Imagine you need to start a fire. Pretty simple with today's technology, with a lighter or with matches, but quite difficult without. Still, humankind would not be able to produce lighters and matches without harnessing fire first. History shows that the early methods of starting fires (Fig. 1) were quite exhausting but, when successful, the effort was very rewarding.
In fusion research, we aim at releasing and controlling energy almost a million times more powerful than fire: the energy that drives stars. The task is at the very limits of present technology, but it is almost within our grasp. JET can achieve the extreme conditions (namely extreme temperatures) under which massive fusion energy can be briefly released. Let us take a look at the power sources needed to get us to this point....
JET is capable of producing hydrogen plasmas (completely ionised gas) with temperatures of hundreds of millions of Kelvins (or degrees Celsius). Obtaining such high temperatures requires extraordinarily powerful heating (See Fig. 2). Powerful heating is also needed to sustain these temperatures, otherwise the plasma would rapidly cool down due to inevitable heat losses via radiation and heat convection/conduction. Given that the temperature gradient from the vessel wall to the plasma centre is about one million degrees per centimetre, it is easy to see that the plasma can lose energy very quickly unless it is well insulated. Standard thermal insulation methods are totally inadequate so JET uses a magnetic confinement system to retain heat in the plasma by using magnetic fields to keep the plasma away from the vessel walls.
Plasma heating is not the biggest consumer of energy at JET. In reality, a significant amount of power is needed to feed the large coils (see figure) which produce the strong magnetic fields to keep the plasma under control and away from the vessel walls. Because the coils have resistance, the large currents in the coils cause them to heat and they need to be water-cooled as a consequence. The energy is mostly dissipated to atmosphere via special cooling towers. Some fusion experiments, like Tore Supra (France), LHD (Japan) EAST (China), Wendelstein 7-X (under construction in Germany), KSTAR (under construction in Korea) or the future ITER use superconducting coils that avoid this energy loss but at the cost of running them at very low temperatures, around -270 °C, using liquid helium.
Every individual plasma experiment at JET (called a JET "pulse") lasts several tens of seconds and during experimental campaigns there are some 30 pulses a day. In other words, most of the JET power consumption is concentrated in short bursts, which is quite demanding on the electricity grid and on electrical engineering in general. Moreover, even during a single pulse, the power requirements are not constant - the pulse startup (magnetic field set-up and initial plasma heating) needs more power than the "plateau", the sustaining phase. The toroidal field coils (Fig. 3) are the largest single load on JET. The poloidal field system, on the other hand, has complex switching and control requirements. After the plasma has been created, its position and shape is feedback-controlled by taking sensitive magnetic measurements and supplying additional power to the vertical and horizontal poloidal field amplifiers according to plasma behaviour.
As a comparison, the typical power of a central heating boiler in a family house is around 25 kW (kilowatt = thousand watts). Running a JET pulse requires around 500 MW (megawatt = million watts) of power, of which more than a half goes to the toroidal field coils (Fig. 4). Around 100 MW of power is needed to run the poloidal field system (ohmic heating and plasma shaping coils) and the rest (~150 MW) runs the additional heating sources (neutral beams and RF heating).
Of our plasma heating systems, the total input power to all the neutral beam heating systems can be up to 140 MW, and for the Ion Cyclotron Resonance Heating can be up to 90 MW. Additionally, the Lower Hybrid Current Drive system can support the JET plasma current, and the installed input power of this system is several tens of MW.
The energy conversion efficiencies of all heating and current drive systems limits the power that the plasma receives. The installed output power of the neutral beam heating system is 23 MW, and that of the Ion Cyclotron Resonance Heating is 32 MW of RF power. Lower Hybrid Current Drive can achieve 12 MW of microwave power. However in most JET pulses only part of these installed capacities is exploited, depending on experimental scenarios. Last but not least, the plasma also gets a few MW of power from ohmic heating, ie the heating due to electric current induced in the plasma by the inner poloidal coils. In total, JET plasmas usually consume a few tens of megawatts.
The plasma accumulates only a fraction of the consumed energy. The rest leaks away via radiation, heat conduction and particle losses (Fig. 5). The "energy confinement time" is a simple measure of our ability to reduce those leaks. The time is equal to the ratio between the total plasma energy (Joules) and the total heating power (Watts = Joules per second) needed to sustain such plasma energy. In the case of JET, the energy confinement time is usually close to one second. That is, with power consumption well above 10 MW, the total heat energy of the typical JET plasma is more than 10 MJ (10 million Joules). The heat energy of a hot truck engine is comparable to this number - but keep in mind that its weight and temperature differ a lot from the JET plasma. The latter weighs only tens of milligrams but is at a temperature of hundreds of millions of degrees.
Perhaps you may be thinking that JET's fusion research facilities are inefficient. That wouldn't be fair - they are efficient in their task, which is achieving the extreme conditions required to initiate fusion and producing plasmas on which to perform research. We are back to our first picture now: the priority of that Iron Age human was to start a fire, not to spare the energy of his body. In clear parallel to him, we are confident that our efforts will eventually pay off. In fact, we cannot imagine sustainable progress of humankind without first mastering fusion energy (Fig 6 & 7). |
Fig. 1 Primitive methods of starting fires were energy demanding
Fig. 2 Plasma and its heating
Fig. 3 JET's coils and plasma
Fig. 5 Plasma energy balance
Fig. 6 Fusion-powered spaceships are anticipated to explore outer Space (image courtesy of NASA)
Fig. 7 Conceptual design of a future fusion power station
|
In total, JET's power supply system has an installed capacity approaching 1400 MW (MW = megawatt = one million watts), a significant proportion of the maximum output of a large power station. However, not all of the installed capacity is necessary when an experiment is run - the different systems rather serve as a portfolio of many options on how we may produce various plasma conditions. Even more importantly, most of the power requirements are concentrated into short time periods of plasma pulses, followed by much longer quiet periods of machine cooling and data processing. This interval can be used to accumulate stored energy on our research site, thus providing a powerful local source that can considerably reduce the national grid load during the subsequent plasma pulse.
Two methods of energy storage are applied in fusion research facilities. Large banks of capacitors are used on small and middle-sized machines with short (flash-like) experimental pulses. On big machines, energy may be stored using massive flywheels. JET, a large tokamak with pulses extending 20 seconds and more, is an obvious candidate for the flywheel solution.
Each JET pulse consumes around 10 GJ (GJ = gigajoule = one thousand million joules) of energy, with the peak power requirements exceeding 1000 MW. This amount of power cannot be taken directly from the UK National Grid so two massive flywheel generators (Fig. 9) are used to supply the additional energy needs. The rotating part (rotor) of each generator is 9 metres in diameter and weighs 775 tons (!), much of which is concentrated on the rim to form a large flywheel. For experts - the total moment of inertia is 13.5 million kg.m2 per flywheel!
Before each pulse the flywheel is accelerated by its 8.8 MW electric motor - even high-speed trains like Eurostar or TGV have less powerful motors...
Each flywheel can be spun up to 225 rpm (3.75 Hz) so that the edge of the flywheel rotates at the speed of 380 km/h (236 mph)! That is where the rotor carries the pole windings. Positioned as closely as possible to these rotating windings are stationary pole windings mounted on the stator, which is the fixed cylindrical construction around the flywheel.
When power is needed for a JET pulse, the rotor pole windings are energised. In other words, electric current is sent to the rotor windings so that they start producing a strong magnetic field. The magnetic field immediately interacts with the stationary windings. According to the laws of electrodynamics, the stationary windings start producing massive electric power at the expense of the kinetic energy of rotor gyration: magnetic forces act as a powerful brake that slows the flywheel down to approximately 112 rpm.
Each flywheel generator is capable of providing 3750 MJ of energy for the JET pulsed power systems, with a maximum of 400 MW power output. One generator supplies toroidal field coils, another the poloidal field coils. The power output from the generator is alternating current (AC) but JET pulses need direct current (DC) so each generator features four large AC/DC semiconductor converter units.
The remaining power required during the pulse - namely part of the toroidal coils' consumption and all the additional heating - is obtained directly from the national grid. Again, semiconductors (diodes and thyristors) must be used to convert the grid AC power into a dynamic DC form suitable for JET. The power used for pulsed operation is supplied from the 400 kV grid. In addition, continuous electrical power is provided by two 132 kV supplies to run the ancillary equipment.
An important advantage of the Culham site for the JET facility has been the vicinity of the Didcot Power Plants (Didcot 'A' and Didcot 'B'). This huge enterprise with coal power plant and combined cycle gas power plant, with total installed electric power of 3400 MW, is located only some 5 km (3 miles) away from JET.
However, when the UK public's electricity consumption hits a peak, the National Grid operator can quickly inhibit the pulsed operation of JET in order to prevent overload of its power plants. Our scientists, being naturally very curious people, have tried to find out when these periods of "JET blackouts" are likely to occur. To our surprise, the intervals of TV advertising spots that are broadcast in the middle of highly popular programmes (eg Coronation Street, football finals etc) are common causes of delays in JET's evening operations. Presumably adverts cause millions of viewers to switch on their kettles all at the same time! |
Fig. 8 JET power supplies and their connection to the National Grid (the arrow shows camera position for the next photograph)
Fig. 9 A bird's eye view of the JET Power Supplies
Fig. 10 One of the two JET flywheel generators during construction
Fig. 11 Side-view of the complete generator
(the arrow shows camera position for the next photograph)
Fig. 12 Inside the generator: Stator and rotor pole windings
Fig. 13 Sectional view of a flywheel rotor
Fig. 14 The JET connection to 400kV National Grid from Didcot/Cowley
Fig. 15 90° Panoramic view of the Didcot Power Plants and JET (5km apart)
|
