Fullscreen HTML5Med Res (??MB)HTML4

The Story Of Electrical Energy, Pt.8 Since AC power transmission first began, there has been a trend to higher and higher voltages, to keep resistance losses to a minimum. But there is a limit to the use of very high voltage AC and when this is reached, DC transmission must be used. By BRYAN MAHER The year was 1983, the location Foz do Iguacu, in Brazil the day overcast and hot. Watched by an assembly of invited guests, a VIP mounted a decorated dais to ceremoniously throw a small control switch. In response, giant transformer circuit breakers in an adjacent hall slammed closed, bringing on line generators of the Brazilian Itaipu hydroelectric system. Thus was initiated the first stage of the world's greatest power line which when completed would operate at 1.2 million volts and carry up to 6.3 gigawatts of electrical power. The receiving substation for this project was 800 kilometres distant on the opposite side of the country at Sao Roque, a suburb of Sao Paulo. Question: how can you economically transmit these large quantities of electrical energy over such great distances? And sell power to neighbouring nations as well? The partial answer is to use very high voltages. We have seen previously in this series the need for high voltages in long power lines. Not only are transmission losses reduced but also the corridor needed can be narrower for a given power to be carried. To transfer 7.5GW at 330kV would require about seven double circuit lines in parallel. These would occupy a corridor 250 metres wide. The same quantity of 7 __'_J I 333MVA ISOLATED , - - - TRANSFORMER I CORE AND CASE 330kV INPUT 2.255MV SUPPLY TO TEST LINE .,. FIG.1: THE TEST ARRANGEMENT used for the ASEA transformer set at the test range at Lakerville, USA. The core & case of the final transformer is alive, at 345kV above ground potential. A test cage allowed various weather conditions such as rain & mist to be simulated. 82 SILICON CHIP power could be carried by one 1200kV 3-phase line which would fit within a corridor 90 metres wide. UHV problems Ultra high voltage AC systems are, however, beset with problems. The extra insulation required can be mostly provided by the simple extension of known technology. But other difficulties emerge. Some countries, notably Russia, Sweden, the USA and Italy have been using experimental megavolt research installations for the p2.st 15 years. In the USA, west of Lakeville, Indiana, ASEA (now ABB) established a UHV test line of five spans on towers 61 metres high. Voltages up to 2.255MV are produced by a triple bank of cascaded ASEA transformers. The final transformer has its core and case alive, 345kV above ground potential, as shown in Fig. l. This complete transformer, weighing 290 tonnes , was mounted on a giant insulated pedestal. Conductor bundles of more than 1.2 metres in diameter, consisting of 18 subconductors, each 30mm diameter, have been tested. A 2.255MV disconnect switch and a test cage wherein various weather conditions such as rain and mist can be produced at will, are provided. Instruments measure voltage surges, control operation, effects of conductor height, RFI, ozone, corona power loss , audible noise, high frequency components and weather conditions. Switching surges of up to 100% overvoltage have been observed in some cases. The Ohio Brass Company contributed its laboratory facilities for insulator testing and other UHV research institutes from Canada, Italy and France cooperated. The ASEA company has always led the field in power line practice above 400kV, both in AC and DC applica- THIS HISTORIC PHOTOGRAPH shows the world's first 10kV mercury arc rectifier. It was developed at the ASEA laboratory in Ludvika, Sweden, from 1929-33. tions. That company inspired the world's first 400kV national grid system in Sweden as far back as 1952. EPRI research The Electrical Power Research Institute (EPRI) of USA contracted with the General Electric Company to conduct UHV tests at GE's facility at Pittsfield, Massachusetts. Threephase lines of 1.5MV (1500kV) and 500 metres long are used to investigate optimum design. A wide range of factors must be optimised: height, phase spacing, bundle diameter and the number of parallel subconductors. Measurements made included RFI in the broadcast band and corona effects using different types of corona shields. The test line uses gantry span towers, 71 metres wide and 21.5 metres high, with the three phases suspended below the cross member on V-shaped suspension insulator strings bf glass, porcelain and other materials. Soviet UHV The Russians possess test rigs for designs up to 5MV or even 7.5MV (7500kV). Much research and development was done into 1150kV circuit breaker design , protection systems, transformers and electric field intensity. Due to transformer leakage reactance and circuit capacitance, voltage ringing oscillations occur when switches are opened. In one Russian case, sinewaves of 400kV at a frequency of 1800Hz were found superimposed on a 1150kV 3-phase line during switching operations. The insulation on lines, equipment and particularly transformer windings must be sufficient to cope with these overvoltages. But three fundamental problems of UHV 3-phase AC lines are harder to alleviate. These are line inductance, line capacitance and the noise generated by corona discharge. Inductance and capacitance As we have previously seen, any cable has inductance which produces an AC voltage drop proportional to the product of current, inductance and frequency. Also, we have noted THE ASEA/AEP 2.255MV research substation. The high-voltage transformer is mounted on an insulated platform (bottom of photo), while the 2.255MV busbars are mounted 23 metres above ground. MARCH 1991 83 ':Qt!Q!:!Yll!:!:.!.~ 1oc-- 3-PHASE GRID SYSTEM CONTROLLED INVERTER L TRA~Wci~~ER .,. STEP-DOWN TRANSFORMER --1oc SENDING END - - - - - - i t------- RECEIVING END -------t FIG.2: BASIC SCHEME for a high-voltage DC transmission system. Either 50Hz or 60Hz AC power is generated in the usual fashion at the transmission end & this is then rectified to DC for transmission over the line. At the receiving end, the DC is converted back to AC & transformed down so that it can be fed into the system grid. that the natural phase to phase and phase to ground capacitance demands a charging curreBt. This can run to many hundreds of amperes and is also proportional to frequency. Capacitance charging current is an insurmountable problem in AC underground and submarine power cables. The close spacings of conductors and earthed shields results in huge capacitance values. In fact, inductive and capacitive effects greatly exceed the line voltage loss due to ohmic resistance of the conductors. Compensators Excessive inductance in an AC power line can be compensated for by inserting capacitors in series with the line at intervals. However, this is very expensive. Compensation by this method becomes uneconomical for lines carrying lGW or more over distances exceeding 500 kilometres. To correct for the high values of shunt capacitance in long high voltage underground cables, inductors must be used. These are connected in parallel with the line at intervals along its length. But again, the cost escalates way above the cost of cable and simple trenches. For long undersea cables though, this form of compensation is impossible. In wet and foggy weather, the corona phenomenon generates a lot of noise as well as a characteristic blue glow around the conductors. The noise , based on the system frequency (50 or 60Hz) and its harmonics, can be a loud buzzing and sputtering with components from 50Hz to many kilohertz. The DC solution Wouldn't it be nice if we could remove all these frequency dependent problems? Well, we can - just reduce the frequency to zero! Then line inductance would not be important during steady current flow. Further, line capacitance would produce no further effects after initially being energised and charged. Corona would still produce a pretty blue glow around conductors in the rain, but the generated noise levels would be vastly reduced, down to a faint hissing sound. The answer, of course, is to use high voltage DC. Thomas Edison would stand up and cheer were he still alive. Nearly a century ago, AC transmis- ISOLATED AC SUPPLY / WATER COOLER AND PUMP VACUUM TIGHT HVBUSHING - TANK ALIVE AT POSITIVE DC POTENTIAL - EVACUATED WATER COOLED STEEL TANK +15DOVDC LOAD . PERHAPS. 5kA FLOOR Corona noise Corona, as we saw in past chapters, does not cause large losses, though it does increase dramatically in damp weather, especially in wet snow. The deleterious effects of corona are RFI and acoustic noise. 84 SILICON CHIP FIG.3: A MERCURY ARC rectifier consists of an evacuated steel tank containing a mercury pool, an anode & a starting electrode. A starting current is used to vaporise the mercury to produce electrons & positive ions. When the anode swings positive, the electrons quickly accelerate towards it & a high current flows. However, when the anode swings negative, the heavier ions accelerate towards it quite slowly & so only a very small back current is produced. sion was chosen in preference to the DC systems advocated by Edison; purely because of the ease of transforming an AC voltage. The world has now turned full circle; we are facing the difficulties engendered by the frequency factor as we build higher and higher voltage AC systems. But note that it is the AC line fre quency, not the voltage, that gives trouble. Therefore, DC was chosen for that monstrous 6.3GW power line mentioned at the start of this episode. AC-DC-AC Nobody in his right mind would attempt to generate huge quantities of high voltage DC in rotating machinery. No! We still generate AC at 50Hz or 60Hz in normal alternators, at voltage in the 1 lkV to 33kV range. Transformers then raise the voltage to UHV (ultra high voltage) values around the sub-megavolt region. This is then rectified to DC for transmission over long distances. At the far end of the line, more equipment inverts the DC currents back to AC at normal system frequency (50 or 60Hz, depending on the country). This AC is usually transformed down (in a normal transformer) to feed into the system grid at the load end. The block diagram of Fig.2 explains the concept. Rectifier valves The transformers used are just extensions of known designs. But what was needed was the development of some types of valves to rectify these high AC voltages to DC, then subsequently to invert back to AC at the far end. The story starts way back before any of us were born. Though Thomas Edison discovered the vacuum diode rectifier in 1885, he saw no applications for it; nor did he give the modern theory of its operation. Physicists like Richardson (1902) explained valve rectification of AC to DC as being due to flows of negative charges to a positive anode . But a negative anode would support no such flow. So AC supply could sustain current flow in one direction only through such a valve. Rectification of AC to DC was thus achieved. As we survey the evolution of HVDC systems, it is interesting to note how early some of the techniques were known. Richardson gave the THIS IS ASEA's high-voltage DC laboratory at Trollhatten, Sweden, in 1944. Note the steam produced by the water-cooled load resistor when testing HVDC valves at full power. name "thermions" to his mobile negative charges within a valve. Today we simply call them "electrons", the Greek word for amber. His vacuum tube diode rectifiers in 1902 were called kenotrons. Around the turn of the century, scientists had perfected the use of these tubes in rectifying 100kV AC to high voltage DC. The small currents available were used for x-ray experiments. By 1932, Cockroft and Walton were producing 700kV DC supplies from voltage quadrupling diode rectifiers. Other engineers in the 1920s were using gaseous valves to rectify AC INPUT STEEL TANK ~ 'POSITIVEGRIO --,- ·PULSE SUPPLY ANO TIME'R - PULSE TRANSFORMER CONTROL GRID LOAD FIG.4: THE OUTPUT VOLTAGE of a mercury arc rectifier can be varied by interposing a control grid in the electron stream between the mercury cathode and the anode. If a sufficiently large negative potential is applied to this grid, all electrons in the mercury plasma gas will be repelled and none will pass to the anode. The valve is then in the cutoff state and no current can flow. MARCH 1991 85 V4 V6 pelled and none will pass to the anode. The valve is then in the cutoff state and no current can flow. In this way, a mercury arc rectifier can be used as a controlled rectifier, similar to an SCR. V2 Switched operation f.~ ~ I-- V1 :!,p_HASE OUTPUT TRIGGER PULSES TO EACH VALVE V3 V5 SYNC SIGNAL INPUT --i □ C HVDC LINE NEGATIVE RETURN FIG.5: BLOCK DIAGRAM of a DC-AC inverter at the receiving end. V1-V6 are controlled high-voltage mercury arc valves which are switched in turn to provide current waveforms to a 3-phase output transformer. The trigger pulse generator turns off each rectifier at the correct time by injecting a large negative pulse to the anode via a capacitor for a sufficient time to allow the arc to cease, thus giving control back to the grid much greater currents in 600V to 3kV circuits. One of the gases used was mercury vapour. Mercury arc valves Soon this technique led to larger valves in which the cathode was simply a pool of mercury at the bottom. In the late 1920s, these mercury arc rectifiers were widely used for supplying rail traction currents at . voltages in the 600 to 1500V range. The original glass envelope had then been superseded by steel tank models. Fig.3 shows a typical mercury arc rectifier which uses an evacuated steel tank. To start the rectifier, a current is passed through the mercury pool and an initial arc drawn. This vaporises some mercury to a heavy gaseous plasma of electrons and po_sitively charged mercury ions. The AC supply is connected to an anode of iron, carbon or patented alloys. Whenever the AC on the anode swings positive, it attracts electrons and, because these have only a tiny mass, they accelerate very quickly, flowing during the whole positive 86 SILICON CHIP half-cycle. Moving electrons constitute an electric current, and so many thousands of amperes can flow through the rectifier. When the AC supply on the anode swings negative, it attracts the heavy positive gaseous mercury ions. But as these ions have very great mass, they accelerate only slowly. The result is that the negative half cycle is over before any appreciable number of positive ions arrive at the anode. This back current can be measured in microamps or milliamps. Thus, a mercury arc valve rectifier passes useful current only when the anode is positive. The output is taken from the steel (cathode) case of the rectifier. Control grid The output voltage of a mercury arc rectifier can be varied by interposing a control grid in the electron stream between the mercury cathode and the anode, as shown in Fig.4. If a sufficiently large negative potential is applied to this grid, all electrons in the mercury plasma gas will be re- Once in conducting mode , the mercury valve continues passing current as long as the anode is positive, without any regard to the grid potential. We thus need a method of interrupting this current flow and this can · be achieved in two ways. The first , employed where the receiving end has no local AC supply, uses a transformer or inductance in series with the DC line to each valve as shown in Fig.5. To stop the valve conducting, the continuous DC supply must be interrupted momentarily and the grid held at cutoff negative bias. To achieve the currents in the three output phases, the appropriate pair of valves is switched on by releasing the negative bias on them at the correct timing. Then, when the output current in that phase is to be stopped, the grid is taken negative beyond cutoff and a large negative pulse injected at the valve anode. This momentarily makes the anode negative for a sufficient time for the mercury arc to cease, giving control back to the grid. Six timing circuits (3 for the anodes & 3 for grids) are required. Instead of transformers, series inductance and capacitance coupled anode pulses may be used. These methods are seldom used nowadays as systems expand. First HV mercury valve Building on their own experience and earlier German and English results in 1.5kV traction rectifiers , the ASEA company of Sweden developed a lOkV prototype from 1929-33. This, the world's first high voltage mercury arc valve, was set up in their laboratory cJ,t Ludvika for testing in 1933. Though mercury arc valves give no problems in the 600V to 3kV range, the idea of rectifying voltages up to a megavolt is frightening. In the conducting mode, the forward drop is only 15V and so there are no problems here. However, in the reverse mode, with the anode negative at hundreds of kilovolts , you would 100kV AC SUPPLY FROM TRANSFORMER ample power was available. Mercury arc valves were tested using a steaming water resistor as a high power load. METAL 1-r-..---...-.-1---- ANODE World's first HVDC line C EVACUATED - - CERAMIC HOUSING FREQUENCY COMPENSATED VOLTAGE DIVIDER TIMER SIGNAL --------++UPPER COOLING SYSTEM NEGATIVE GRID BIAS AND POSITIVE 1----1---1-- _ TRIGGER PULSE AND TIMER SYSTEM CONTROL GRID I ..!_ _ _ CIRCULATING WATER JACKET FIG.6: TO OVERCOME BREAKDOWN problems when rectifying high voltages, ASEA developed a mercury arc rectifier with a series of nine intermediate electrodes between the anode & control grid. A voltage divider connected to these electrodes thus provides 10 steps of lOkV each, which means that the field experienced by any one positive ion is drastically reduced (10kV vs. lOOkV). naturally expect voltage breakdown under such a tremendous electric field. To overcome this problem, ASEA researchers developed a more advanced valve which reduced the electric field experienced by any positive mercury ion. This was achieved by having the electron stream pass through a series of nine intermediate electrodes, each at equal smaller increments of voltage, obtained from a frequency compensated voltage divider - see Fig.6. The whole compound-anode assembly was mounted inside a ceramic extension of the water cooled steel tank. A valve with nine intermediate grating-like electrodes plus one final solid anode would have the full voltage, say lO0kV, applied to the top electrode. The voltage divider provides 10 steps of lOkV each in the space between intermediate anodes. This is sufficiently low to prevent voltage flashover or breakdown during the negative half cycle when the valve is in the non-conducting state. In 1943, an assembly consisting of four parallel mercury valves was tested in the ASEA laboratory at Ludvika. Each valve contained nine intermediate electrodes, and the whole arrangement ran successfully on 40kV at a group current up to 200 amps. Trollhattan Hydro Under a 1944 agreement with the Swedish state power board, ASEA conducted further high power mercury valve tests at Trollhattan, close to the hydroelectric power station. A point readers may not have considered is: how do you conduct a full power test on any power equipment unless you have that much power available? For this reason, ASEA shifted its research and development facility to this new location where In 1946, the many years ofresearch by ASEA engineers and scientists came to fruition. In that year, the world's first high voltage DC power line was built and put into operation. This 60 kilometre feeder operated at 90kV DC and carried 6.5 megawatts of power from Trollhattan power station northwards to Mellerud on the shores of the Vanern . So successful was this line that the concept of mercury arc rectifiers, inverters and HVDC power transmission was becoming a reality. Subsequently, a power supply was needed from mainland Sweden to Gotland, an offshore island in the Baltic Sea. An undersea cable was proposed from Vastervik on the mainland to Visby on the island, a distance of 105 kilometres. As the seabed along the route does not exceed 100 metres in depth , laying the cable was not a problem. Specifications called for 20 megawatts to be carried at lO0kV. At such a high voltage, an AC undersea cable would have been impossible due to capacitance effects. So, in 1950, ASEA was contracted by the Swedish State Power Board to develop and install suitable mercury arc valves. As there already existed a power station on the island, the DC link could in principle carry power in either direction. Thus, rectifiers and inverters were made as identical twins, and the timing sequences to the valve grids would decide the power flow direction. The Gotland Link made history in 1954, being the world's first high voltage DC undersea cable. The fame of that line is understandable as it opened the way for cross-channel power transfer anywhere in the world, a concept totally foreign to AC designs. Exhaustive tests on 140 mercury arc valve designs occupied four years of intense development. To carry the required 200 amps DC current, ASEA installed a bridge circuit at each end of the line. Each diode therein consisted of a grid controlled mercury arc rectifier having two parallel anode assemblies rated at over lO0kV. MARCH 1991 87 DEVELOPED BY ASEA at Ludvika in 1943, these four high voltage mercury arc valves contained nine intermediate electrodes (see Fig.9) & ran successfully on 40kV with a group current (anodes paralleled) of 200 amps. The nine intermediate electrodes of each rectifier are mounted inside a ceramic tube which sits atop the water-cooled steel tank. Each anode contained at least 10 intermediate electrodes. Undersea cable The single-core undersea cable was specially designed and manufactured. To achieve the greatest active conductor area in a given space, each copper strand was squashed from circular to roughly hexagonal cross section. Thus, 60 copper strands were laid up in a 5-layer pattern. The surrounding 150kV insulation was itself encircled by a neutral shielding layer, aluminium tubing, layers of steel armour stranding and a waterproof covering. Note that you cannot operate a steelarmoured single core cable on AC. If you do, the current carried by the copper will set up strong AC magnetic fields in the steel. The resulting 88 SILICON CHIP eddy currents in the armouring will overheat the cable, damage the insulation and cause breakdbwn. But the same cable carrying DC gives no problem. Sure, there are strong magnetic fields in the steel, but DC fields do not induce eddy currents! Sea return Another innovation inspired by the historic Gotland DC submarine cable is the idea of sea re.turn. In this scheme, the copper cable carries current for one half of the circuit, while the return current flows through the seawater. This practice does more than just save a copper return cable. As the ocean has an almost infinite cross section, the sea return path has almost zero resistance. Thus, cable losses are half that which would re- sult if a copper return cable were used. Electrodes of large surface area implanted in the sea at both ends of the line provide the connection to the salt water. ASEA continued their research and development of controlled mercury arc rectifiers and inverters for a further 17 years. Later valves carrying 1000 amps per anode at voltages up to 135kV established the HVDC concept beyond doubt. Installations all over the world followed and then solid state inverters appeared on the scene. We hope to cover these in a future issue. SC Acknowledgements Special thanks to ABB Australia and Sweden for supplying historic photographs and data; to ASEA Journal and Action; to General Electric and IEEE Spectrum.