Superconducting magnets are known for use as energy storage devices. However, a major problem with utilising a superconducting magnet is that the very rapid energy dissipation from and supply to the magnet results in significant stresses and strains on the windings with the result that these move, causing heat generation and the generation of a quench condition.

In accordance with one aspect of the present invention, an energy storage device comprises an array of superconducting electrical coils made from low T _c material, the axes of the coils being substantially parallel and the coils being laterally spaced apart in a non-overlapping manner. We have found a way of practically embodying an energy storage device, for example an electrical surge protection device, using low T _c superconducting materials which overcomes _.he quench problem and thus enables rapid release and supply of energy from and to the device. In practice, energy is required for a relatively short time, typically less than one second and, following dissipation, the device must be reenergized usually within about ten seconds. This becomes possible with devices according to the invention.

Further advantages are that arrays can be assembled to provide any field profile that a continuous track can create. "Further, they can provide profiles that in practice could not be achieved with continuous tracks if theory demanded the continuous track followed very small radii of curvature. Stress concentration and local strain arise in conductor near sharp bends increasing the chance of wire movement under e/m forces. As an example, it might be an advantage in some circumstances to shield the array using a "box permanent" magnet design using "cladding" coils. Each coil can be simply supported mechanically against e/m forces and thermal strain. Because of the, typically cylindrical, shape of the individual elements, windings may

SUBSTITUTESHEET

be simply clamped and "glued" into position without much risk of discontinuities. The short continuous length in each coil reduces the chance of discontinuous strain due to thermally induced volume changes in the coil. There is a greatly increased possibility to design the array for optimum distribution of strain energy such that small disturbances do not release a cascade of micro movement down a strain energy gradient.

The invention allows systems to be flexibly connected to match load time of charge discharge demands. For example, a more rapid partial discharge could be achieved taking all of the energy out of one section than partial extraction of energy from the whole array. The array approach allows for elemental failures if parallel connection of subsets of arrays is used giving the system damage tolerance. During service elements could be individually replaced.

The geometry of an array of coils lends itself to inter-connection of protection circuits and early quench sensors which trigger active protection systems.

The array of coils can provide more options for mounting cooling channels than in a massive one-piece winding. Cooling can be increased locally to assist in quench protection. In order to minimise heat generation within the superconductor filaments of an LTSC composite during field changes, very small filament diameters are required. Standard filamentary conductors have filaments around 70 micron diameter. For micro SMES 2 micron diameter filaments are needed.

The co-processing of 2 micron filaments of niobium titanium in copper matrix conductor is much more likely to lead to conductor fracture during drawing. Fractures occur most often as the conductor approaches its final size, and therefore, it has high work cost. The ability to relax length requirements could be a major source of cost

reduction (with joint arrays joints can be made efficiently as the joints can be located in low field regions) .

In a coil array, "more flux threads more turns" resulting in an increase in the inductance per unit volume of the system. This means lower refrigeration costs per unit of inductance and more compact structures that are easily sited and shielded (prevention of fringe field hazard will be mandatory in view of the rapid rate of charge of field) . The coils may be fabricated from a conventional low temperature (T _c ) superconductor where the coil superconductor is at liquid helium temperatures but is also applicable to higher temperature superconducting materials but in general only materials which superconduct at or below liquid nitrogen temperatures.

Typically, the electrical coils will carry working currents in the same sense.

The coils may be connected in series or in parallel to input and output connections. The parallel arrangement enables selective use to be made of the electrical coils and thus is able to achieve partial energy release if this is required.

In some cases, some or all of the coils could include further coils nested within them. An example of an energy storage device according to the invention will now be described with reference to the accompanying drawings, in which:-

Figure 1 is a cross-section through the device showing the coil arrangement but with supporting and cryostat structures omitted;

Figure 2 is a longitudinal section through one of the coils shown in Figure 1;

Figure 3 illustrates graphically the variation of magnetic field profile along a radius of the device shown in Figure 1;

Figure 4 illustrates graphically the magnetic field profile due to a single elementary coil, and along a radius

R of an array of seven such coils and the array of thirty seven coils shown in Figure 1 respectively;

Figure 5 illustrates the magnetic field profile in the radial direction for a single large coil equivalent in size to the coils shown in Figure 1;

Figure 6 illustrates the radial magnetic field profile of the array shown in Figure 1 with the outer members weakened;

Figure 7 illustrates the magnetic field profile around the inner circumference of the central coil in Figure 1; Figure 8 illustrates the magnetic field profile around the outer circumference of the central coil in Figure 1;

Figure 9 illustrates the magnetic field profile around the inner circumference of an outer coil of the array shown in Figure 1; and.

Figure 10 illustrates the magnetic field profile around the outer circumference of an outer coil shown in Figure 1.

Figure 1 illustrates a hexagonal array of 37 coils 1- 37. Each individual coil in the array has the dimensions shown in the longitudinal cross-section shown in Figure 2. The inner radius is 0.9 metres, the outer radius 1.0 metres and the length 2 metres. For the purposes of this example, the current density in the windings is assumed to be 7.5 x 10 Am ^" . The coils are wound from conventional low T _c copper wire on formers and within a cryostat which have been omitted for clarity. The coils may include additional coils nested within although this is not shown in this example. Magnetic fields have been calculated for this system of coils using well-known techniques using Ampere's Law and involving numerical integration over the conductor volume. These calculations have been repeated with several different step sizes used in the integration to ensure that the results are sufficiently accurate.

Stored energy has been calculated from the integral of the square of the magnetic field intensity over all the

volume of space in which the field is appreciable. To confirm accuracy, these calculations were repeated with different spatial limits for the integration.

To illustrate the development of this example system, calculations have been made for

1. A single coil element;

2. seven such elements in a hexagonal array (those numbered 1 to 7 in Figure 1.) ;

3. The complete array of 37 coils as shown in Figure 1;

4. A similar array of 37 coils in which the outermost members (numbers 9, 10, 18, 15, 16, 23 etc) have been weakened by reducing their winding cross-section. This was done by changing the inner winding radius from 0.9 metres to 0.93 metres;

5. A single large coil 38, whose inner radius is 6.9 metres, outer radius 7 metres and length 2 metres. This coil, when viewed from a distance which is large compared with the size of the small elementary coils can be shown to be equivalent to the array of 37 coils, and indeed the energy calculation showed it to have very nearly the same stored energy.

Figure 3 shows the magnetic field profile along a radius R of the array of 37 identical coils 1-37. Figure 4 represents three similar profiles 40,41,42 superimposed, for respectively a single elementary coil, for an array of seven, and for 37. (Cases 1, 2 and 3 above) .

Figure 5 is the field profile for the equivalent large single coil 38, (case 5) presented for comparison.

Figure 6 is the field profile of the array of 37 with the outer members weakened. (Case 4) .

It can be seen that the arrays have similar field profiles to the large single coil, but with small regions of negative field near the outer edges of the windings.

Within the array, away from the edge, the return flux of neighbouring coils tends to reduce the field inside a

particular coil to a more moderate level than would be the case for an isolated coil. At the outer radii of the coils, the return flux reinforces the negative field, thereby producing the regions of strong negative field. At the edges of the array, these mechanisms do not occur. If the (positive) field strength at the inner winding radius of the outer coils is unacceptable (either because of the bursting forces produced, or in the case of a superconducting winding because of reduction of the current-carrying-capacity of the conductor) it can be reduced by weakening the outermost coils slightly, as shown in Figure 6.

Figure 7 and Figure 8 show the azimuthal variation in field strength around the inner and outer circumferences respectively of a coil at or near the centre of the array (eg one of the coils 1-7) . The field variation is small and as is to be expected has 6-fold symmetry. Figures 9 and 10 show the equivalent data for a coil at the edge of the array (eg one of the coils 9, 15, 20, 25, 30 or 35). Here the variation is much greater. It has been found that the detailed form depends on the shape of the coils. Relatively long coils such as these individually have uniform field profiles, and the result is a smooth curve dominated by a first order variation. Short hoops have less uniform fields and the resultant azimuthal field profile for a coil at the edge has both first and easily discernable third order components.

These detailed differences do not affect the principles of the invention, but it should be remembered that in a superconductor the current-carrying capacity will be determined by the maximum field value, but the hoop stresses, in any conductor, by the mean value obtained by averaging around the circumference.

The stress in the conductors can be calculated from the Lorentz force. For a conductor on the mid-plane, this force is J _Θ .B _Z per unit length of conductor and is directed radially outward. In the absence of any other support,

this force will be contained by a hoop stress in the conductor, given by S = J.R.B _Z .

Where additional support is provided, it is difficult to transfer all the forces to an external structure, because of the interposing of low-elastic-modulus materials such as insulators and cooling channels, as well as other conductors, and so substantial hoop stresses will still appear in the conductors.

The present invention can offer two benefits to the support of the electromagnetic forces. In the first place, the use of coils with small winding radius directly results in a lower hoop stress arising from a given Lorentz force. In the second place, the return flux from neighbouring coils results in a compressive component to the stress. This is particularly desirable where the conductor is a ώrittle superconducting compound in a ductile matrix. In this way, the coils provide mutual support. The structure is analogous to "honeycomb" composites used in producing stiffness without weight in aerospace and other fields of mechanical engineering.

The following table summarises the properties of the variations on the coil system and also those of the equivalent single coil.

The coils 1-37 may be connected in series or in parallel to terminals which enable the device to be connected via a switch to an electrical circuit. Operation of that switch will be as with conventional storage devices.

TABLE

37 elements 2.99 -6.20 -131 3.35 -5.95 -110 4.68 -4.56 -13 4.48 -1.94 16 weakened outers E » 550 V = 38.0

Single Large 5.49 -3.7 450 Coil

E - 780 V 8.73

Coil 1 is the central member of the array, coil 4 is an outermost coil, coils 2 and 3 are intermediate.

"al" refers to the inner winding radius, ⁿ a2" to the outer winding radius.

Field strengths are in Tesla, stresses in Megapascals.

E is the stored energy in kW-hours, and V is the volume of conductor used in the system in cubic metres.