Publication: The Electrical Engineer
New York, NY, United States
THE NIAGARA POWER TRANSMISSION LINE.¹
BY J. G. WHITE.
MY first personal knowledge as to the then proposed transmission line between Niagara Falls and Buffalo dates from the autumn of 1894 when the White-Crosby Company was asked to prepare and submit detailed plans, specifications and proposals for its construction. It was then found that the engineers of the Westinghouse and General Electric companies had both recommended the construction of circuits of three wires, adapted to the three-phase system, each wire having an area of about 330,000 circular mils. Beyond this no definite plans had then been determined on. The engineers of the Cataract Construction Company at the time considered it advisable to have the line built entirely of iron poles and accordingly detailed drawings and specifications were prepared and submitted showing alternative plans. The first included two entirely independent lines of iron poles, not less than 40 feet long, weighing something over two thousand pounds each. The second consisted of lines of poles of about the same height, weighting one thousand pounds each, set in pairs and tied together by deep trusses, which served both as braces against the terrific wind storms which sometimes sweep across Lake Erie and as a means of carrying part of the wires. A steel truss fulfilling admirably this double function was designed, which weighed about six hundred pounds, the company's right of way being 30 feet wide, and the lines of poles located 15 feet between centers. Several wooden trusses were designed for the same purpose, but, while better from an electrical standpoint, these were all so clumsy in comparison to the iron poles that preference was given to the steel truss. Various designs of poles were considered, including several built up from "rolled shapes," final preference being given to plain tubular poles on account of their ability to withstand equally strains from all directions, their appearance and the ease with which they can be kept painted.
Numerous designs for cross arms were also considered, including those made from "rolled shapes," composite arms (part wood and part steel), and those made entirely of cast iron or of wood. The latter two were preferred on account of the ease with which they could be designed to accommodate either wood or iron pins.
In determining the details for these alternative plans it was assumed that poles would be set in concrete and spaced one hundred feet apart; that each wire would weigh one pound per lineal foot; that the three wires composing each circuit were to be placed at the corners of an equilateral triangle having sides at least three feet long; and that the lateral strength of the line was to be not less than three times the forces produced by wind acting with a pressure at right angles with the direction of the lines, equal to thirty pounds per square foot on the projected surface of the poles, arms, insulators and wires, when the latter were covered with a coating of ice one-half inch thick.
As a matter of fact the heat generated by the current passing through the wires, together with the static effect tending to repel all particles of moisture coming in contact with the wires, this effect being quite noticeable on a ten thousand volt line, would combine to prevent the formation of any such coat of ice unless at a time when current was off the line. Any error thus introduced was on the safe side and consequently not objectionable.
Full proposals with detailed plans and specifications for the construction above outlined were submitted October 11, 1894, and the amended proposition for carrying out the construction on the same general lines was submitted March 13, 1895. Nothing further developed in the matter until June, 1896, when new proposals were asked for the construction of the line on the assumption that white cedar instead of iron poles would be used throughout. Such proposal was submitted June 18 and accepted some days later. In working out the details of the line as built, the same general assumed data above given was used.
When the route for the line was finally determined on its length was found to exceed 27 miles instead of being 25 miles, as previously assumed, and consequently the area of the wire was increased from 330,000 to 350,000 circular mils. The wire actually erected is composed of 19 strands, having a combined area of full 350,000 circular mils, and weighs nearly 6,000 pounds per mile.
In designing a transmission line, the three most important factors probably are: 1. Its ability to carry its full load continuously and without interruption. 2. Cost. 3. Efficiency.
The first cost of power used to develop current for a transmission line is usually low, wherefore the efficiency of the line is not of primary, although of great, importance. Those of us familiar with the development of the street railway motors have seen in practice an illustration of the fact that efficiency is not always a controlling factor. There has probably never been in general use another street railway motor so efficient as the Sprague No. 6. With units of an ordinary size a short circuit on a line carrying ten thousand volts, even if through a defective insulator, a wooden cross arm and a wooden pole would make itself manifest at the power station by the opening of a circuit breaker, the blowing of a fuse, or some similar method. With the huge generators which furnish power for the line the effects are different. During one night last fall, while an attempt was being made to operate the line on temporary insulators, the best obtainable at the time, the ends of no less than five of the large cross arms used on this line were burned entirely off; and this too without any manifestation having been made at the station that anything unusual had occurred. This naturally raises the questions whether it is possible to procure insulators which can be depended on to maintain the insulation on a circuit carrying 10,000 volts or over, and whether it is not good practice to have fewer poles, and consequently less weak spots in the form of insulators.
There are two sides to this question, and both deserve serious consideration in designing any transmission line. Let us assume that poles are set 100 feet apart and allow a sag in wires between supports of twenty inches, or one-sixtieth of the length of the span. We find that the area of the wire in use on this line is .267 of one square inch, and that its tensile strength, even assuming a high value for soft copper is about 10,000 pounds. Allowing the same deflection, one-sixtieth of the length of the span, this determines the maximum safe distance between poles as 178 feet, allowing a factor of safety of four, and shows that the cables might be expected to break if the span were lengthened to 712 feet, no allowing for wind pressure or extra load due to ice.
Assuming tensile strength of 8,000 pounds per square inch for yellow pine, we find that the larger cross arms used on the line, which are twelve feet long and nearly five inches wide by six inches high, would support a load of 270 pounds on each end without bracing, and of 360 pounds on each end with the steel angle braces used, and this too with a factor of safety of ten, an unnecessary margin when we consider that the arms are all specially selected heart, long leaf Georgia pine. Similarly these cross arms would have the same factor of safety carrying three power cables on each side with braces. Besides giving this added strength, the braces used on this line prevent such vibration and oscillation as usually takes place with the ordinary strap iron braces, such oscillation being the cause of many of the petty troubles on ordinary lines. These braces were each made from a single piece of steel angle 2 x 2 x ¼ inches bent to shape and forming when finished a truss 18 inches deep and five feet in extreme length, their weight being a little over 20 pounds each. Assuming again that poles are set 100 feet apart, we find that 12 wires with cross arms, insulators, etc., would present an area to the wind aggregating about 67 square feet, and that consequently pressure was 30 pounds per square foot, of about 2010 pounds. A sound 50-foot cedar pole, 8 inches diameter at top and 18 ground, would be capable of withstanding before breaking a side pressure nearly its top of only about 4,900 pounds and or only about 3,400 pounds when a layer 2 inches thick had decayed around its circumference.
With spans of 100 feet the pole would, therefore, have a factor of safety of only about 2-1/2 when new, while the wires would have a factor of safety of about 7, and the cross arms of about 26. The advantages of having cross arms amply strong are so evident, and the possible reduction in cost such an insignificant part of any ordinary line carrying much copper, that it would seem foolish to reduce the strength size of these in order to bring their strength down to correspond with other parts of the line.
It is evident from the above that the weakest point of this line mechanically is the pole, in spite of the fact that as advised by one of the prominent members of your Executive Committee final decision was made in favor of spacing the poles 75 feet apart on straight, and proportionately closer on curved parts of the route. The only reasonably safe and practicable method of decreasing the number of weak spots furnished by the insulators would be use poles larger than eight-inch tops, or to brace the poles to withstand this wind strain. Fifty-foot poles having tops greater than eight inches are now hard to find, and although 35-foot poles were used on a part of the company's private right of way, nevertheless many 50-foot and even a few 65-foot poles were required to avoid obstructions and for crossing railroads and highways. The only feasible plan, therefore, would seem to be to brace the poles laterally, which can readily be done if set in pairs, but which would be very difficult to accomplish in a satisfactory manner with a single line.
This naturally brings us back to the question, Are insulators such insurmountably weak links in such a chain, and is it not possible to get insulators which can be depended on even when supporting wires under a pressure of 10,000 volts? This can be most satisfactorily answered by again narrating the experience with this line. During the past eight months insulators have been sent to Niagara Falls by four of the works which are among the first six in this country in the production of porcelain for electrical use. Of a sample lot of ten received a few days ago for test from one of these factories, one had broken in transit, eight broke under the strain of electrical pressure varying from 16,000 to 36,000 volts, and the last broke under 44,000 volts strain. This latter would seem to be a minimum safe test limit for any insulator expected to sustain a regular strain of 10,000 volts and is a test which any mechanically good, well vitrified insulator of ordinary design will pass. As several previous lots from the same factory showed even poorer results, the manager of that company states that he hopes within a few weeks to be able to furnish insulators which will stand a 40,000 volt test. It is only fair to state that these insulators were of a smaller type than, and different design from those in use in the Niagara line. Several lots of somewhat similar insulators from another factory gave about the same results.
Several thousand insulators of a diameter almost equal to the round type now on the line, but of a design somewhat different in details, were furnished by a third porcelain works. These were all supposed to have been tested and to have successfully withstood a pressure of 44,000 volts at the factory before shipment. When, however, these were tested at Niagara Falls by Mr. Lincoln, the electrical superintendent of the Cataract Construction Company, it was found that a large majority of them broke down under a 40,000 volt test, illustrating that a dry test, such as had previously been made, is use-less for practical purposes.
The method of test used at Niagara Fails was as follows: The insulators were set inverted in a shallow iron pan is lots of about 20, the bottom of the pan being covered with an inch or two of water containing a little salt. A little of the same brine was poured into the pin hole of each insulator, and into this was thrust a small piece of metal such as an ordinary iron spike or the small round zinc rod from an ordinary sal ammoniac battery, this being connected to one aide of the testing circuit, the other being connected to the pan containing the insulators. After the metal rod had been placed in the brine in the pin hole of an insulator, the primary circuit of the testing transformers, specially built for the purpose, was closed; and if the insulator was weak, this was quickly manifested by a series of sparks through the punctured porcelain. Experiments made with pure water and with brine showed that there was no difference in the results; but that any weakness was manifested a little more quickly with brine, besides which the salt imparted the characteristic bright sodium color to sparks otherwise almost colorless and difficult to detect.
As it was important that the line should be ready to deliver current by a specified date, the test was reduced on these insulators to 20,000 volts, and all which withstood this pressure were passed for temporary use. These insulators were later replaced by some of those now on the line, all of which successfully passed a 40,000 volt test made as above described. A lot of the temporary insulators illustrated the old saying that every rule has its exception, for when after removal from the line they were tested under 40,000 volt pressure, a solitary insulator from a total of 1,150 was able to pass muster.
Of the insulators shipped by the fourth factory, and of the two types now on the line, about 25 to 40 per cent. were usually found to be detective, breaking down under 40,000 volts, this percentage decreasing in the last shipments received. It is, however, worthy of special note that since the last of the temporary insulators were moved from the line, there has not been one minute's suspension of current supply due to defective insulators. During this time, some three or four months, three insulators have been replaced, none of these being of the oval type. Two of these three had been broken while being put in place, and the third was broken by a stone or other missile. In all three cases only the outer petticoat was broken, and the insulator continued to do satisfactory service until such time as it could be replaced without affecting the operation of the line.
Apparently, therefore, it has been demonstrated that it is possible to secure insulators which are reliable. That there have been no greater troubles in the past from defective insulators is probably due to the fact that most of the large transmission plants in operation under high voltage up to within a few months have been in the far West, where the climate prevails very different from that natural to this immediate region.
In the above and other experiments with insulators some interesting facts have been developed and are worthy of note. The insulating strength of porcelain depends almost entirely on the thoroughness of its vitrification and very little on its thickness, a thin china tea cup having successfully withstood a pressure of 60,000 volts, while a coarse piece of porcelain two inches thick was readily pierced by 20,000 volts. It is therefore, practically unnecessary to met electrically any insulator which, when broken, will not pass a good absorption test, using red ink or other fluid.
It is quite, if not entirely, impossible to puncture a glass insulator, even an ordinary pony telegraph insulator, not-withstanding any pressure which can be applied, the last being determined by the pressure which will send an arc around the insulator. The objection to using glass insulators in the past has been due to the difficulty in getting a well annealed and mechanically strong insulator, of such massive design as is needed for this work and to the hygroscopic property of glass, which is not shared by porcelain. The first can unquestionably be overcome by care in manufacture. The importance of the second has probably been exaggerated in most calculations made in the past, due to an inadequate appreciation of the static effects of 10,000 volts in warding off snow flakes and drops of rain, and to a less extent of the rapidity with which water falling on such insulators is evaporated by the heat of the current leaking over the surface.
It is consequently reasonable to expect that the use of glass insulators for high voltage lines will greatly increase with improved manufacture. Meantime any lines erected should have the best obtainable porcelain, and every insulator should be subjected to test.
Before closing it is natural to ask, "Is the line as built a genuine success? Can it be depended on, and is it effective?" In answering let me give briefly some of the facts. The line now in operation is over 25 miles long, of which the last 4,000 feet is underground, the current being carried in lead covered cables with rubber insulation, these having been drawn into terra cotta duct conduit built specially as part of this line. These cables successfully withstood a test of 40,000 volts, are guaranteed for five years under working pressure up to 25,000 volts, and were punctured during test only by a pressure, estimated by Mr. Lincoln at about 80,000 volts. They have given no trouble since current was first turned on the line, November 15 last, except at two joints imperfectly made. Except for the short time needed to repair one of these joints there has not been a single abut-down chargeable to the transmission line itself, since the last temporary insulators were removed some three or four months ago.
A number of interruptions to service have occurred during that time, due to derricks used on the work now being done on the Erie Canal hitting wires, undermining of poles and conduit by this work and to allow new lightning arresters to be put in circuit at transformer houses. Except for these extraneous and unusual causes the service has been perfect with the one slight exception above noted. One short interruption early last winter was due to the dead limb of a tree blowing across the wires, illustrating the fact that all trees should be cut down for some considerable distance on both sides of any high voltage line. The line shows an insulation resistance of some 250,000 to 300,000 ohms on wet, and about 1,000.000 ohms on dry days, this being between any one of the three wires and the ground, the insulation therefore varying from six to twenty-five million ohms per mile of wire.
The actual working efficiency as shown by the wattmeters in the low tension alternating circuits at Niagara and the direct current 500-volt circuit at Buffalo was 79.6 per cent., this being for a considerable period and a fluctuating load. This efficiency included loss in step-up transformers,