HIGH VOLTAGE DIRECT CURRENT POWER TRANSMISSION

By M. van der Spuy (Member) - The Certificated Engineer Nov 1967

ABSTRACT

With the considerable increase in bulk transmission of electric power, especially at extra high voltages, it has become necessary to examine critically the various transmission methods available. After a brief technical description of high voltage direct current transmission systems, the relative merits of direct current and alternating current systems are discussed and world research programs are reviewed.

1. INTRODUCTION

Direct current transmission of energy is not new to the electricity distribution industry; the industry started with this method in 1882. Until 1886, all central station systems generated direct current. The electric companies supplied their customers with low voltage direct current but could not deliver the power for more than a few thousand feet from the station. The arc-light companies, with a higher range of voltages, could reach out several miles but could supply only arc-lights for outdoor illumination because the higher voltages involved were not considered safe in the home and shop. In some cases, the arc-lamp was hung over the sidewalks in front of shops and commercial houses.

When William Stanley demonstrated the practicability of alternating current distribution in 1886, a healthy technical controversy as to the most efficient method of developing electric power systems arose. There were the proponents of direct current, such as Thomas Edison, and those who favoured alternating current, but who could not agree as to the most suitable frequency, which varied from 15 to 133 cycles per second.

The first transmission was accomplished at the generator voltage. With the development of the transformer, alternating current was freed from this voltage limitation. It was possible to advance as high as 30 000 volts on a 100-mile line in Germany and a 75-mile line in the U.S.A before 1900. As a means of providing high voltage for direct current, the Thury system was developed in France, which consisted of cascading many direct current machines in series at the sending and receiving ends. For a 112 mile line from Montiers to Lyon which was constructed in 1905, several generators were placed in series to provide 57000 volts direct current. This line continued in service until 1937.

During the latter half of the nineteen-thirties, the General Electric Co. operated an experimental d.c. line in Western Massachusetts at 27000 volts, utilizing mercury vapour converter tubes.

Following World War II, practically all d.c. systems had been supplanted by a.c. systems, although engineers recognized the inherent simplicity and economy of D.C., especially where long-distance submarine or underground cables were involved. Alternating current is more versatile because it can readily be stepped up to a high voltage with a simple, efficient transformer. Direct current cannot be handled by a transformer, and high voltage devices for converting from AC. to D.C. and from D.C. to AC. were not available in the power ratings now necessary for transmission. Present trends and future needs indicate, however, that much of the coming expansion of transmission lines will be in the form of circuits designed to carry 500000 kW (500 MW) or more.

2. NATURAL LOAD OF A.C. TRANSMISSION SYSTEMS

Power transmission over long distances using A.C. is complicated by the inductance and capacitance of the line. The lagging inductive volt-amps of a line = 12WL and the leading capacitive volt-amps of a line = V2WC. The characteristic impedance of a line, Z0, is L/C. The corresponding load for a three­phase line is thus V2/Z0 watts/phase or kV2/Z0 MW for a three-phase line - which is termed the natural load. Any long power line must be designed with its rated load approximately equal to its natural load.

Typical values for lines of various voltages are shown in Table I, from which it is clear that high power lines are inevitably operated at high voltages and high currents.

Even higher voltages than are shown in the table are to be used on lines under construction, e.g. in Canada. The Hydro-Electric Power Commission of Quebec is constructing three 700 kV lines from the Maniconagan River project to Quebec City and Montreal, a distance of 430 miles. The carrying capacity of each line will be 2 million kV (2000 MW).

To illustrate the size of transformer units that have been developed during recent years, Figure 1 shows one of the world's largest banks of transformers, comprising 3 single-phase transformers with a total output of 1 000 000 kVA, which was put into operation in 1959.

3. RECENT D.C. TRANSMISSION SYSTEMS

Within the last decade, successful development in Sweden was a mercury arc "valve" capable of operating at voltages in the order of 100 kV or more and of handling currents in the order of a thousand amperes. This has renewed the interest in D.C. as a supplement to A.C. systems. The first full-scale commercial D.C. transmission installation was completed between the Swedish-main-land and the Island of Gotland (a distance of 60 miles) in 1954. This consisted of a single conductor 100 000 V submarine cable, the ocean being used as an earth return. It transmitted 20 000 kW of power. The excellent record of performance of this link demonstrated the feasibility of high voltage rectification and inversion.

During 1961, a second underwater D.C. project was placed in service between England and France, a voltage of 200 000 V being utilized. A power of 160 000 kW was transferred over a distance of 40 miles. Experimental overhead D.C. lines were built in Russia with the object of transmitting power from hydro-electric stations over long distances. Today, many more projects are under construction and are being commissioned, while other projects are in the planning stage.

Table II indicates that over 2 million kilowatts of D.C. transmission facilities will be in use by the end of 1967.

Before discussing the relative merits of the two systems, a brief description of the technical characteristics of a D.C. system is given.

4. THE D.C. SYSTEM - TECHNICAL CHARACTERISTICS

A typical D.C. system is shown in figures 2 and 3.4 The stations are situated on the Mainland of Sweden and the J land of Gotland 60 miles away. The incoming A.C. supply, 130 kV three-phase in this case, feeds to the transformers via circuit breakers, and then to the grid controlled mercury-arc valves.

A.C. circuit breaks are required for switching the circuit in normal or emergency conditions; there is as yet no device commercially available for switching high power D.C. circuits. However, it is one of the advantages of the D.C. arrangement that, by employing a bypass valve, the transmission may be momentarily suppressed when backfires or line faults occur, and supply may be restored without switch operation and with as little noticeable effect as would be experienced with high-speed circuit-breaker reclosing in an A.C. connection.

The transformers are on-load tap change units and are necessary for voltage adjustment between the valves and the adjacent system and for increasing the number of phases applied to the valves in order to produce reasonably smooth direct current.

Mercury arc grid-controlled valves are at present the only practicable device for current conversion at high voltages. The valves at the two ends are identical in type. They may be used as required for rectification or inversion by adjusting the timing of the grid-control signals.

In addition to the above items, D.C. reactors are required for smoothing, and filters are necessary to reduce the possibility of harmonic currents resulting from valve operation causing interference or resonance with the adjacent power-supply systems. Where inversion takes place, capacitors, which may be synchronous or static depending on the associated system characteristics, are necessary for supplying the reactive power for the process of inversion as well as the reactive component of the A.C. load supplied. Current commutation in the valves may also give rise to radio interference, and, in the cross-channel project, it has been thought essential to accommodate the valves and associated equipment in a building completely metal-screened. The terminal equipment is, therefore, complex in type and arrangement and care is necessary for building the equipment and in co­ordinating its operation. It speaks well for the designers and development engineers to find that no major faults or problems have been found in the systems at present in operation. Power is transferred by a single core cable with the sea used as the return. The earth electrodes in the sea are located 10 kilometers from the converter stations.

The mercury arc valve for high D.C. voltage is similar in principle to that of an ordinary mercury arc-rectifier for medium voltage. Figure 4 gives a diagrammatic section through a typical valve.

The valve consists of an evacuated tank, at the bottom of which is the mercury pool cathode, with its associated ignition and excitation electrodes. The electrodes maintain a cathode spot on the mercury surface which is a source of electron emission and thus the source of the mercury arc between anode and cathode. Because it is the only point where current can enter a metallic surface, the arc is conducting only for current passing from the anode to the cathode. This is the simple basis of the unidirectional valve action of the ionic valve.

A control grid in the arc path from anode to cathode serves when given a negative potential, to prevent the formation of an arc even when the anode is positive relative to the cathode. By reversing the potential of the grid to a positive value, the phase of each cycle at which conduction through the valve begins can be controlled at will.

Two such anode structures working in parallel were used on the Gotland valves. The grading electrodes, interposed between the anode head and the cathode vessel, are a characteristic of the high voltage valve. Their connectors are led through the wall of the porcelain tube-and outside this they are connected to a resistive capacitive potential divider. The grading electrodes serve to help the valve to withstand the high inverse voltage which is impressed on it during a part of the A.C. cycle and also serves to withstand the high inverse voltage which is impressed on it during another part of the cycle when the control grid is acting to prevent the valve from picking up the current.

The valves shown in Figures 5 and 6 are as used for the English channel system. They are rated at 100 000 V 800 A and as can be seen, each valve has four anodes connected in parallel.

5. CONVERTER STATIONS

The terminal stations of a D.C. transmission line differ from normal transformer stations, not only in that they have mercury-arc valves and certain other special equipment, but also in that there is a difference in principle in the basic arrangement. In an A.C. transformer station and a power station, it is general practice to connect in parallel the number of units required to cover the total power of the station. On the other hand, in a converter station, the converter units are connected in series on the D.C. side as shown in Figure 7.

The main reason for this dissimilarity is that with A.C. it is simplest to connect up or disconnect transformers by means of circuit breakers; whereas with D.C. the switching problem is far more difficult and series connection provides a means of getting around the difficulties. It is, of course, possible to block the passage of current through a valve group by means of control grids, but in the case of certain internal faults of the valve, this method is not reliable. On the other hand, it is always possible to take a converter unit connected in series with other converter units out of service merely by means of grid control. This can be done without disturbing the operation of the other units. For this purpose, each valve group is equipped with a "bypass" valve which can short circuit the D.C. terminals of the blocked converter, thus leading the current past the converter. For the English Channel and the New Zealand systems the valves are connected in groups of seven, each group forming a 3-phase bridge arrangement with a bypass valve. The two poles of the D.C. system are so arranged that they will form two almost entirely separate transmission systems and any fault in either conversion equipment of the transmission circuit, will affect only one pole.

The neutral point between the two poles is earthed. During normal operation there will be no earth current, but in the event of a fault on a section, half of the transmission capacity would be available by using the earth return.

Despite the complexities of the D.C. system, there are, however, three regions of use where D.C. transmission has a definite technical and economic advantage over A.C. transmission. These will now be considered.

6. CABLES: UNDERGROUND AND UNDERWATER

An insulated cable has the characteristic of storing electrical energy. The current which flows into the cable as the voltage rises is called the "charging" current. The magnitude of the charge is nearly proportional to the length of the cable. With alternating current circuits, this charging current flows into and out of the cable as the voltage reverses, 50 times per sec. With extra-high voltage cables, these capacity effects which cause dielectric losses reach the stage where the full current carrying capacity of the cable (in heating effects) is required merely to charge the insulation. This imposes a definite limitation on the permissible voltage and length of the underground and underwater circuits when an alternating current is employed. In contrast, a D.C. circuit maintains essentially constant voltage, so that after the initial charging current, no further cyclic charging and discharging takes place and no dielectric heating takes place. There are also no sheath or armour currents and consequently no sheath losses. In the case of A.C. single-core cables, such currents are practically equal to the powerful currents in the conductors.

In addition, the insulation under D.C. conditions can be subjected to much higher voltages than when used for A.C. Each core of a cable suitable for say 132 kV AC. can operate satisfactorily at 200 kV D.C. to earth so that for equal currents, the ratio of power that can be transmitted by the two methods is approximately 1:2.6.

In the D.C. scheme for the English cross-channel connection, which was estimated to cost in all about 9 million rands, there was a saving of some 4 million rand on the cables as compared with those required for the A.C. scheme. With the D.C. scheme, two cables were adequate for transmitting the total power of 160 000 kW, whereas four single-core cables (one of which was a spare one) would have been necessary for the A.C. circuit.

In the New Zealand plan, where the power to be transmitted is 600000 kW at -+- 250 kV D.C., the cables to be installed are three in number, one being a spare and the other two operating at opposite polarities. For the comparable A.C. scheme, the voltage limit was thought to be 150 kV and therefore three 200 000 kW cable circuits would have been required, involving nine single core cables for service operation and two spare cables for emergency use.

In such a case the saving in cost of the cable installation completely off-set the cost of the conversion and auxiliary equipment at the terminal stations.

The reduction in the number of cables also requires less space on the seafloor, since the separation of a few hundred yards is necessary between the individual cables in order to facilitate identification and recovery if a repair should become necessary.

For the English cross-channel connection it would have been possible to transmit 160 MW on one D.C. cable, using the sea as a return path as was done in Sweden and Italy. Two cables of opposite polarity have, in fact, been laid side by side, and the supply authorities have agreed to make no use of sea-return for power interchange. The intention of this arrangement is to ensure that there is no interference with the compasses of ships passing over the cables. The magnetic compass deviation of about 4 degrees that would have been caused is small, but having in mind the concentration of shipping in this region, it was felt desirable to remove the interference completely.

Some apprehension has been voiced about the possible corrosion of underground facilities due to earth currents resulting from the direct current transmission. This problem was studied extensively in Sweden and Italy. Two cables of opposite polarity benefits to be gained from single pole systems such as in the Gotland and Sardinia installations. It was proved that the electrode current is dissipated into the interior of the earth and that the current at the surface are confined to the immediate vicinity of the electrodes without the one electrode influencing the current distribution of the other. Ground return systems were thus found to be feasible provided that the station grounds are adequately designed and sufficiently remote from other facilities, such as pipelines which might suffer harmful effects. If ground terminations are installed 6 to 7 miles from the vicinity of such pipelines and cables, no harmful effects will be incurred.

7. OVERHEAD LINES

The electrical design of a D.C. overhead line is in some respects simpler because only resistance and not reactance and capacitance need be considered in voltage regulation and in calculating power loss. On direct current lines, due to the difference in the ionization mechanism around the positive and negative conductors, the corresponding corona is also different. It is a characteristic of D.C. corona that the charge released must be carried to the ground or to a conductor of opposite polarity which is not the case for alternating current. Thus a space charge, which reduces the voltage gradient, will be formed around the conductor. The effect is so pronounced that it is usually more reasonable to characterize the corona performance by the line voltage than by the voltage gradient. Actual results have shown that corona loss5 in kW/kilometer on a ±250 KV D.C. line was 1.8 compared with 4.0 for a similar 380 kV A.C. line. Due to the fact that conductor diameter does not affect corona losses in D.C., the choice of conductor arrange-additional costs of the terminals

In addition, insulation requirements for nominal voltage are less with D.C. than with A.C., and a considerable saving in materials is brought about because only two conductors are required as compared with three for alternating current.

The saving in cost in a D.C. line are lower but transmission costs must, however, be off-set against the additional costs of the terminals.

Figure 8 shows a comparison of the cost of A.C. and D.C. systems related to the distance of transmission. Three voltage levels have been chosen, 345 kV A.C./ ± 250 kV D.C., 500 kV A.C./± 375 kV D.C. and 700 kV A.C./±500 kV D.C. and in each case the natural load for the alternating current compared with a similar direct current load.

Under present conditions and costs, in terms of distance alone, for a distance of less than 600 miles, A.C. transmission by overhead line is cheaper than the corresponding D.C. transmission. This distance is not inordinately large today. Planning is already underway for an 800 mile 2 600 MW line in the U.S.A. and for a 2500 mile 3 000-5 000 MW line in Russia.

8. SYSTEM INTERCONNECTION

Another reason for preferring direct to alternating current exists where, as in the case of the English cross-channel cable, the circuit acts as an interconnector between two large power systems. A D.C. circuit is an asynchronous connection, and each supply system can operate in accordance with its own requirements without special regard for the frequency and service performance of the other. The load to be transferred is determined by the settings of the control gear and is comparatively unaffected by the changes in the system conditions at either end.

An A.C. interconnecting circuit also has the disadvantage that it would add to the short-circuit infeed from one system to the other in fault conditions. A D.C. system does not transfer short circuit power between the two systems.

In the Japanese scheme, the eastern part of the Island has a frequency of 50 cis with a system size of 10 000 MW, while the western part of the islands has a frequency of 60 cis with a system size of 15 000 MW. These systems will be linked with a 300 MW 250 kV D.C. connection. By pooling power economically and by interconnecting the two systems, the, saving on reserve capacity and operating cost is estimated at 20 million yen/year.

9. OVERLOAD CAPABILITIES

The question of overload capability has often been raised and It is alleged that mercury arc rectifiers have 'very inverse" overload characteristics. Actually, as is illustrated in Figure 9, the capabilities vary little from those of power system capacitors.

Since series capacitors are necessary to obtain maximum performance tr m long transmission lines, the apparently inverse overload capacities appear to apply equally In. both cases. It must be pointed out, however, that Improvements in the overload characteristics of the D.C. valves are obtainable with only a nominal increase in the cost of the equipment, whereas a sub­stantial increase in cost results from a similar improvement in des capacitors.

10. THE FUTURE

Long term surveys of electricity requirements are regularly undertaken by the responsible supply authorities throughout the world. The possibilities of using high voltage direct current are receiving most careful attention in an increasingly favourable atmosphere. Reliability of the equipment is being proved beyond any doubt. Because of the increased demand and use, costs are becoming substantially less and continual reduction of the losses in the converter station is being brought about. Losses are now of the order of 2 per cent in the converter stations.

Figure 10 illustrates the growth of the D.C. transmission system from 20 MW to 2280 MW for the years 1954 to 1965. Contracts signed during 1965 boost this figure to a total of 6 880 MW.

The modern trend for the bulk transport of electric power over long distances favours the use of direct current. This is of special interest in the Republic of South Africa where the Cape load centre is situated nearly a thousand miles from the major coalfields.

11. WORLD RESEARCH PROGRAMMES

High voltage D.C. research programmes have been in progress intermittently since 1936. During the early 1940's however, major research programmes were initiated in Sweden which led to experimental lines 3 miles long with a capacity of 6 500 kW by 1945. In Russia, the Direct Current Institute was formed in the years immediately following the 2nd World War and a line with a length of 112 kilometers and a rating of 200 kV and 30000 kW was commissioned in 1950.

During 1963 the Big Eddy substation of the Bonneville Power Administration of America was brought into commission." Facilities include a five mile long overhead line and supply systems up to about 825 kV D.C. The aims of the test project are:

(1) To study radio and television interference problems.

(2) To study insulation co-ordination.

(3) To study earth electrodes and earth currents.

(4) To study control system performance as related to system requirements.

Radiated radio interference and conducted radio noise will be recorded, along with weather variables, each half hour, in digital form on a paper tape.

The facility will also include a flash-over test area where full-scale towers will be erected to study electrostatic field conditions around the cross-arm, insulator string and conductor attachment. Various types of insulator string assemblies will be erected and checked under either clean, polluted, wet, dry or foggy conditions with voltages up to 815 kV D.C. This research has become necessary because long transmission lines are proposed and planned in the U.S.A.

The power supply for test purposes will be erected in an inflated house or bubble of fabric which will be held up by positive air pressure. Figure 11 shows a model of high voltage d.c. the power supply in the "air house".

During 1964 a large experimental installation for research relating to high voltage D.C. converters has completed at the Marchwood Engineering Laboratories of the Central Electricity Generating Board. Methods of carrying out prolonged synthetic tests at full ratings are being studied, as well as network conditions when cable systems are used.

Britain has under active consideration 55 miles of cable transmission from Kingsworth to London for delivering 500 MW initially and 2 000 MW finally at 500 kV D.C.

12. RESEARCH IN THE REPUBLIC OF SOUTH AFRICA

To enable engineers in South Africa to keep abreast of the results of fundamental and applied research in high voltage D.C. transmission abroad, and also to undertake research on special problems peculiar to geography and weather conditions of this country, proposals have been made that a high voltage research programme of a similar nature to the Big Eddy project in the U.S.A., but on a smaller scale, be initiated in South Africa.

Special problems of lightning activity, height above sea-level, moisture and dust content of air, and which occur in unique combinations in South Africa, require detailed investigations in order to provide accurate and reliable data for our design engineers.

13. CONCLUSION

Despite predictions of failure by the more conservative engineers, direct current transmission has been proved not only feasible, but also reliable and economical. In the light of recent developments it is advisable that every electrical engineer be fully aware of the circumstances under which D.C. systems could be used to best advantage. It is hoped that the information given will assist engineers in their understanding of a subject that is still new to South Africa.

REFERENCES

(1) M. G. SAY, The Electrical Engineers Reference Book, Newness, 1964.

(2) ASEA Journal, Vol. 36, o. 7, 1963.

(3) Federal Power Commission, National Power Survey, 1964. Part 1.

(4) ASEA, The High Voltage D.C. Transmission to Gotland, 7313, 1954.

(5) N. HYL TEN -CA V ALLIUS, Insulation requirements, corona losses, and corona interference for high-voltage D.C. lines, Paper 63-998, IEEE, 1963.

(6) Direct Current, Vol. 9, No.1, 1964.