[Trade Journal]
Publication: Journal of the Institute of Electrical Engineers
New York, NY, United States
vol. 49, p. 235-267, col. 1
HIGH-TENSION PORCELAIN LINE INSULATORS.
By J. LUSTGARTEN, M.Sc., Associate Member.
(Paper received 6th February, 1912, received in final form 27th February, 1912, and read before the MANCHESTER LOCAL SECTION On 12th March, 1912.)
SUMMARY OF CONTENTS.
Electrical Porcelain.
General Remarks—Manufacture—General properties—Electrical properties--Mechanical properties.
Line Insulators.
General remarks.
Pin Insulators.
Development—Electrical considerations in design—Forms of electric discharge: (a) In the dry state; (b) In the wet state—Protection against arc effects—Metal shed insulators—Shackle insulators.
Suspension Insulators.
General remarks—Advantages--Types: (a) Interlink type; (b) Cemented type—Electrical considerations in design—Distribution of Potential gradient of the series—Protection against arc effects.
Testing of Insulators. Measurement of voltage—Puncture tests—Testing under artificial rain—Insulators under natural weather conditions.
Appendix.
Potential gradient and density of electric charge: (a) Case of capacity with a dielectric between electrodes; (b) Case of capacity with a compound dielectric, i.e., porcelain and air, between electrodes.
ELECTRICAL PORCELAIN.
General Remarks.—Electrical porcelain is eminently suitable for permanent exposed insulation, for which it best fulfils the necessary requirements. It is not hygroscopic nor inflammable, and does not alter in composition with time; it is unaffected by atmospheric action, and by practically all chemical substances. Under electrical stress its resistance to puncture is high, and even with prolonged application it undergoes only a slight rise in temperature, which is of little consequence. The material has one bad characteristic—it is brittle. This, and the difficulty of manufacturing very intricate pieces, militate against its universal adoption in the electrical industry. Still, by proper design the objections caused by brittleness can to a large extent be overcome.
The essential components of the insulator "body" are kaolin (china clay), some form of silica (such as quartz or flint), and felspar. The mixture is raised to the vitrifying temperature, causing chemical interaction between the silica and silicates, which ultimately produces a homogeneous mass. Ordinary felspar is potassium aluminum silicate. Pure kaolin is hydrated aluminium silicate produced by decomposition of felspar under the action of weather, but even in the best natural state it is contaminated with undecomposed felspar, grains of quartz, mica, etc. Quartz and flint are practically silica.
The different character of the various ingredients used in making the insulators in England, Germany, and America necessitates a difference in the process of manufacture. In England the body is made up of kaolin, plastic or ball clay, flint, Cornish stone, and felspar. The ball clay, which is in a preponderating amount, gives plasticity, makes the mixture easier to work, and renders the finished insulator less liable to mechanical damage—i.e., less brittle. The Cornish stone is a fusible natural rock consisting mainly of felspar and quartz. On the Continent the mixtures generally contain only kaolin, felspar, and quartz, and the resulting porcelain is whiter, though this is not so much due to the composition as to the method of firing, while the electric strength is probably greater.
Manufacture.—After calcination and preliminary crushing, the flint, felspar, and other rocky ingredients are water-ground until they are fine enough to pass through silk lawns 100 strands to the inch. The clays are beaten up with water in a large iron churn or " blunger" to a creamy fluid, known technically as "slip." All the ingredients are then mixed in proper proportions in the presence of water, to obtain the intimate mixture necessary for proper manufacture. The milky fluid or slip is passed through sieves of brass wire, phosphor bronze, etc., and then through a trough containing a row of magnets to arrest iron particles, after which it is pumped into a filter press under a pressure of about 4-1/2 atmospheres. Some manufacturers then store the resulting sheets in cool, damp cellars. The next process is to knead the sheets in a pug mill, which solidifies the mass and drives out any enclosed air-bubbles. The clay is cut into slices, which are thrown heavily one upon the other on a "wedging" slab. The several processes described are for the purpose of rendering the mass perfectly plastic and homogeneous.
In English practice the first operation in the manufacture of the insulator is to manipulate on the potter's wheel a clay ball of proper weight to the shape either of a dish or a solid cylinder. After this the work passes to the first drying-room, which is kept at a temperature of 80 deg F., to expel a portion of the moisture. The turner thus receives the pieces at such a consistency that they can be placed in a cup chuck or on a "dicing" lathe. The bolt-hole is then tapped in the interior of the solid pin insulator, or the lowest shed of a multi-piece insulator. The insulator or its component parts go to a second drying-room to be slowly and perfectly dried before going to the kiln to be fired. They undergo scrutiny by “fettlers”; the slightest sign of flaw condemns the piece. At this stage they are as strong as blackboard chalk. The articles are placed in coarse fire-clay boxes or saggars, which are stacked concentrically in a vertical reverberatory furnace; the doorway is walled up and heating commenced—at first gently to expel the remaining moisture, then at a dull red heat to drive off the water of constitution of the clay, and the material is then raised to a vitrifying temperature of 1,250° C. to 1,300° C., depending on the composition of the mass. This "bisque fire" occupies from two to three or four days, when the kiln is allowed to cool for a similar period. The "biscuit" insulator, which has shrunk about 16 per cent., linearly, is examined before glazing. In this biscuit state the insulator, though hard and of good electric strength, has a surface which attracts moisture, though without sensibly absorbing it; this surface would gather dirt and soot, and owing to the superficial irregularities would be very difficult to clean. For this reason, and no other, a coating of smooth and durable glaze is used. The ingredients of this glaze are kaolin, borax, felspar, whiting, and lead monoxide, which, being readily fusible, produces the glassy surface. The glaze may contain different metallic oxides to give either a yellow, brown, green, blue, or black colour. A brown colour can also be obtained by using a red marl in the clay body, such as is employed on vessels for mast shackle insulators (see Fig. 23). These oxides are non-conducting and therefore do not affect the insulating properties of the porcelain. The coloured glazes have the disadvantage of being more fusible than the white, and therefore exhibit uneven depth of colour, especially at the edges. Muffled kilns for coloured glazes appear to produce better results. Parts intended to be cemented to metal are left unglazed by wiping the glaze off with a sponge before firing. The glazing or glost oven is fired to a lower temperature than the first oven.
In German practice there is a difference in the methods following the wedging stage. The clay is put into a plaster of Paris mould of two halves of the shape of the finished part required, but about 20 per cent. larger. The clay is pressed together and the mould is allowed to stand until the clay has shrunk somewhat owing to moisture being absorbed by the plaster. The mould is opened and the shell allowed to dry without employing heat preparatory to burning in the first oven. The shells taken from the mould can be stuck together by the help of fluid clay or "slip," if intended to form a solid mass, otherwise the shells are cemented together after the final firing. Fig. 1 shows an insulator made up of three sections stuck together, the neck and bolt hole being cut after the drying stage. The first oven is at a temperature of only 800 to 900 degC., which serves mainly to expel water of constitution from the clay, but leaves an unvitrified substance which is only about as hard as a clay tobacco pipe, the surface being rough and porous to water, the porcelain is then dipped into the glazer which in this case consists only of felspathic materials and quartz, an is subjected to a second firing at a higher temperature, about 1,400 to 1,500° C., which finally vitrifies the body and matures the glaze at o operation. Of the coloured glazes, brown and green seem to be most favoured, as they are the only ones that can be successfully produced under these manufacturing conditions.
The American method is very similar to the German The conventional design of American insulators is by nesting thinner and more shells together than in the European design. Each shell is made from a mould, which is placed on a potter's wheel and the clay on the top shaped by a forming tool called a "jigger." Suspension types are also made in this way. The shell after being taken out of the mould is allowed to dry for about two weeks. The thread for the bolt-hole is cut out, or the holes fur the interlink types are made (see Fig. 26), which for the latter is a difficult operation. The insulator is then dipped into a glaze and fired to vitrifying temperature.
General Properties.—The shrinkage of the porcelain insulator depends upon whether it is turned or moulded, and also upon the composition, size, and thickness. From the stage before the bisque-firing to the finished stage it is 13 to 17 per cent. for lathe-turned, and for moulded insulator parts slightly less.
The density of porcelain is about 2.3 to 24. English electric porcelains will not absorb water whether glazed or unglazed. To detect absorption dip a broken piece of the insulator into water coloured with anilin or fuchsin. No colouration should be maintained at the break. Bad porcelain can be detected by its adhering to the tongue.
The linear expansion coefficient of hard porcelain is between 0.0000045 and 0.0000065 (C.), being less than that of glass. This is a recommendation, as the porcelain will withstand changes of temperature better. The more felspar in it, the greater the linear coefficient; thin has the opposite effect. The relative heat conductivity to that of silver (taken as 100 per cent.) is 0.045 per cent. Its specific heat is O.17.
Electrical Properties.—The specific resistance can scarcely be given as a property of the material, but rather as the property of the surface conditions existing on the test-piece. The surface leakage is more important than the actual leakage through the material, and the former has nothing to do with the material itself, but depends upon the humidity of the air and other circumstances affecting the surface. With clean glazed porcelain at normal temperature and humidity the specific resistance on a test-piece amounted to 2 X 10-12 megohms. As with all silicates, porcelain becomes conducting when raised to a red heat.
The dielectric constant obtained with direct current is about 5.3. With alternating current the value for a frequency of 50 is about to per cent. less, diminishing a further 3 per cent. for a frequency of 100.
The electric strength of porcelain will depend on the manner in which the clay has been worked and the disposition of the thick parts. The requirement for a high electric strength is a thoroughly homogeneous and vitrified mass, which, however, is difficult to attain with increasing thickness. The curve of puncture voltages for various thicknesses of porcelain plates tested between a sphere and a plate is shown in Fig. 2. In the English method of manufacture the glaze does not add to the electric strength, but in the German method the full strength is not attained till after the glazing, when the whole mass is fully vitrified. Should the glaze be of poor material it will in time become influenced by weather conditions. To obtain the necessary thickness for resisting puncture and to provide for a factor of safety, two or more shells arc cemented together, the insulator tint obtaining a more thorough vitrification.
Mechanical Properties.—The greatest mechanical strength of porcelain is in compression, being about 30 tons per square inch, only 30 per cent. less than cast iron. The exact tensile strength is difficult to obtain, owing to the small distortion of the test rod during firing, which subjects it also to a bending stress. From a large number of results the mean is about 10 tons per square inch. The tensile strength of porcelain from experiments on bending is about 3 tons per square inch. The shearing stress of good cement is about 1,600 lbs. per square inch.
Porcelain is harder than the rocks from which it originates. Its surface can be penetrated by a diamond point only under great load; the hardness of the vitrified Continental porcelain being about 5 to 10 per cent. greater than the outside glaze. This hardness gives the surface its weather-resisting property and its permanency; even the brush and spark discharges do not affect its insulating power.
LINE INSULATORS.
General Remarks.—An insulator has to satisfy two .main conditions. First, it must withstand the mechanical stresses necessary to support the conductor, and, secondly, it must withstand the electrical stresses necessary to insulate it. Additional requirements are: (1) It should he able to resist atmospheric influences in service; (2) it should not be easily broken by stone-throwing, bullets, etc., nor in transport; and (3) its weight and cost should be as low as possible. The properties to be held in view are therefore great mechanical strength, high electric strength, small conductivity, small surface leakage, a size and shape to exclude electric discharges, and to fulfil the three other requirements mentioned above.
The long spans which have come into use subject the insulator to very great stresses. These stresses are due to the following causes (1) The weight of the wire coated with snow and ice (especially with aluminium conductors, which require a large diameter); (2) wind pressure and extreme cold; and (3) the horizontal pull of the wire. The latter stress is the most important, especially when the wire breaks. The stresses are exceptionally great at corners and dead-ends.
The insulator withstands a compression test best; hence in the pin-type the pin is threaded up into the head of the insulator, so that the porcelain is only in compression but not in shear. The insulator can be designed to withstand such heavy testing loads as 3 to 4 tons, the pin bending before fracture of the porcelain commences. The suspension cemented type (see page 260) can be designed to withstand a continued shear and tension up to 5 tons. In practice, conditions are arranged so that the wire will break or the pin will bend before the porcelain gives way.
The shutdown of a line owing to a punctured insulator, the waste of time locating it, especially when covered with snow in high mountainous regions, and the actual loss of the insulator have called attention to the necessity for a high electric strength, and a design in which the insulator must rather flash-over titan puncture. A safeguard against the porcelain fracturing in case of flashing over is discussed later. With small insulators the safety against puncture and flash-over is so high that the more important consideration is that of mechanical strength.
The conductivity of porcelain is so low that leakage through the material needs no special consideration in design.
The surface resistance must be sufficiently high to prevent neighbouring lines (especially telegraph and telephone) being influenced by the leakage current. In general, the insulator will have a high surface resistance it ample provision has been made against electric discharges on the insulator under adverse climatic conditions.
The provision against discharge will be best considered in the brief description of the development of the modern pin insulator.
PIN INSULATORS.
Development.—The prototype of the modern high-tension insulator was a petticoated insulator of the same cylindrical form as the low-voltage telegraph insulator, but somewhat larger. With a line pressure of only a few thousand volts the necessity for providing a safety factor against brush discharges and sparking over was unknown. Surface insulation was considered the most important point, and the insulator was designed accordingly. To increase the surface resistance a triple-petticoated insulator of the same form was tried, but as the narrow spaces soon became filled with insects, etc., a reversion to the original type was made. On raising the pressure the narrow spaces became tilled with a glow and brush discharge, these constituting a loss. Under rain this takes place earlier, and a brush discharge starts from the wet edge of the outer petticoat to the bolt; the spark which follows bridges the flanges. The action is assisted by drops of water from the outer petticoat, being pulled inwards by the strong field. In high-tension insulators the surface resistance is of subsidiary importance. The following experiment will show that sparking between two electrodes on a surface does not depend upon the surface resistance (except when moisture is deposited). Fig. 3 shows a porcelain plate on the centre of which is a disc electrode and at the edge a tinfoil electrode. A metal rod touches the latter and can be inclined at various angles 0 to the plate. For a constant sparking distance d of 6 in. the value of the spark voltages for various inclinations of the rod are given in Table I.
Here we have the surface resistance constant, but the spark voltage varies on account of the altered flux density at the centre electrode.
The next development was to incline the petticoats outwards in the form of a cone, thus moving the outer wet flange farther from the bolt whilst still keeping the inside surface dry. An increase in the dimensions also made this insulator comparatively heavy. Investigations at this time showed that the drops of water (which influence the brush discharge and flash-over) could be driven outwards by spreading the outer shed, giving it an umbrella shape; the weight was thus reduced. (Fig. 4 shows an early American type.)
With the use of higher pressures the solid insulator frequently punctured during test. In the attempt to obtain large sparking distances, the required thickness to prevent puncture was too great to be without flaws. The next development was an insulator made of two pieces cemented together. Fig. 5 shows an American type, and Fig. 6 a delta insulator produced about 1897. The wooden pin was early recognised as delaying the production of brush discharge. About this time the Paderno insulator of Ginori was introduced. The type is similar to that of Fig. 7 A, and virtually consists of two similar insulators of the early form one above the other. The type(1) has been largely retained with but slight modifications in the head-piece and the pin-piece. Lengthening the pin shed and extending the outer shed horizontally to increase the striking distance resulted in the insulator shown ill Fig. 8 (Bay Counties, 1900). This insulator endeavours to combine the alternative ways by which insulation of two electrodes for given spark distance can be effected. The pin-piece alone would be better than the above telegraph insulator, and the head-piece alone be equally ineffective. The combination of the two still has a considerable wet surface leading to brush discharges, as Nature does not confine herself to vertical downpour but also furnishes driving rain. It is useful for places near the sea, where salt deposits accumulating on insulator act as a conducting surface—the cleansing action of rain is therefore a necessity.
To minimise the wetting of the lower surface the intermediate petticoat or shed was retained. The first European 40-k.v. line (Gromo-Nembro, 1903) was equipped with the insulator similar to that shown in Fig. 1. This shape has been largely retained for pressures up to 40 k.v. (Fig. 7 B). In America the multi-shed insulator (Fig. 7 C) was quickly developed for higher pressures. Fig. 9 shows the Locke insulator for the Guanojuato transmission line (1904), and Fig. 17 an insulator with long shells designed by Mershon for the Niagara, Lockport, and Ontario 60-k.v. line. Fig. 7 D is the Hermsdorf 1908 type with a greater curved pin-piece and reduced length of bolt-hole. A similar 3-piece insulator is used on the 66-k.v. European line, Molinar-Madrid (1909).
The intermediate shed takes various forms. It may assume that of a projection on the pin-piece itself, as in the Rosenthal insulator (Fig. 10), which also shows a petticoated head-piece.
The Semenza (Italian) insulator (Fig. 11) has an exactly opposite function to the insulator of Fig. 8. The intention is to preserve a dry state of the surface. It is protected by a porcelain cover of a cheaper kind of porcelain, screwed or cemented to the insulator proper. Its wet flash-over voltage is thus considerably increased, thus enabling a smaller size of insulator to be used for a given line pressure. The metal-covered insulators discussed on page 252 serve the same purpose, from many points of view, in a better manner.
Electrical Considerations in Design.--Improvements in the process of manufacture have made it possible to produce a thickness of shell of 3/4 in. to withstand 100 k.v.;(2) above this small cracks or flaws in the interior arc likely to occur. The practice in this country and America is to make 2-piece insulators for line pressures from 10 to 30 k.v., and 3-piece from 30 to 50 k.v.; above this in America 4-piece insulators are made up to 70 k.v.; but suspension insulators are rapidly superseding the pin type from 6 k.v. On the Continent the single-piece insulator is general up to 20 k.v., and the 2-piece to 60 k.v.
The necessity for providing large air-spaces is seen from the results in the Appendix. Consider the flux passing from the line wire and the tie wires at the neck of the insulator to the bolt. At the top of the bolt the flux passes wholly through porcelain; but further from the bolt it must pass partly through air. The flux density will depend upon the thicknesses of porcelain and air encountered in the path of the flux. At the neck the distance between the electrodes being a minimum, the flux density is greatest. The starting of a glow at the neck depends on this value, which can be reduced by increasing the thickness of porcelain. For other flux-paths the density is smaller; and when these paths traverse air and porcelain, this flux density will be smaller the smaller the thickness of porcelain encountered—i.e., the greater the length of air-path. The insulator with large masses of porcelain and narrow air-spaces will not only require a greater charging current, but will also have a tendency to glow in the air-pockets.
As the pin shell has the smallest diameter at the neck of the insulator, the flux penetrating will produce a greater potential gradient than the others. Thus it will have a greater tendency to puncture, as is shown by the following results : Out of 10,480 3-piece insulators tested to a dry flash-over voltage of 195 k.v. for 3 minutes, 4,172 failed, of which 2,317 failed in the pin shell, namely, about 55.5 per cent. Of the failures, 81.6 involved the pin shell.(3)
Taking the flux-path in air constituting the spark distance for the insulator in the dry state (Fig. 24), the flux density is still somewhat influenced by the porcelain flanges. The following experiment (Table II.) shows the influence of a medium of greater dielectric constant on the flux density of a spark-gap. A sheet of porcelain was gradually-brought up between two pointed electrodes separated 3.32 cm. apart, and the spark voltage was observed for different positions of the plate.
(The reduction was more marked with plate electrodes. For the same distance apart the flash-over voltage was reduced from 36 to 31.8 k.v.)
The necessity for a large cylindrical air-space between the bolt and the lowest shed is evident from the conclusions in the Appendix. The nearer the porcelain is to the bolt—viz., the electrode—the greater is the total flux. A narrow air-space may lead to brush discharges within, especially when the lowest shed is wet. Curving the lowest shed outwards, a greater striking distance between it and the bolt is provided.
Curved sheds have, on the whole, greater advantages than straight. Rain falling at 450 can be deflected away from straight sheds, but with curved sheds there is a splashing action, with a consequent slight wetting of under surfaces. With the straight-shed insulator insects and dust may lodge where driving rain cannot reach, but this is not the case with the curved.
A slight conical formation of the head-piece helps to keep the under surfaces dry. With the head-piece wet, the total flux is increased, For paths near the flange the flux density will be considerably increased, owing to the point action of the drops of water. If the under surface of the head-piece and the upper surface of the intermediate become wet, then the distance between the electrodes is considerably reduced. In the air-spaces between the intermediate and the pinshed and between the latter and the pin, the flux density may produce discharges. The shells concerned are subjected to greater stress. Finally, when all surfaces are wet except the inner of the bottom shed, the latter will undergo maximum stress. The flux density from this flange to the bolt will be greatly increased. Both leakage and displacement (charging) currents will become great, resulting in great loss (see page 272). The latter is an exceptional condition which might only occur with thawing snow or hoar-frost deposited between the shells.
Long insulator parts, as in Fig. 17, have a large capacity—viz., produce a large total flux—since porcelain replaces air in the long flux-paths. In these long paths the flux tends to seek stronger fields, and thus a crowding takes place, putting greater stress on the parts, especially the pin-piece. Experience shows that long insulator parts do not resist sudden stresses well.(4) Shorter insulators give a better distribution of the flux, and enable the bolt to be shorter, thus securing additional mechanical strength. Large diameters give increased capacity. Wide and high insulators both give increased electric charge, and flash-over takes place by surface sparks (see page 276).
FORMS OF ELECTRIC DISCHARGE.
(a) In The Dry State.—The two forms of electric discharge—the brush and surface spark—arc responsible for the flash-over of an insulator in the dry state, the brush predominating in the case of small pin insulators, the surface spark for very large insulators (Fig. 12), and a combination of the two forms for medium-sized insulators.
As the pressure on the insulator is gradually increased a glow appears at the neck and extends to a surface brush down the shed, as in Fig. 13. The glow may also commence within the pin shed and at joints if the flux densities are high and are sufficient to produce the requisite ionisation.(5) With increase of pressure on the top shed streamers or surface sparks will form. The requisite pressure is dependent upon the thickness of porcelain, length, and disposition of the sheds. Meanwhile brush discharges start from the bolt and the line wire, commencing earlier from the wire if placed in the side instead of the top groove of the neck. On small insulators these brush discharges produce the pilot spark, which may reach to the neck via the surface brush or direct to the line wire, as in Fig. 13, according to the potential gradient acting along these respective paths. Thus for small insulators neglecting the reduction produced by the porcelain flanges, the curve for spark voltage and spark distance will be of the same order as the spark voltage distance curve for the line wire and bolt with surfaces at right angles. With medium-sized insulators the surface sparks commence on the top shed and the pin shed; a brush discharge completes the flash-over by reaching from the top flange to the surface sparks on the pin shed. The glow at places of high density, as in the joints, gives rise to surface sparks with the largest insulators, and as the length of these sparks increases rapidly as the cube or fourth power of the voltage (see page 276), flash-over rarely takes place between the flanges, especially with very long sheds. The spark-voltage distance curve for large insulators—i.e., large distances—will rapidly bend over, as in Fig. 14 (see also page 256).
The length of the surface brush(6) is influenced by the humidity of the air, though not sufficiently to affect the flash-over of small insulators. For medium-sized insulators the flash-over voltage is increased 3 to 9 per cent., and for the largest constructed the maximum increase is about 20 per cent. with increase of humidity. When the humidity is 100 per cent. moisture is deposited on the surfaces, and a lowering for all sizes will be produced. Temperature has little effect, as the spark voltage varies inversely with the absolute temperature(7). A variation of 10 mm. in atmospheric pressure produces a difference of 2.4 per cent. in the spark voltage between small electrodes,(8) so this will hold approximately with line insulators.
(b) In the Wel Stale.—Rain alters the appearance of the discharges, the alteration depending upon the amount and the direction of rainfall. With a fine drizzle and no wind the top surface becomes wet and the glow and surface brush disappear, but give place to glowing drops of water at the edge of the top flange. These point in the direction of the flux, and, becoming charged, are forced away with a velocity increased by the flux density and diminished by their size.
The flash-over voltage is reduced on account of the diminished spark distance and the point action of the elongated drops. With a more intense rainfall the brush discharges take place from glowing raindrops of one shed to the moist part of another, and the spark selects the shortest path to these conducting places (Fig. 24). If the under surface of the top shed be also wet, there will be no glowing drops at this shed. On lowering the pressure and completely wetting the sheds, except inside the pin shed. there is no other discharge but the brush discharge to the bolt. With most of the surfaces wet the total flux and the capacity are considerably increased, and if the pressure is brought up quickly whilst the insulator is still under rain, the flash-over will take place by surface sparks as in Fig. 15,(9) although the normal flash-over under rain takes place mostly through air between the sheds. When the rain stops the moist surfaces are soon dried by the heating effect of the 'leakage current and by the action of the electric field. While this is going on the fluctuating surface streamers may be seen between the moist parts. In a mist the diving action would be constantly going on. This is proved by the following experiment (see Fig. 16): A pressure of 26 k.v. was applied to a 13-k.v. insulator in a damp atmosphere. The leakage loss fell from 40 to 3.5 watts in 10 minutes. The pressure was increased by 50 per cent., and maintained for 5 minutes. Then, on reducing the voltage again to 26 k.v., the leakage loss was less than 3 watts, but gradually crept up to that value. Hence the insulation resistance of the insulator in a damp atmosphere is a function of the voltage(10). If steam be blown against charged electrodes it is repelled away, the action being greater with pointed electrodes. Insulators with the bottom shed close to the bolt have a fairly dry space within, so that with this shed moist on the outside a brush discharge may start in the bolt-hole at the line pressure. This has been observed in this country on damp days with the older form of insulator. An insulator which becomes wet by being in a damp atmosphere will show when quickly subjected to the line pressure a glow over the whole of the surface owing to the leakage current, but the drying action comes rapidly into play.
All these forms of discharge—namely, glow, brush, and spark—have a reddish tinge: they are less intense than those obtained with the insulator dry, because of the resistance of the water in series. Their actinic value being lower, much longer photographic exposures are necessary.
PROTECTION AGAINST ARC EFFECTS.
A high potential surge due to a direct lightning stroke or to induction in the neighborhood of the line may cause one or more insulators to flash over, on account of the difficulty with which the charges travel along the line to the protective devices (arresters, etc.). The potential stresses setup may be sufficient to puncture one or more sheds of the insulator; this puncture results in the formation of the arc. There have been cases, however, where a shed of. a large and long insulator has been shattered by lightning without puncturing the insulator and without producing a flash-over—the effect being similar to that caused by a hammer blow.(11) The arc tends to be drawn to an intense part of the electrostatic field, as can be seen in Fig. 13. Hence the end of the arc striking the pin tends to run up within the pin shed and, with great plant powers, the heat produced breaks the porcelain sheds from the bottom upwards. Since insulators are now designed to flash over —rather than to puncture—the question of affording complete protection against damage by arcs is important. A simple method has been devised by Nicholson. He uses two metal rings concentric with the insulator, the lower (of greater diameter than the insulator) attached to the pin, and the upper ring (somewhat larger than the neck of the insulator) suspended from and connected to the transmission line (Fig. 17). The rings serve as a safety-gap for the arc. The lower ring can be placed so high and close to the insulator as to reduce the flash over voltage, thus forcing the arc to pass to it from the upper ring or the cable. Fig. 18 shows the transference of the arc to the lower ring --no upper ring being employed. There is still the risk of damage to the head-piece, which might be obviated by the use of a larger upper ring instead of the small one suggested by Nicholson. Fig. 19 shows the manner of using the two rings. The dry flash-over voltage has been reduced from 114 to 102 k.v., and the wet flash-over voltage under standard precipitation is not less than the normal value without the ring. viz., 88 k.v. That this device has been effective has been shown he tests made on the 60-k.v. 3-phase transmission line of the Niagara, Lockeport, and Ontario Power Company with a 30,000-k.w. arc at the line pressure. In no instance was an insulator damaged. With or without the small upper ring the arc would travel out on the cable a distance of 12 ft. in 4 seconds with a breeze of 3 miles per hour, but the scarring would be greatest where the arc first started. A large upper ring would save the cable or tie wires from the initial hurtling.
It is interesting to give the performance of the insulator used because the experience on this line has affected subsequent designs. The insulator(12) consists of 3 shells cemented together and has the following dimensions:—
Diameter of head-piece -- 14 in.
Diameter of intermediate shell – 13 in.
Diameter of pin shell – 11 in.
The total height from the edge of the bottom shell to the top of the head, 19-1/2 in.; length of intermediate, 12 in.; and of bottom shell, 17 in.
Each shell had withstood a factory test of 75 k.v. for 3 minutes before assembling, but the complete insulator was not tested. Its dry flash-over voltage is 195 k.v., and wet 12.k.v. On testing with 195 k.v. there were many failures due to the high stress on the porcelain. It be noted that the test voltage is more than three times the line voltage—a safety factor much higher than used at the present day for 60-k.v. line pressures. The lower ring was then used; it reduced the dry flash-over to 160 k.v. and considerably diminished the tendency to puncture which generally takes place in the pin shell. Tests were made on 800 insulators first with and second without the ring—the flash-over voltages were 160 and 195 k.v. respectively; the pin shell failures with the ring were 3 per cent. and without 22 per cent. The ring employed was 26 in. in diameter, 3/8 in. thick, and placed 2 in. above the edge of the bottom shed. This size and height was chosen so that under normal precipitation the arc would strike the ring and give a flash-over voltage of the same value as without the lower ring, and also a dry flash-over voltage of 160 k.v. as indicated. The ring placed 21 in. above the base reduces the effective length of the insulator. Therefore a shorter insulator has been designed, which replaces the topmost insulator of the line. The experience on this and on other lines employing mast-top pin insulators, without the overhead grounded wire, and without arcing rings, shows that more than three-fourths of the insulators broken were top insulators. The new design is of the same diameter but 6 in. shorter, and consists of four shells. Its dry flash-over voltage is 190 k.v., and wet 105 k.v. It is interesting to note that the dry flash-over is altered very little by lessening the length. The surface sparks on account of the longer pin shed and the smaller thickness of porcelain between the pin and line (giving a greater capacity), reduce the effective air distance from wire to bolt. An arcing ring provided on the new design reduces the dry test value to 160 k.v. (Fig. 20).
The earthed metal ring effects a redistribution of the field. It takes up part of the flux which would otherwise enter the pin shed, thus producing a less flux density and causing a more uniform distribution among the other shells, with less danger to all the shells. The large upper ring shown in Fig. 19 distributes the flux still more uniformly, and will relieve a surge of charging current due to sudden disturbances.
METAL SHED INSULATORS.
The use of a large tipper ring in Fig. 19 suggests its extension to a metal shed. Fig. 21 shows the Hermsdorf patent metal-shed insulator. The upper surface of the ordinary insulator is practically conducting when wet. Its function is, besides the mechanical strengthening of the head of the insulator, to protect the other sheds against rain, and to provide a dry under surface. The metal shed of large diameter effectively protects the porcelain parts against wetting, and in dispensing with the weight of the porcelain head and also being shorter for a given line voltage, its weight and price arc less. These advantages are greatest at voltages between 40 and 70 k.v., but below 25 k.v. they are not maintained. The diminution in leakage-path in this insulator is counterbalanced by the protection it offers to rain. Further, at the sharp metal edge the flux density is greater than with the corresponding porcelain head-piece, so that drops of water become more conical and will be urged with greater velocity outwards (see Fig. 42). The wet flash-over voltage is influenced by the direction in which the drops fall, and we should expect that the metal shed insulator will give a value not much less than when dry. The insulator in Fig. 21 (without the lower ring) has a dry flash-over voltage of 75 k.v. and wet 70 k.v. A larger insulator for 44 to 50-k.v. line pressure has a dry flash-over voltage of 110 to k.v. and a wet value of 98 k.v. Even in the heaviest rain the pilot spark starts from the inside surface rather than from the edge. At the edge of the head-piece of the all porcelain insulator there is a fluctuating variation in the flux density, as some points are wetter than others, but at the edge of the metal shed insulator the flux distribution is practically uniform. Thus the wet flash-over voltage for the metal shed insulator is more definite than for the all-porcelain insulator. The flux in passing from the bolt to the top shed will be more uniform with the metal shed insulator than with the other. In a comparative wet test of the two types for 80 k.v. line pressure, the ordinary insulator flashes over at 130 k.v., whereas the other shows only the initial discharges. Heavy surface sparks occur on the ordinary type, and are preceded by brush discharges at a much lower voltage than on the metal shed insulator. Since the brush discharges occur later on the latter, and since the production of surface sparks is minimized (being less wet), the insulator can be made shorter. The dimensions and weight for 80- and 55 k.v. line pressures are given below in Table III.
The distribution is also better for the metal shed insulator in the dry state. The flux density in the neck and top of the bolt-hole is much lessened, as the flux starts from a wider area, viz., from the inside of the shed, thus increasing the safety factor against puncture and hampering the production of surface brush discharge. The following test may be cited to show the improvement in the electrostatic conditions. With a line wire simply resting in the neck of an ordinary insulator the dry flash-over voltage was 109 k.v. Attaching a tie-wire round the neck increases the value to 116 k.v., and with a metal cap to the head 123 k.v. Replacing the porcelain shed by a wide metal cover to give the same sparking distance, the flash-over value rose to 126 k.v.
The shape of the intermediate porcelain shell—namely, its convexity to the metal—is explained by the discussion on design, pages 245 and 246. The nearer the porcelain shell is to the bolt, where the flux density is greater than adjacent to the shed, the greater will be the total flux produced. Its convexity gives it also an excellent protection against rain.
The loss due to displacement current is not greater than for the ordinary porcelain insulator of the same voltage.
An arcing ring, of about the same diameter as the metal shed, fixed to the bolt, will protect the porcelain against power arcs. The large cooling surface of the metal shed would prevent its being fused by the arc. Used in conjunction with a metal basket, instead of a ring, it would offer a good protection against breakage by stone-throwing.
Shackle Insulators.—Shackle insulators are used for curves, dead-ends, and similar points in a transmission line where the pull is too great for the ordinary pin insulator. Fig. 22 is for a line pressure of 30 k.v., and withstands a pull of 12,000 lbs. For heavier loads the insulators are used in multiple. To obtain maximum strength of this type, if the pin used be allowed free play in the hole and not fit snugly, only the porcelain at the point between the wire and the pin will be under compression. For higher voltages shackle insulators arc used in series. On a 20-k.v. transmission line (12) at Lobito, Portuguese East Africa, three 10-k.v. shackle insulators are used in series for dead-ending, etc. The suspension strain type is very suitable for points where the shackle insulator would be employed (see page 257). A simple form of suspension strain insulator shown in Fig. 23 is used in the British Navy for insulating the aerial wires in connection with wireless telegraphy. It withstands a compression test of 11 tons.
General Remarks.—The increase in line voltage to 110 k.v.(13) has brought about the introduction and development of the suspended type of insulator on account of the large increase in weight and cost of the pin insulator required. Above a pressure of 50 k.v. the weight and size of the pin insulator increase rapidly. The weights of insulators of some typical European transmission lines are given below:--
The Heimbach-Aachen line was an early equipment. The Spanish transmission line (Molinar-Madrid) is equipped with 3-piece insulators 14 in. diameter and 15 in. high.(14) On page 251 the details of the Niagara, Lockport, and Ontario insulator arc given; its weight is 59 lbs.
The large increase in size is necessary to provide a sufficient spark distance to counterbalance the influence of surface sparks which arc produced by large capacities. In Fig. 14 the curve shows the relation of spark distance (d of Fig. 24) and the flash-over voltage of insulators of practically the same type from English, German, and Spanish firms.(15) Curves for needle-points and 4-in. spheres are also plotted for comparison. It was pointed out on page 247 that for intermediate sizes a combination of the brush discharge and surface sparks causes flash-over, and for the large—the surface sparks. As the latter vary in length as the third and fourth power of the voltage, they will cause the curve in Fig. 14 to bend over rapidly. (The curve will thus cut the curve for needle-points.) The size and the weight of the insulator will increase almost at the same rate, namely, the third power of the voltage. The price will have a similar relationship. The following are the weights and prices of pin insulators from 50 to 80 k.v.(16):--
The weight and price of the suspension units, on the other hand, will increase linearly with increasing voltage (see page 266), each disc added for increased line pressure being identical. If the weight-voltage curves for the two kinds—pin and suspension—be plotted, they will intersect between 50 and 60 k.v. (according to the types of suspension units selected). The intersection of price-voltage curves is between 6o and 70 k.v. (The price of the suspension units for pressures less than 60 k.v. is greater on account of the metal parts.) Taking into consideration the necessity of a higher power, the extra cost of which is counterbalanced by the less cost for the cross-arms and smaller maintenance cost,(17) 60 k.v. may be taken as the pressure at which the suspension insulator becomes economical. As an instance of its adoption for pressures between 60 and 70 k.v., the Groba (Saxony) transmission line under construction will employ 3-unit suspension insulators. For the 66-k.v. Northern Hydro-Electric Company (North Wisconsin) line the same number of suspension units is being used.(18) Fig. 27 shows the suspension insulators in use. They hang vertically beneath the cross-arm and are allowed to swing freely. At intervals of about 1 mile the line is anchored to strain towers. There are about 5 to 10 ordinary towers to one strain tower. The strain insulator employed is usually larger, as it undergoes greater mechanical stresses. Fig. 25 shows the method of attachment, with the line wire hanging below the insulators. On account of the surfaces being wetter under rain, a greater number are used in series than in the suspension group. These strain insulators arc used also at corners and for dead-ending the line.
Advantages—The suspension type has other advantages. The unit construction is of paramount advantage, for if at any time it is desired to increase the line pressure the only change is to add the standard unit. Localities such as sea-coasts and near chemical factories requiring extra insulation can have an extra safety factor by addition of a unit. Further, for such places the suspension type is better cleansed by rain. This is not so with the pin type ; it necessitates a change of form and size. Moreover, with the latter, if a shed of the insulator is damaged the whole insulator must be replaced, but with the suspension group only the damaged unit requires removing. And whilst the damaged pin insulator may cause a shutdown, the safety factor of the group is sufficient to prevent it. Its simple construction and its comparatively small size give no difficulty in manufacture. Hence there is the advantage of shorter time of delivery and of the possibility of keeping a large number in stock.
The suspension group has a smaller capacity than the pin type, this capacity diminishing as each unit is added, but increasing for the large sizes of the pin type for higher line pressures. A greater flash-over distance gives a higher ratio of flash-over voltage to line voltage for a suspension group than for a pin insulator. Although a wide pin insulator gives a better rain protection than the narrower but larger suspension group, yet the number of dry surfaces of the latter gives a smaller surface leakage loss.
The suspension type has none of the mechanical difficulties of the pin and the cross-arm in the pin type. The manner of suspension allows the line wire to move freely under wind pressure. The flexible connection between the conductor and cross-arm should minimise any tendency to crystallisation of the line wire (especially with aluminium), which has been known to take place where it is fastened to the neck of the insulator.(19)
The suspension of the wire below the earthed cross-arm causes less disturbances of the line due to lightning. Many lines which have changed from the lower pressure with the. pin type to the higher pressure with the suspension suffer less from lightning. The addition of the earthed wire between the masts has lessened the disturbances still further. This wire in sonic cases is used to take the longitudinal strain of flexible mast construction.(20)
Types.—There are two distinct types in practice, from the manner in which the units are supported. One type—the link insulator—has two interlinked, semicircular holes for tic wires. Fig. 26 is one form. The second type has the metal parts concentric (Fig. 32, 34, and 35). These consist of a metal cap and a bolt cemented to the porcelain. A third type of suspension insulator, which has not come into use—at any rate for line work—is similar to the first type, but the tie wires are not interlinked. Electrically the type is better because of the lessened flux density and capacity, but mechanically they are much weaker. In the latter the porcelain is in tension, but in the former in compression.
(d) Interlink Type.—With the first type the interlinking has the advantage that in case an insulator breaks the tie wires will still hold the other units together, and, with the ample safety factor, prevent a shutdown. But against this there is the possibility that the tie wires may not interlink when one or more units puncture, and an arc may destroy the links. The tie wires may break from the constant rubbing in the holes and from a possible corrosion by ozone produced by any electric discharge in the holes (see Fig. 41).
Fig. 26, the Hewlett form of the General Electric Company, is a flanged disc with enlarged central portion.(19) It is made in two diameters, 6 in. and 10 in., of lengths 2-1/2 in. and 2-3/4 in. respectively. The larger unit has a dry flash-over of about 80 to 95 k.v., and wet50 to 56 k.v. It is rated for a line voltage of 25 k.v. Four in series (Fig. 27) are being used on the 80-k.v. Victoria Falls and Transvaal Power Company's transmission line (length 45 miles) in South Africa; but on a new line of the same company five are being specified. On a 100-k.v. transmission line of the Central Colorado Power Company (22) four are employed, and for the Grand Rapids and Muskegon Power Company five in series for 110 k.v. The 135-k.v. line, (23) Cook Falls to Flint, and Battle Creek, Michigan (total length of 190 miles), construction, will have 8 units(24) and a group height of 52 in. The 6-in. diameter unit gives a dry flash-over of about 50-k.v., and wet 30 k.v. Two units in series are used at Hayle, Cornwall, on a 10-k.v. line of Edmundson's Electricity Corporation.
Fig. 28 (25) shows a recent production of the interlink type, used on the 60-k.v. transmission line, Dessau-Bitterfeld. It has the advantages of increased spark distance, better rain protection, and the use of wire ribbon, which is easier to thread through the holes than the ordinary circular wire. The strain insulator (Fig. 29) of the Hewlett "fishtail" pattern is made in the same diameters. In all the above lines the number of strain units and suspension units is the same. Both suspension and strain insulators of the largest diameter are tested to 3 tons (compression), and the smaller to 1-1/2 tons. Details of the clamps, etc., for the 80-k.v. Victoria Falls line mentioned are shown in Figs. 30 and 31.
(b) Cemented Type.—Figs. 32, 35, and 37 show the various forms of this type. The design of the cemented type can be made so that the porcelain is partly in compression and partly in tension. When an insulator of this type fails under mechanical stress, it is by shear and tension combined. By altering the relation of the bolt and bolt-hole surfaces, the bolt may be made to pull out when the shearing strength of the cement is reached. The following experiments conducted with the insulator shown in Fig. 35 show the physical characteristics of the cemented type. Immersed in a freezing mixture at —7 C. and then plunged into hot water at 90° C. showed no effect on its strength electrically and mechanically; nor when changing the order of immersion. There was no difference when immersed for 5 minutes at 0° C. and then for the same duration of time at 80° C.—the operation repeated a hundred times. Even heating the metal cap with a Bunsen flame to 130° C. and then immersing in water at 0° C. produced no difference. When all the sheds were knocked off the mechanical strength was practically unaltered.(26)
Fig. 32 is the Ohio Brass Company pattern,(27) and is similar to that described in a previous Institution paper.(28) It is 11 in. in diameter, and has a dry flash-over of 85 to 90 k.v., and wet 50 to 56 k.v. Under oil it punctures at about 130 to 140 k.v. The dry test on the four units is given as 300 k.v., and wet about 200 k.v. The first 100-k.v. line in India (44 miles in length)—that of the Tata Hydro-Electric Power Supply Company, Bombay—will use six 10-in. diameter insulators of this pattern for suspension on the intermediate towers and six 10-in. diameter interlink (Fig. 32) insulators on the anchor towers; details of suspension are shown in Figs. 30, 31, and 33(27)The ultimate break-ing stress of these units is about 5 tons, the routine test being 3 tons. For the 100.k.v. line of the Central Mexico Light and Power Company(28), 6 units are used. Taylor gives three as that used on a 60-k.v. line. On the 110-k.v. line of the Hvdro-Electric Power Commission of Ontario, 8 units of similar design(29) and 10 for strain, are employed. The latter have a more massive cap (about 3/4 in. longer) and have been tested to 4-1/2 tons. The dry flash-over voltage of the suspension unit is given as 84 k.v., wet 50 k.v.; each unit adds 70 k.v., and 40 k.v. wet respectively.
Fig. 34 is the Duncan insulator (30) of the Locke Company, made up of two shells. It has the advantage of increased dry surfaces in unfavourable weather. The largest insulator has an outside diameter of 14 in. and an inner of 6 in. The shells are tested with 90 and 60 k.v. respectively, the unit being specified for a 25-k.v. line pressure. Four units are used in series on the 110-k.v. line of the Stanislaus Electric Power Company, San Francisco,(31) and four on the 100-k.v. line of the Great Western I'ower Company, California.
Fig. 35—the Hermsdorf form—is like the conventional pin insulator. It is made in three sizes for line pressures of 20, 25, and 30 k.v., but the last two sizes are the most suitable for high pressures. The intermediate size has a dry flash-over voltage of 88 k.v., wet 51 k.v. TWO units are sufficient for 50 k.v., and three for 70 k.v. The largest size has a dry flash-over of 100 k.v., and wet 61 k.v., and is used on a 70 k.v. transmission line in Guadalajara (Mexico), three units in series.
Fig. 36 is a later Hermsdorf type, with a sunken cap, which gives the unit greater mechanical strength and enables the units to be brought nearer together. A mechanical test shows that the first fracturing of the porcelain occurs at 5-1/2, tons: at 8 tons the bolt is torn out. Its dry flash-over voltage is slightly less than that of the previous type but its wet value is better with increasing number in a group. It is to be employed five in series on the first European 110-k.v. transmission line—Lauchhammer-Groditz. The total height of the group will be 3 ft. 9 in., whereas on the American lines the height of suspension groups is 4 to 6 ft. Fig. 37 is a metal-shed form of Fig. 35. Its dry flash-over is the same, but its wet flash-over is 68 k.v. In comparison with the all-porcelain insulator the wet flash-over voltage improves with the number in series. Fig. 38 is the newest metal-shed form of Fig. 36. By curving the metal shed better mechanical protection and provision against weather are obtained.
Electrical Considerations in Design.—The principles of design of pin insulators discussed on pages 245 and 246 are in the main applicable to the suspension insulator, for the latter may be considered as a modification of the former. If small discs of metal, in duplicate, be interposed at the cemented joints of the insulator shells, the flux distribution will not be sensibly altered, as the equipotential surfaces before introducing the metal are practically coincident with the cement layers. On separating the shells, with the discs attached on either side, and electrically connecting the discs, we get the series formation of suspension units. It is only necessary to make the shells of uniform size and to modify the electrode to produce the types shown in Figs. 26 to 38.
The separation of the shells has introduced large air-spaces, thus the total capacity and charging current decreased considerably. The distribution of flux density is more uniform than before separation, and therefore there is less liability to puncture (see page 252). For a given voltage the flux density in the shortest distance between the electrodes of the unit depends on the total thickness of porcelain for the n units of the series; the larger number of units which can he used in contrast with the number of sheds in the pin type diminishes the flux, the flux density, and the puncture risk. Each unit must itself have sufficient thickness, so that flash-over takes place before puncture. This is one reason for the employment of two shells in the Duncan form (Fig. 34). The difficulty of boring the interlink type makes the thickness of porcelain between the links variable and produces a greater percentage of punctures than in the cemented type. This greater percentage, together with the difficulty of boring, brings up the price of the interlink type.
As the separation of the units increases, the flux density at. the electrodes diminishes and then becomes constant. The distance at which it becomes constant depends on the shape of the unit. If we suppose that the flux, for a given voltage, varies inversely with the n units,(32) then to attain the same flux (and flux density) as for one unit a voltage n times is required. An equivalent statement is that, the capacity being reduced n times necessitates a voltage n times as great to produce the same displacement current. That is, as the line pressures are increased the addition of units gives still the same displacement current. Contrast this with the pin-type insulator, which increases in capacity for the higher line pressures. The relative size of the electrodes is a factor influencing the flux density at the neck of the unit. The metal cap in the cemented type tends to produce a better flux distribution than in the interlink type. (The sunken metal cap, Fig. 36, removes the high gradient from the top shed to the under surface of the pin shed—hence the wide airspace to counteract it.) The metal sheds of Figs. 37 and 38 distribute the flux better still, lessening the tendency for glow formation. The smaller thickness of porcelain and smaller size of the electrodes of the interlink type start the glow earlier than in the cemented type. Its capacity is also greater. Under rain the initial discharges occur sooner than with the other type. Narrow air-spaces between the shells (as in the Duncan unit) and between the sheds (in the petticoated or multi-shed unit), where the flux densities are high, tend to formation of glow.
In the early designs of suspension unit a disc type without petticoats was employed. To obtain a high flash-over large diameters and great separation of the units were tried, but it was readily seen that the full sparking distance could not be utilised. This is on account of the surface sparks(33) produced on the upper and lower surfaces (see Fig. 39). A small thickness of the shell and a large diameter tend to establish these sparks, and thus the unit flashes over, for the same sparking distance, at a less voltage than when caused by a brush discharge in air-path direct. It is obvious from our discussion on page 247 that if the unit flashes over by the surface spark it will be true also for the series; especiallv when we take into consideration that the spark distance (over surfaces) varies with the third to fourth power of the spark voltage. In the Duncan unit the two shells force the brush streamer to take an air-path, but the large diameter (14-1/2 in.) gives rise to the surface spark. The addition(34) of petticoats, as in the type of Fig. 32, raises the sparking voltage of the unit (35) for instance, with a 10-in. diameter disc the petticoats increase it 40 per cent., in addition to increasing the surface resistance; in wet weather especially its characteristics are better than for the plain disc. Under rain, brush discharges taking place between the petticoats (as in Fig. 15 for the pin insulator) cause flashing over.
For the same sparking distance the inclination of the shed of the unit has little effect for small diameters, but lowers the dry flash-over voltage for large. In the case of the latter, it disadvantageously increases the height of the series.
Distribution of Potential Gradient of the Series.—In passing from the lowest unit to the topmost, the flux (especially that part taking the long air-paths) reaching successive electrodes is not the same. Evidently for the middle units the flux density is less than for the outer units, and the divergence will be greater the greater the number in series. The divergence will also depend upon the distance apart of the units—namely, upon whether the ultimate flash-over takes place over the whole series as in Fig. 40, or across each individually as in Fig. 41. The flash-over voltage for one unit (Fig. 32) is 90 k.v., the addition of another increases it by 70 k.v. (which is explained by a diminution(36) of spark distance, but at the sixth unit the increment is only 30 k.v. If the distance between the units be increased so that flash-over takes place across each unit, the spark voltage will be higher and the diminution of the increments is not so marked. The metal shed(37) unit causes a less sunken cap (Fig. 38), as the units can be placed nearer together. The curve of flash-over voltage and number of units is practically linear, which is the case with the ordinary porcelain units only when wet. The difference between the dry and wet flash-over voltages with the latter amounts to 4o per cent., and this is even greater when the under surfaces are wet with driven snow. With protecting rings the difference is 20 per cent, and with these curved(38) metal shed units 12 to 15 per cent.
TESTING OF INSULATORS.
~~Measurement of Voltage.!!—The equipment of the testing laboratories of electrical porcelain works should be such as to be able to carry out two kinds of tests. The one consists of investigation work on the characteristics of new designs when subjected to the conditions in practice; and the other comprises guarantee tests on representative sampler selected by the purchaser, and puncture tests on duplicates. The laboratories should be equipped for mechanical and also chemical tests, the apparatus for the former to deal with compression, tensile bending strengths of the insulator, and for the latter to examine the raw materials and to lead to further improvement in the porcelain. Various plants have been described in the technical press.(39) In view of the increasing demand for high pressure many have installed transformers up to 300 k.v. The Victor Works of the Locke Company possess three transformers, each of 250 k.v. and 200-k.w. capacity ; the Hermsdorf includes three of 200 k.v. and a fourth of 200 k.v., and are extending to a fifth of 500 k.v. In the high-tension laboratory of the Manchester School of Technology a transformer of 120 k.v. and 30 k.w. is employed for research. The equipment for control and measurement of the pressure have been previously described in an Institution paper.(40)
There are three methods of measuring the pressure: (1) by transformer ratio, (2) by the Kelvin electrostatic voltmeter reading to 100 k.v. and (3) by a spark-gap. In the first method the pressure of the primary to obtained on a Kelvin electrostatic voltmeter (with mirror attachment) and a scale length of 3 ft., corresponding to 120 volts. The range of voltmeter is increased to 600 volts by a resistance of 100,000 ohms placed across the primary and tapped in five sections. A high resistance (20,000 ohms per kilovolt) placed on the secondary winding of transformer enables the same voltmeter to read up to 60 k.v. Oscillograph records and the spark-gap enable the maximum values of pressure to be determined.
With regard to the spark-gap, there is a tendency to measure voltage by the needle-point spark-gap standardised by the American Institution of Electric Engineers.
(1) The more recent "Normale" type for higher pressures is constructed on the American pattern.
(2) Under oil, a plate of this thickness punctures at about 105—110 k.v.
(3) Transactions of the American Institute of Electrical Engineers, vol. 29, part I, p. 590. 1910.
(4) Nicholson, "Protection of Insulators from Lightning Effects," Transactions of the American Institute of Electrical Engineers, vol. 29, part 1, p 573 (1910).
(5) J. Lustgarten," Flash-over Voltages," Electrician, vol. 62, p. 374, 1908.
(6) With increase of humidity the gaseous ion becomes the focus of neutral molecules water vapour which, on account of the "loading," requires a greater potential gradient to produce the necessary ionisation for tile brush discharge.
(7) Kemp and Stevens, Journal of the Institution of Electrical Engineers, vol. 45, p. 689, 1910.
(8) Weicker, "Dissertation," Konigliche Sachsische Technische Hochschule, Dresden, 1910.
(9) The sparks coming out black on the white porcelain surface in Fig. 15 is an interesting example of photographic reversal.
(10) Weicker, "Dissertation," Dresden, 1910.
(11) Nicholson, Transactions of the American Institute of Electrical Engineers, vol. 29, part 1, p. 573, 1910.
(12) Equipped by Messrs. Johnson and Phillips.
(13) A line of 135 k.v. is at present under construction (see page 259), and higher pressures will be only a question of time until corona on the line wires will be the next question to consider.
(14) Weicker, Helios, vol. 17, Nos. 28-30, 1911.
(15) Taylor, Tunnicliffe & Co.; Bullers, Ltd.; Hermsdorf Porcelain Company; and Berenguer.
(16) Hermsdorf Works.
(17) W. T. Taylor, "Modern Long-distance Transmission of Electrical Energy," Journal the Institution of Electrical Engineers, vol. 47, p. 174, 1911.
(18) The 44.-k.v. transmission line from Lockport to Western Avenue (Chicago) has added suspension insulators to secure additional carrying capacity for the increasing loads. The original lines in duplicate possess six pin insulators. Two 3-unit suspension type are hung from the bottom cross-arm, and a pin insulator placed at the peak of poles gives now three separate circuits. The last wire takes the position of the original ground wire. The arrangement should also substantiate the claims of those who advocate the use of a ground wire to prevent lightning disturbances in the next lightning season. The Nicholson arcing-ring device has been adapted to all the insulators on the south side of the pole line. (See Electrical World, vol. 57, p. 171, 1911)
(19) Buck, Transactions of American Institute of Electrical Engineers, vol. 26, p. 981, 1907.
(20) Matthews and Wilkinson, Journal of the Institution of Electrical Engineers, vol. 46, p. 562, 1911.
(21) Manufactured in England by Messrs. Butlers, Ltd.
(22) Electrical World, vol. 55, p. 202, 1910.
(23) Ibid., vol. 56, p. 98, 1910.
(24) Since the reading of the paper, the author notes that the line is operated at 140 k.v. and is equipped with ten 10-in. Ohio Brass Co. type insulators (Fig. 32). Ibid vol. 59, p. 795
(25) Made by Hermsdorf Works.
(26) Weicker, Elehtrotchnisch2r Zeitschrift, December 14 and 21, 1911.
(27) Made by Butlers. Ltd.
(28) W. T. Taylor, 'journal of the Institution of Electrical Engineers, vol. 47, p. 174, 1911.
(29) Made by Hermsdorf Works.
(30) Transactions of the American Institute of Electrical Engineers, VOL 29, part 1, p. 615.
(31) Mathews and Wilkinson, ~Journal of the Institution of Electrical Engineers~~, vol. 46, p. 562, 1911.
(32) This is not strictly true (see page 265), but when the top sheds are wet or are of metal it is, after the first unit, approximately true, the metal sheds acting as parallel equipotential planes between cross arm and line wire.
(33) Austin, Proceedings of the American Institute of Electrical Engineers, vo1. 30, p. 1320, 1911, wrongly attributes the surface sparking characteristic to surface resistance.
(34) See results with sooty insulator, Fig. 47 A and page 241.)
(35) Though the addition of the petticoats increases the total flux, for a given voltage, before discharges take place, it is afterwards smaller than for a plain disc.
(36) For one unit 90 k.v. is for the flash-over distance from cap to bolt; for the two units 160 k.v. from cap of first unit to bolt of second across the flanges. The spark distance from the cap of one unit to the cap of the next is slightly less. The bottom most unit in all cases, has the shortest spark distance.
(37) A metal sheet dividing an all-porcelain series in two halves would cause a more uniform distribution and increase the increments; incidentally it would raise the voltage and protect the lower units from rain.
(38) Weicker, International Congress of Applied Electricity, Turin, September, 1911.
(39) Electrical Review, vol. 62, p. 401, 1908. Messrs. Taylor, Tunnicliffe & Co. Weicker, Electrician, vol. 61, p. 51, 1908, Hermsdorf Porzellanfabrik. Brady, Electrician, vol. 61, pp. 322, 1908, Messrs. Bullers, Ltd.
(40) Journal of the Institution of Electrical Engineers, vol. 45, p. 685, 1910. Electrician, vol. 55, p. 809, 1905.
