[Trade Journal]
Publication: American Institute Of Electrical Engineers
New York, NY, United States
p. 1863-1870, col. 1
INSULATOR DEPRECIATION AND EFFECT ON OPERATION
BY A. O. AUSTIN
ABSTRACT OF PAPER
Investigation shows that insulator trouble increasing with time is not due to fatigue in the material under applied working loads, but rather to depreciation caused by the absorption of water by porous material or by the cracking of the dielectric from high internal mechanical stress set up by uneven temperature in the dielectric, or by greater expansion of cement or metal, or stress from a combination of these.
The shape of the dielectric may cause high maximum stresses under comparatively mild conditions, necessitating the working of material with a lower factor of safety than that permissible even in steel work. The high maximum internal stress under which insulators operate will cause considerable depreciation in some types through cracking, necessitating a careful study of the effect of depreciation upon the operation of the system. Trouble comes largely through the matching up of faulty parts so that the remainder of the insulator will be destroyed by a comparatively mild surge. Applying the theory of probability, it is then possible to obtain a relative operating hazard for the insulator under the same conditions or for varying degrees of depreciation.
An equation for the operating hazard may be developed which gives a good idea of the relative economic importance of the number of sections in the insulator, the magnitude of the switching surge and the rate of depreciation as affecting the reliability of the system.
The study of depreciation shows that routine tests which will tend to eliminate future depreciation, or refinements in the mechanical features of the insulator, are of far more importance in producing reliability than the designing of insulators to withstand extremely severe dielectric design tests, for insulators which may have extremely high dielectric strength will cause trouble through cracking from internal stress.
THE growing investments dependent upon the transmission line make the study of factors governing reliability in the insulator of ever increasing economic importance. In view of this, it is hoped that this article will throw some light on insulator failures and be of assistance in deciding on the refinements and size of insulator necessary for a given system.
Even with more severe conditions placed on the insulator, recent lines have continued to make a much better showing. Some systems have had no insulator failures while others have operated for a period of several years only to run into increasing trouble, in some instances necessitating the entire reinsulating of the line.
Where insulator trouble increases with time, it is only natural to attribute the trouble to fatigue in the insulator. Upon looking into the subject, however, it is evident that this is not necessarily the case, for the failures may be, due to a number of entirely different causes.
Suspension insulators give the best available data, although failures of pin type insulators have been far more numerous, in some instances a visual inspection showing cracks. developed in over 60 per cent of the insulators.
A few pin type insulators which have cracked after several years' service are shown in Fig. 1. An examination of this class of failure shows that there must be a strong radial force which splits the outer parts or shears the head of the cylinder from the side wall. This force can come from several sources, but as the porcelain does not seem to be appreciably affected either mechanically or electrically outside of the cracks it would appear that the breakdown is due more to a rather high stress, than to a low stress acting for a long time.
It is quite likely that the stress set up is that due to uneven expansion from varying temperature and from an expansion of the cement.
Porcelain, like most dielectrics, is a very poor conductor of heat, so if the insulator has become hot in the sun or from sur-face leakage, and is then suddenly cooled by rain there will be a considerable difference in temperature between the outer and inner parts. This difference in temperature may set up an average tensile stress of several hundred pounds in the outer members.
This average stress is hardly high enough to cause any damage but the shape of the dielectric may be such that a very high maximum stress is set up, causing a crack, as along AB in Fig. 2.
There is no doubt that the defect comes from internal stress, but as this stress may be set up by uneven temperature and expansion in the dielectric or by cement expansion, it is not easy to determine the cause.
Tests on the insulators show that there may be considerable difference in temperature between inner and outer surfaces which may account for destructive stresses in some instances, but hardly accounts for failures in other cases, particularly where insulators have been known to crack in storerooms and in protected places where there has been no sudden change of temperature.
It is also significant that trouble usually develops at railway crossings first. While it is possible that the insulators may be-come hotter, owing to the black surface or increased leakage, it appears that the sulphur fumes in the smoke attack the cement and increase the crystalline growth presumably of calcium sulphate. This crystallization causes expansion, setting up a stress which, combined with that due to difference in temperature, may produce exceedingly high strains.
It is not surprising that there should be some -trouble from cement expansion, for in general, little attention has been given to this rather difficult subject, and even where there has been, it is difficult to predict expansion from the formation of acids by static discharge in the cement or effect of continued weathering.
Owing to deficiency in mechanical strength, European porcelain does not seem to withstand the mechanical stress very satisfactorily, and it has been general practise abroad to avoid cementing, or to use special cements at greatly increased cost.
A microscopic examination of the cement sometimes shows ,a marked crystalline growth in the cement, particularly where the insulators are near railroad crossings, or in cement exposed to continued weathering.
Fig. 4 is reproduced from a photomicrograph of cement taken from the insulator on the right of Fig. 1. Taken at a magnification of 49 diameters a few crystal growths are visible. Fig. 5 shows a marked growth in the cement from a strain insulator. These insulators gave trouble in a little over a year, so it is evident that cement expansion played an important part in causing failure.
Fig. 2 shows a section of an ordinary pin-type insulator which usually cracks-along the line AB. It is readily seen that a contraction of the outer part or expansion from the cement will set up a very high stress along AB. As this stress is highly concentrated in most insulators, owing to two components at nearly right angles, the break often occurs through a very thick part.
The large cement spaces and shape of the insulator greatly increases the hazard or danger of cracking in the insulator shown in Fig. 2. In Fig. 3 is shown one of the later designs of insulators which have proved to be practically proof against lightning and it is evident that the stress set up by uneven heating or cement will not only be less, but the strength of the parts resisting these stresses will be greater.
This type of insulator has low working stresses, so it is possible that high-frequency disturbances which would cause a considerable heating of the cement in the insulator shown in Fig. 2 would have little or no effect on the insulator shown in Fig. 3. Also refinements in cement, or the elimination of a portion of the stress by dipping the ends of the shells in an elastic varnish or wax or the insertion of a cushion which would be beneficial in Fig. 2, would be entirely unnecessary in Fig. 3.
The material is so distributed in the later types of insulators that not only are exceedingly high tests obtained on the parts, but a small protecting air path is provided between conductor and pin to act as a safety valve for surges.
Fig. 6 shows an improved insulator, a section of which is shown in. Fig. 3, flashing over at 216 kv. for a striking distance of 14 in. (35.5 cm.), and having a total part test of over 300 kv.
There is nothing in practise to show that anything would be gained by increasing the conductivity of the cement in insulators of this type or metallising the surfaces to reduce heating from flow of charging current under high-frequency surges although tests on the oscillatory transformer may show that this is beneficial in the ordinary insulator in its high charging current.
Fig. 7 shows typical breaks in two different suspension insulators. On the left is shown a disk with crack in the groove at the shoulder from combined expansion of the cap and cement exposed to weathering. The stress is concentrated by the shape of the insulator and the cement groove. This groove affects the strength somewhat like a scratch in a pane of glass. Insulators may also be cracked slightly at this point by too high mechanical tests or rough handling, although there is no outward sign. The elimination of the cement next to the flange and the substitution of a sanded surface in place of the grooves for holding the cement greatly reduces the maximum stress.
To the right of Fig. 7 is shown a failure in an insulator of similar design but of different material. These insulators have transverse cracks in the bottom of the grooves, showing that the insulators would not withstand the elongation of the metal under the highest working temperature. A higher assembly temperature would improve this or preferably a sanded surface in place of grooves, for the latter would be free from the objection of high shearing stresses in cold weather.
A magnified view of a sanded surface is shown in Fig. 8. In addition to eliminating grooves which tend to concentrate both electrical and mechanical stress, this surface is of equal gripping efficiency in any direction. This latter property greatly reduces the maximum stress set up in insulators working under heavy mechanical loads.
As there are records of insulators working under loads which set up stresses of at least 50 per cent of the ultimate, it seems quite probable that most mechanical failures are due to stresses very much higher than have been thought possible.
It is possible that vibratory stresses which are most severe in dead-end insulators may cause a breakdown of the dielectric structure which, with the greater weathering, may account for the very much poorer showing made by dead-end insulators in some cases.
It is certain that, where the maximum stress is very high and fluctuating with temperature, it will be only a matter of time before the molecular structure of the dielectric will be destroyed. Where this is combined with an increasing stress from cement expansion it is apparent that failures may be very serious in time, although there is little evidence of this during the early years of operation.
While only defects from mechanical stresses have been considered, there are others of an electrical nature that are always present and in many instances are far more serious.
Good porcelain will withstand considerable heat, so material which acts as a resistor may pass electrical tests successfully, only to depreciate very rapidly in service owing to the absorption of water, which greatly lowers the resistance.
Porous material or that which lacks vitrification, while withstanding high voltages at time of installation, gradually loses its ability to carry electrical stress. Where the absorption is slight the insulator may have an appreciable amount of dielectric strength after a number of years, but where the absorption is large there will be little insulation in a year, or so. On some of the earlier lines there was a large percentage of porous ware, which accounts for the poor operation until this material was weeded out.
An investigation of one of the large systems showed that out of 2 per cent of insulators shown to be weak by the megger, at least 1.4 per cent were poor owing to lack of vitrification, and could be detected by the trained eye.
Of the remainder, part were defective owing to porous streaks or the developing of faults left by the burning out of lint or other impurities. This really made the percentage of insulators which were poor, owing to conduction, over 1.4 per cent. It was not possible to classify some of the remainder outside of those having failed by cracking.
Fig. 9 shows a small fault detected by the megger and later burned by a very small current.
Porous insulators, while forming by far the largest percentage of defective members in the better designed insulators, can be detected to a large extent by Mr. Gaby's megger method and removed from the line.
For the factory use, it is advisable to use a more sensitive instrument than the megger in order to detect insulators which have very minute defects or are only very slightly deficient in insulation.
Fig. 10 shows a galvanometer which may be worked on a very high direct voltage obtained by rectifying and charging from the peak of the wave. Surface leakage has given considerable trouble, but it is hoped that improved means for shunting this surface leakage current will make this method very valuable.
Failures from electrical stress may be gradual, for it is possible to puncture porcelain several times, where the flow of current is extremely small, before complete failure occurs. For this reason surges sometimes do considerable damage to a system and their effect is not noticeable until the cumulative damage causes a complete breakdown.
Since the static breakdown is in the nature of a very slight mechanical fault, it is not surprising that a comparatively low electrical or a mechanical stress will cause complete breakdown in time.
The success of some of the later types of insulators, however, shows that there is little to be feared from static puncture as compared to some of the other defects in the insulator.
In well-insulated lines, trouble comes not by the depreciation of the insulators as a whole, but by the matching up of defective parts in a single insulator or string such that the normal voltage or a switching surge causes the insulator to spill over or puncture. Just what the factors governing this are, was not apparent until some recent investigations caused this matter to be investigated and more thoroughly analyzed.
The investigation of some of the systems shows that there is practically no successive breakdown in the insulator, and that faulty members are about equally distributed throughout the string. This, with the total absence of punctures during several years' operation on lines like the Shawinigan Water and Power Company, Fig. 11, and the 140-kv. Au Sable Electric Company's line, Fig. 12, shows that there is little danger of puncture from high-frequency surges where the insulator is made up of six or more closely spaced, well tested sections.
Lines of this class have also shown that the modern line is practically lightning-proof, for these systems have not averaged one kickout in two years from the spilling of insulators, from lightning or any other cause. This performance has, of course, been much better than was thought possible when the lines were insulated, and shows that spill-overs which have caused so much trouble on some lines can be prevented on a new system at a comparatively low cost for insulation.
The big problem in line insulation is not so much to design for high-frequency but first of all to prevent depreciation as far as possible, or at least to minimize the danger due to matching up of faulty parts in the insulator, for it is evident that if an insulator has a large number of parts which become bad through absorption or cracking, trouble is sure to follow regardless of the fine showing of an insulator on high-frequency tests.
Since the line trouble on the better insulated lines will come from the matching up of parts in the insulators which have become bad with time, rather than from lack of dielectric strength provided by the design, it is necessary that we recognize these conditions in order that the good operation of the system be maintained or established economically. To this end it is well to consider the probability of trouble from this source and analyze the problem, in order that the relative importance of the factors governing the matching up of the faulty parts to the danger point may be obtained.
To be presented at the 302d Meeting of the American Institute of Electrical Engineers, New York, December 11, 1914.
