LARGE HIGH VOLTAGE TRANSFORMERS
Reproduced from Machinery Lloyd and Electrical Engineering (Vol.37, No.1 2nd January) with the kind permission of the publishers - The Certificated Engineer July 1965.
Transmission voltages of 700 kV and above are becoming increasingly accepted as a necessity, as the demand for power increases. To mark its successful testing of the first of three 735 kV transformers which will be installed in the Quebec hydroelectric scheme in Canada, the English Electric Co. Ltd. held a symposium at its Stafford works towards the end of 1964. The object of the symposium was to examine the design and manufacture of high voltage transformers as the company believes the 735 kV transformer to be the highest voltage rating transformer so far manufactured and tested in the world.
The English Electric Company's experience of high voltage transformer design extends over many years and, apart from overseas activities, it traces the development of high voltage transmission in Great Britain. At the commissioning of the first 132 kV grid line in Great Britain in 1930, the transformers at one end were made in Stafford. The same applied in 1952 when the 275 kV grid was commissioned and again in 1963 when the first experimental 400 kV line was energised. Significantly, transmission at 400 kV became a reality so soon after the 275 kV grid was commissioned, and it is unlikely that 700 kV will be more than a passing stage. Engineers must be considering the problems associated with working at 1 000 kV or above. Unhappily, the designing of larger and higher voltage transformers is not just a case of scaling up proven designs. New electrical and mechanical problems arise and even such matters as transport impose strict limitations which have to be overcome.
The problem of transport
As units become larger, complicated problems of design, manufacture and economics arise. For a typical generator transformer, the capital cost can be shown to be related to the losses and in recent years the heaviest design available has tended to be the most economic on this basis of calculation, but this results in a tendency for the most economic design to exceed the weight limitations. Various solutions to this problem have been advanced.
One solution receiving close attention at present is the use of single-phase transformers. Although this idea is often rejected on the grounds of price by comparing the cost of three single-phase units with one three-phase unit, the calculation is not quite so simple. In a large power station, there should always be a spare transformer and the choice may well be between say five three-phase units and thirteen single-phase units. From the manufacturer's point of view, the building of thirteen identical units of manageable size might be more attractive than the smaller number of large units.
The second solution involves raising the weight limitation by using a ship with drive-on and drive-off facilities for vehicles carrying transformers weighing up to 300 tons and strengthening bridges and roads between manufacturers' works and the dockside. One school of thought believes that the cost of this expenditure on roads may be justifiable in terms of the economics of using larger units.
A further solution is building on site. On the supposition that the desired size of the transformer is already above 200 tons, will soon pa s 300 tons, and will continue to rise, the possibility of complete site construction needs to be examined. In generating stations where there are adequate covered areas, power supplies, crane facilities, etc., it would be a relatively simple task to plan a complete site-building operation. The problem is quite different in the case of transmission transformers on remote sites.
If site erection becomes necessary manufacturers can develop portable workshops, test equipment and other gear such as the demountable drying oven which the English Electric Company has already. At present, the economics of site-built transformers require further consideration but there is already evidence of their technical acceptability.
The fourth solution is a compromise involving the building of the transformers at the manufacturers' works, dividing it into transportable units and reassembling them at the site. Various methods have been proposed but the one favoured by the English Electric Company is the vertical separation of the core. It is embodying this method in a 180 MV A transformer which it is building. In this, the total shipping weight of the three-part design is 177 tons compared with 155 tons for the conventional three-phase unit. This technique, which avoids the use of heavy lifting tackle, can provide the low losses required at a reasonable cost.
In this development, the objectives have been to use techniques and materials of proven effectiveness, to avoid a reassembly procedure needing large cranes, and to avoid exposure of major insulation to atmospheric conditions. The method used is to make the transformer tank and core in three parts, split vertically, each part of the core standing on rails in its part of the tank and the three tank sections being on rollers.
The yokes are split, the two parts of each yoke being connected by multi-leaf overlapped inserts which can be positioned on site and clamped tight by a man inside the tank for the lower yokes or working through an access hole for the upper yokes.
On assembly during manufacture barriers are fitted between sections to isolate the windings from the atmosphere, and when the tank is closed up the whole unit is placed in a portable oven and processed, by applying a vacuum to the tank and circulating hot air around it.
After testing the oil is drained and replaced by dry air and the transformer is split into its three components for transport. On arrival at the site, one end section is positioned and the others have drawn on to it in turn, the yoke connections are made, the flanges bolted up and the barriers removed. When the joints have been made the tank may be treated as one piece-in fact the complete transformer may be jacked up by the ends only.
Insulation and cooling
Although the company has investigated new materials which have been claimed by their manufacturers to have superior insulation properties, it has reaffirmed its belief that the conventional basic insulation system of paper, oil and fuller board is the best.
The major insulation between windings on the 735 kV transformer is of the English Electric 'Multiduct' construction which was introduced in 1962. Doubt about the use of solid insulation had by then become serious for voltages of this order as it was known that thermal instability was certain unless some method was used to remove the heat produced by dielectric losses. The 'Multiduct' system used consists of a large number of small oil ducts, narrow enough to have high dielectric strength, yet able to allow adequate heat flow. The resultant capacitance between the windings is very much lower than with solid insulation whilst the 'open' structure allows faster removal of moisture during dry-out and faster impregnation after oiling up.
Away from the high-stress regions some of the Multiduct is cut away to improve the cooling oil flow in the windings. At Stafford, there has been a careful investigation of the location of hot spots and the effect of oil flow. For this purpose, a test rig has been devised in which transformer windings can be simulated by a simple-to-produce folded stack in which the conductors, etc., are full size. In this way, scale-effects have been eliminated. It is considered feasible to use copper rating as high a 40 W/lb by careful design to avoid hot spots.
At the ends of the windings, the Multiduct fuller-board layers are flanged and interleaved with angle rings to provide both puncture strength and creep distance to the yokes and between windings. These arrangements and the designs of the components which form them, spring from the research work carried out in preparation for very high voltage work.
For example, the use of flat sheets of pressboard for interphase insulation even was able to withstand the test voltage, should the oil itself break down, is now known to be inadequate. Appreciably higher interphase stresses than formerly can be allowed by using correctly shaped and spaced barriers which follow the shapes of the equipotential surfaces. In this work, the electrolytic tank is used extensively.
In the insulation of coil ends and leads to the correct use of barriers is vital. The electrode itself is covered either by wrapping with paper tape or pressboard or coating with a synthetic resin where the shape is awkward. This has the primary effect of suppressing any possible discharges arising from minute irregularities on the electrode surface which become increasingly important at higher voltages. On electrodes of small radius, a fair build-up of paper will enormously reduce the maximum stress of the innermost layers of the oil. The use of an epoxide resin coating is well illustrated by the stress shield on the bottom of the high voltage bushing of the 735 kV transformer. By shaping it to give minimum surface stress a complex member resulted which was difficult to insulate with paper tapings. After a series of breakdown tests for comparison purposes, a resin coating was adopted and applied to the casting by the fluidised bed process.
Another problem which has been investigated in a specially prepared rig is the inter-influence of three or more electrodes and its effect on the possibility of breakdown. Any pair considered alone may be quite safe but the compounding of stresses by a third or subsequent electrode may raise the stress from one of them too well beyond the breakdown value. In the same rig, work has been done on creep and the possibility of failure by tracking across the surface of a solid. On many tests with modern materials, flash-over has taken place in the oil surrounding, and not across the surface under test. In such cases, the surface is very smooth and free from imperfections. English Electric considers that a creep surface must be treated as a weak link and broken as much as possible and with barriers inserted wherever the voltage level or stress demands it.
Windings
In the 735 kV, transformer interleaved windings have been used on both the 735 kV and 230 kV windings. This system, which English Electric introduced in 1947, provides a series capacitance many times greater than that of an ordinary disc winding. Interleaved windings usually have a distribution cons tan between 1 and 5 compared with 10 to 20 for a normal disc winding.
For this particular transformer, the method is chosen results in only half the test volts appearing on one leg, therefore only half the main insulation was necessary. This, of course, gives different values of ground capacitance for the two legs and to give as uniform an impulse distribution as possible down the full 735 kV winding, the product obtained by multiplying the ground capacitance by the series capacitance was kept constant for the two legs. This was achieved by grading the capacitance of the 735 kV winding between the two legs by winding one with two wires in parallel and the other with one wire.
Under test at 5 per cent above normal volts, the sound level measured was 82dB at a flux density corresponding to 18 000 lines/cm3 at 50 cycles. The analysis of this problem of the noise developed in the cores of large transformers and the investigation of ways of reducing it at source or of muffling it is receiving much attention in the English Electric research programme. Some results are already being put into practical use and a roller hearth annealing plant has been installed in the works. In this, the grain orientated core plates can be annealed after cutting and punching. The residence in this plant is so low that a controlled atmosphere is unnecessary.
Another method being studied is the reduction of the piezoelectric effect by tensioning the core plates perhaps by applying a coating to the plates to hold them always under tension. The clamping pressure also is believed to have an optimum value, both above and below which the noise level increases.
Specifications and tests
During the symposium, there was evidence of a growing doubt as to the wisdom of designing only to meet a customer's specification or to pass the more common test conditions. Speakers argued that both of these requirements must be interpreted in the light of service conditions if the best design is to result. Three dangers were pointed out-the designer might feel restricted by some rigid test condition; the competing manufacturers might aim to exceed certain factors where the improvement is irrelevant in service, and lastly, the imposition of unsuitable tests may weaken the transformer.
Overpotential tests, in particular, came in for criticism. It is considered that as at present conducted, they are not effective either as a real safety factor test or in checking the capacity of the transformer to withstand service voltages. English Electric suggests that a longer duration test at say 115 per cent of system highest voltage would prove more effective, especially if run in conjunction with a discharge detection test. Since switching surges are liable to have a peak value over that obtained on the overpotential test a new switching surge test was suggested.