement and the prevention of its re-entering the cement. This was thoroughly done by impregnation under vacuum on a large number of the insulators and was less thoroughly done on a portion of the insulators with the idea of determining the difference in life, if any, due to the difference in thoroughness of the impregnation.
In the endeavor to reach 100 per cent results it was necessary to study the characteristics of Portland cement when impregnated by different substances. One of the commonest impregnating substances of an insulating nature is paraffin. (1) It was found, however, that a chunk of Portland cement that had been thoroughly impregnated with paraffin by vacuum and heat treatment and then broken would still absorb moisture through the broken surfaces. A better material was therefore looked for. Some of the pitches were found to give perfect results. Portland cement impregnated with pitch could be broken into small pieces and soaked in water for days without the slightest indication of absorption of the water.
As already stated, different methods of impregnation were applied. This work was done in the laboratory during 1917 and the spring of 1918. The majority of the insulators were given a thorough vacuum treatment under heat and were impregnated, under pressure with the intention and hope of getting 100 per cent perfection and a life of indefinite length—much more than twenty years. In order to get a forecast of the future, laboratory processes of aging were developed. A large wheel, twelve feet' in diameter, which is known in the laboratory as the Ferris wheel, was made and operated with the axis in a horizontal plane. Around the rim insulators were secured with bolts and these insulators were passed through a cooling box next to the floor and a heating chamber near the ceiling. The insulators were exposed to the air between these two extremes of heat and an actual blast of moist air was turned on them as they slowly moved around. Here was a device then that gave the extremes of temperature and moisture from 20 deg. cent. below zero to 120 deg. cent. above. The range of temperature and moisture was exaggerated and the rate of change of temperature and moisture was also exaggerated above that found in practise with the idea of hastening any effect that might develop in actual practise. It was estimated that the effect of a single revolution on this wheel would give the insulator more severe strain than would take place in six months in practise—in other words, that two revolutions were equivalent to a year. In this way artificial tests were made extending over more years than any of us will live to enjoy. In criticism of this method it should be pointed out that it does not give a thorough test of the distillation of moisture into porous porcelain, but it has already been pointed out that these insulators were absolutely without open pores and that consequently that particular feature was not of interest.
Another set of tests was made on these insulators which was more severe in the rate of change of temperature, namely by changing from boiling water to freezing water every half-hour. By these methods we satisfied ourselves that the inpregnation was effective and that the insulators were ready for installation.
THE MOST IMPORTANT CONCLUSIONS TO DATE
Eleven hundred of these insulators, carefully, marked, were installed on a 66,000-volt transmission line near the seashore in New England in July and August, 1918, and put into operation in November, 1918. Up to the present time there has not been a single failure among the lot.
Thirty-six hundred other insulators of the same design, and built in the same factory, 'were shipped direct from the factory, having been assembled at the insulator factory, and installed on the same line at the same time as the others.
A test made on the whole line about a year after it was put into operation showed 13-1/2 per cent of failures among the insulators which came direct from the insulator factory and ~~no failure at all on those which had received the above described treatment~~. The insulators were then about two years old.
Up to the present time there have been additional failures, amounting to 2 or 3 per cent, among the insulators which' came direct from the factory, but no failures whatever in the insulators in which the cement was treated.
Following are some comments relating to (1) the mechanical strength, (2) electrical tests treatment, (3) line testing, (4) aging of porcelain, and (5) open porosity of porcelain—which seem pertinent to the subject in hand.
1. Mechanical Strength. In discussing the mechanical strength of insulators as affected by impregnation of the cement, it is desirable to avoid the introduction of other factors which are independent of the treatment of the cement. For example, it is well-known that the geometrical dimensions play a leading part in determining the mechanical strength. A large head and a deep pin-hole give more bearing surface for the cement' and thereby increase the ultimate "pull" on the hardware. Therefore a type of disk and hardware will have a fairly definite mechanical strength if the cementing is done in the most favorable way.
If the cement is sufficiently set to transmit the force to give the ultimate strength of porcelain in tension (about 2000 pounds per square inch, 140 kg. per square centimeter) and, furthermore, fills the space between hardware and porcelain surfaces sufficiently snugly to transmit the mechanical force uniformly, what more can be asked' of the cement? The type of insulator will then give its maximum mechanical "pull."
To arrive at this desired condition the manufacturing details depend on the application of a few scientifically determined factors. To illustrate briefly: The problem of attaining the proper conditions of the cement was approached by experimenting first with the two extremes. At one extreme the cement was set rapidly in live steam at a pressure of 120 pounds per square inch (8.5 kg. per square centimeter). Most of the disks came out of this treatment with the porcelain cracked due to the expansion of the Portland cement. At the other extreme the cement was set for a few days only under conditions of normal room temperature. As a result, either the pin slipped out of the cement in the "pull" test (as was usual) or the cap slipped over the cement. Sufficient "set" and "snugness" of the cement may be obtained by a manipulation of the factors at either extreme—less time and steam temperature starting from the upper extreme or more time and temperature starting from the lower extreme.
It is impracticable to give details here in this brief account. The treatment must be adjusted to the type of cement. For example, some of the grey cements seems to have common characteristics. It is possible that refinements in the research would show variations in the cement of the same manufacturer at different times of manufacture. We did not go that far. White cement showed the greatest expansion.
The important point in the proposed solution of this problem is that when the cement is given a desirable "set" and "fit" the impregnation excludes moisture which would otherwise change the optimum conditions as the cement grew older. The impregnating material must, of course, be of such a chemical nature as not to attack the cement. If the factors which cause changes in the cement are excluded why will not the cement remain constant forever?
2. Electrical Tests. The insulators at the factory and also the specially treated ones in the laboratory were given severe high-frequency tests. Those authorities who have expressed their fear of permanent damage to the porcelain body by the severe electrical tests may feel inclined to explain the loss of over 13 per cent in only two years of life by attributing it to the severity of the tests. This would seem hardly tenable ground. The insulators with impregnated cement were also given the severe electrical tests and none has failed. Is it not more probable that in the assembly of porcelain and hardware in the factory the particular cement and the particular treatment of it expanded the cement to a snug fit and the alternate dry and moist winds of the New England coast caused a further expansion of the tightly fitting cement, followed by a resulting breakage of the porcelain?
3. Line Testing. Again it is desirable to emphasize the fact that 100 per cent perfect insulators are a good investment financially, while the extra money expended in improving the insulator is of questionable economy. In one case the expensive and dangerous line tests are not necessary, whereas in the other case the line tests are still necessary although they may be made somewhat less frequently. It takes just as long to test 1000 insulators containing one defective as with a number of defectives. The goal is no testing.
4. Aging of Porcelain. The aging of porcelain is a question frequently brought up. Distinction should be carefully made between the deterioration of insulators and the deterioration of porcelain. There can be no question regarding the former. There is, however, a grave question whether porcelain in itself deteriorates with age. The series of tests of the application of heat, cold, and moisture to normal porcelain, which may be described at a later date, gave not the slightest indication of any change in the porcelain structure. By normal porcelain is meant the combination of clay, feldspar, and flint in the proportion of about 5, 3 and 2 respectively and fired to a vitrification sufficient to do away with open pores. The time and temperature are of the order of 4 to 6 hours at about 1330 deg. cent. The tests used to discover any possible changes in the porcelain were: puncture tests on 60 cycles and 200,000 cycles, hysteresis tests, and impregnation tests of dyes on broken pieces of the porcelain. While it is tenable that improperly made porcelain, using certain grades of quartz instead of flint and introducing other foreign chemicals, such as might come in the ball clay or other ingredients, may deteriorate in time, we have not yet found any such results in our researches. Porcelain protected by glaze is naturally free from moisture. It is therefore, in an insulator, subjected only to the range of atmospheric temperature and the mechanical strains due to the suspended transmission lines. While silica assumes a number of molecular conditions at unusually high temperatures there seems little likelihood that molecular changes at atmospheric temperatures could be a source of deterioration. The chance of deterioration is such a slight factor as compared to more evident conditions that we may properly relegate it to a secondary position for the present in analyzing the faults in porcelain insulators. To be sure, if the porcelain is openly porous and moisture is gradually distilled into the pores and is there frozen, the expansion of the moisture in the cells of the porcelain may increase the porosity. This is not an argument proving the deterioration of porcelain with age, but an argument favoring the proper vitrification of the porcelain, and the glazing of the surfaces exposed to moisture to take care of those inevitable cases of slight under-vitrification.
5. Open Porosity of Porcelain. In conclusion it should be noted that the use of dry impregnated cement solves also the problem of deterioration of resistance of porcelain due to the absorption of moisture.
As an incidental auxiliary matter, the manufacturer, by request, supplied a few special, underfired insulators for comparison and a method of determining the under-fired condition, without destroying the porcelain, was tried. This laboratory test was of the nature of hysteresis loss in the material. A description of it is out of the scope of the present paper. None of these was installed.
Final Comments. Three to four years is not sufficient time to furnish absolute proof of the success of this impregnation method. More reliance is placed on the artificial aging test of the laboratory. There are points for discussion and debate and more work yet to be done. We present the matter as a progressive step.
To be presented at the Annual and Pacific Coast Convention of the A. I. E. E., Salt Lake City, June 21-24, 1921.
(1) U. S. Pat. 1,360, 896.
