GB2491826A - Magnetic bearing arrangement for a large flywheel utilising permanent magnets - Google Patents

Abstract

A pseudo-levitation permanent magnet repulsive bearing and drive system arrangement 1, 2, 3, 15 & 16 for a large diameter composite flywheel 14 for mechanical / electrical energy storage is provided. The system is designed to transmit torque and cope with the expansion of the flywheel 14 when it is running up to its maximum speed. The magnets 13 are arranged in a frustrated cone to centralise the running axis and to subject the upper rotating magnet array to radial centric forces opposing the centrifugal forces acting upon this array to maintain its integrity. This application describes how the levitation system functions utilising multiple concentric rings of a prime number of permanent magnets and special drive spokes, which deflect vertically and radially in concert with the flywheel as it speeds up. To prevent instability the lower fixed magnet arrays are rectangular, with square ends or basically rectangular with semi-circular ends such that the rotating magnets never overhang at either end, the fixed magnet as the flywheel expands due to centrifugal force. The spokes are attached to a hollow central shaft which rotates on ceramic bearings or other bearing types including oil lubricated, air or magnetic bearings, around a fixed central shaft.

Description

Page 1 Large Fjywheel pseudo levitation bearing and drive system

Background

To date the most commonly used technology for large electrical energy storage is pumped storage and historically it has been a dependable source of storage.

However they are: * Very capital intensive * Geographically limited to suitable mountainous areas or sea cliffs * Generally far from major centres of population and need o Long planning cycle o Long constmction lead time * Environmental opposition (Areas of Outstanding Natural Beauty. (AONB)) * Generally minimum parasitic storage losses except in hot dry climates due to evaporation.

Other competing Technologies

CAES

Multiple types of Batteries Ultra Capacitors How is this flywheel storage technology applied Today there is much focus on renewable energy, most common of which is wind energy and large GW of capacity have been and are being installed, with even more in planning. This power source of course varies from zero to a maximum according to a cubic factor of wind velocity and averages only about 30% of the rated capacity.

This is one major area where energy storage can play a part.

The second area where energy storage can and does play a part is in the management of the National Grid and Distribution systems. This can inyolve peak demand sustenance and load levelling for examples.

A third area for storage is to raise the night time generating levels to sustain, the recharging of the energy, which has the effect of possibly rdducing the total generating capacity needed and removing the thermal cycling modes of plants especially those taken off-line during the night and extending their life cycles as a result.

History of Large Flywheels for electrical storage Steel has been used for various flywheels for electrical storage for some time.

Notable examples are: The flywheels at Cuiham Laboratory: very large diameter, segmented steel construction, low speed, large power output for seconds to support the local grid on start-up of the collider.

Page 2 Fuji Electric in Japan in the 80's constructed a 6.6m diameter flywheel from thick steel machined plates, stacked in layers, giving a total weight of 650 tons and capable of 160MW for 30 seconds.

Current flywheel technology is focussed on very high speed carbon fibre epoxy composite cylinders, running with active and passive magnetic bearings in a vacuum I containment chamber. However their storage capacity is relatively small and their current applications are around fine tuning grid frequency control. Beacon is perhaps the best known as a manufacturer of such devices. These are all manufactured in a factory environment.

Problems to overcome in manufacturing large flywheels There is a limit to size that can be transported in terms of weight, height and diameter.

In the UK practical limits of size would be 6m diameter and 5m high on a special low trailer as a permitted load on a motorway / dual carriageway. Hence the probable limits on the Fuji Electric flywheel example above.

1. In energy storage angular velocity is more important than mass, which drives the carbon fibre I epoxy solution or other high strength fibre I resin combination.

To make this viable in terms of size and cost, a large diameter thick wall cylinder has to be manufactured.

2. The only way to achieve this is to manufacture the flywheel on site in a controlled environment, where curing and winding are continuous processes.

3. Practically the axis of rotation would have to be vertical. This enables the use of permanent magnets in repulsive mode to elevate the flywheel from beneath, which has a large surface area to contain all the necessary magnets and enable an almost frictionless bearing solution which is maintenance free.

4. There is the issue that magnets in such a mode cause the rotating mass to be unstable, according to Earnshaw and others. This can be overcome by mechanical means and the use of prime numbers and is known as pseudo-levitation.

5. The complete flywheel assembly must run in a near perfect vacuum to minimise parasitic losses. For example the proposed rim speed is Mach 3. The vacuum chamber / containment system proposed is concrete with a steel lid. Any concrete porosity issues and vacuum loss can be solved by detecting leaks with foam on the exterior and sealing with liquid epoxy at locations under vacuum where foam quickly disappears, indicating a leak. The two part epoxy will migrate into the concrete, cure and seal the ingress leakage path. Concrete construction would also utilize dumbbell waterbars of rubber, used vertically between each separate pour for additional ingress sealing.

6. The containment chamber does not have to be as robust as if steel were used in the flywheel construction. Carbon fibre / epoxy composite has the beneficial property that on failing it will delaminate and break into small pieces.

7. The design uses a completely hermetically sealed chamber by means of a magnetic drive system through the lid of the containment / vacuum vessel. So sealing is not an issue.

8. In addition to the issue raised in 4, which were passive mechanical devices it is also proposed to have an active mechanical device to assist in damping and balancing of the rotating masses.

Page 3 9. The effect of Precession torque from the earth's rotation on bearing design is important.

10. Critical speeds especially in the normal running range of 30 -100% angular velocity should be designed out.

11. Detailed site assembly / construction instructions will be essential.

12. Emergency equipment can be added in the event of instability to gravity feed water, pea gravel, fine dry sand, etc. into the chamber, (Japanese Patent) to dramatically slow flywheel. Water is preferred as this will cause least damage and cool flywheel in the event of a serious vacuum leak. However for this proposal SOLAS (Safety of lives at sea.) technology is deemed best spraying a water based fog into the containment vessel and venting it through automatic or spring closed doors in the lid of the containment vessel.

Proposed technical specification

(But NOT limited to.) * Weight 300 ton * Outer diameter circa 10 metres * Inner diameter circa 6 metres * Height 5 metres * Speed <3042 RPM. (FoS of)'2.) * 60 MWhr total storage. Useable storage 48 MWhr, e.g. 6 MW output for 8 hours.

* 4 pole 6 MW synchronous machine, running at synchronous speed, generating at 11kV to 15 kV or according to site specification and 50 or 60 Hz.

* Hydraulic, magnetic variable speed or hydrodynamic gearbox drive for constant speed.

* Low Voltage Ride Through technology by either drive system.

13. Limited torque coupling between generator and drive system to prevent damage to drive train from grid disturbances.

Additional notes: * Hydraulic drive system will have an accumulator in the circuit. If made very large, this could be additional energy storage capacity.

* Accumulator could assist with "Black Start" capability.

* With small standby generator or UPS for controls, "Black Start" is also possible.

* Generator if uncoupled could act as a synchronous condenser and produce Vars for Power Factor compensation.

* Hydrodynamic drive could raise possibility of 20MW output generator.

* Hydrodynamic drive could also act as uncoupling device to operate generator as synchronous condenser.

System Advantages (In addition to above) 1) Can be located next to major usage centres 2) Can lower the peak power being dispatched from whatever source to give lower transmission / distribution losses Page 4 3) Could be used to eliminate increasing ratings on transmission / distribution systems 4) Very low parasitic losses -estimated at <1% per hour 5) Synchronous generator running at synchronous speed 6) Flywheel overspeed almost impossible, with dual back-up controls.

7) Almost perfect wave form (No power electronics) 8) Will meet all Grid Codes for synchronous machines 9) Can be used as synchronous condenser, when not in use (Discharging or recharging) 10) Simplistic design, each component has its own function e.g. stator winding cooling in a vacuum is not an issue! 11) Unique dynamic and damped self balancing system proposal 12) Does not require major geographical features for operation. No need for sites with suitable storage areas and large elevation between for pumped storage (Mountains or sea cliffs) 13) No need for salt cavern storage CAES system. Strategically in UK, better used to store natural gas 14) Lowers the thennal cycling of aging power plants which is costly 15) Enhances the value of renewable energy 16) Raises productivity of existing distribution / transmission systems 17) Improves security of infrastructure 18) Short lead times for equipment.

19) Commodity Storage: Storing energy generated off peak or from surplus "Green energy" for use during peak demand periods during the day, permitting arbitraging the production price of the periods and a more uniform system load factor for transmission, distribution and generation.

20) Rapid grid connection / disconnection of a few seconds possible.

21) Fits well with National Grid uncertainty forecast 4-6 hours out.

22) Could fulfil short term operating reserves, where UK has a big shortfall for 2020.

Renewable energy resources have two problems.

1) "First, many of the potential power generation sites are located far from load centres. Although wind energy generation facilities can be constructed in less than one year, new transmission facilities must be constructed to bring this new power source to market. Since it can take upwards of 7 years to build these transmission assets, long, lag-time periods can emerge where wind generation is "constrained-off' the system. For many sites this may preclude them from delivering power to existing customers, but it opens the door to powering off-grid markets-an important and growing market." E5C White Paper, May 2002 2) "Second, most of the power that is generated that is accessible to the grids is generated when there is low demand for it. By storing the power from renewable sources from off-peak and releasing it during on-peak, energy storage can transform this low value, unscheduled power into schedulable, high-value products.

Beyond energy sales, with the assured capability of dispatching power into the market, a renewable energy source could also sell capacity into the market through contingency services". ESC White Paper, May 2002 "Modernize Conservation: Energy storage will help modernize conservation practices by optimizing the economic and environmental Page 5 * profiles of fossil and nuclear assets through reducing dispatch and cycling costs and/or providing new electricity products to the market." E5C White Paper, May 2002 * "Modernize Infrastructure: Energy storage will help modernize the nation's electric power infrastructure by enhancing its efficiency, reliability * and security. By operating assets in a more systematic and enhanced role, private sector firms will invest to upgrade the system." E5C White Paper, May 2002 * "Increase Energy Supply: Energy storage will increase the productivity of domestic resources in two ways. First, by increasing the efficiency of existing power plants, our existing re-sources will go farther. Secondly, storage will promote renewable energy use thereby adding domestic resources to the fuel source." E5C White Paper, May 2002 * "Protect the Environment: Storage facilities will help reduce the environmental impact of existing power facilities and more quickly and efficiently integrate renewable ones into the system. By operating existing facilities in a more efficient manner, they will produce less waste and reduce dispatch and cycling costs." E5C White Paper, May 2002 * "Security: Energy storage facilities enhance the reliability of the grid.

Reservoirs of energy stored at dispersed sites are less vulnerable to disruption, and can be called up at a moments notice to: 1) reduce volatility, 2) enhance the reserve margin, and 3) provide the necessary ancillary services critical to the proper workings of the transmission and distribution system." E5C White Paper, May 2002 * CCS: Facilitate optimal practices for Carbon Capture and Storage.

* End of life disposal: No major environmental impact. (Especially compared to battery disposal.) System Efficiency This is a key measurement and can also define the market best suited to the overall efficiency and especially parasitic losses during storage periods.

Generator 4 pole, 98% efficiency at >0.95pf Hydraulic or hydrodynamic or magnetic variable speed drive system, 95-97% Magnetic gearbox used as hermetic seal, 98.5% efficient Step-up I down transformer 98.5% efficiency (11kV / 66kV) Power in 10% losses Power out 10% losses 1. Overall 20% losses * Compared to pumped storage 20-25% losses * No parasitic losses on pumped storage during non-operation. (Exception -evaporation) 2, Parasitic losses are due to: * Minor losses in standard bearings on centre shaft and / or magnetic bearings * Some windage and friction as vacuum is not absolute zero * Magnet iron losses when gearbox acts as clutch * Mechanical damping losses * Losses estimated at <1% per hour Page 6

Description:

It is well documented that passive permanent magnetic bearings are unstable. It is the intent of this application to demonstrate a flywheel bearing system with active mechanical stabilising mechanisms and design of elevating permanent magnet bearing system, pseudo-levitation, for very large hollow cylindrical flywheels.

See Drawing 1.

The elevating permanent magnets forming the main portion of the arrangement are arranged in concentric rings on a hollow frustrated cone #1. The lower set #2 is stationary and upper set #3 fastened to the flywheel base.

See Drawing 2.

The purpose of this is two fold: 1. The included angle of the cone acts as a centralising device and an elevating device. #4.

2. It also acts as a repulsive force #5 against the centrifugal force of the upper rotating magnet array, #3. The angle can be such that it neutralises all or some of the centrifugal forces dependant on the rotational speed and the strength of the magnet upper mounting blocks.

The repulsive force is shown by #6 and elevating force by #7.

Reviewing the stability of the basic arrangement in the various planes: Vertical plane and induced wobble due to prime numbers: See Drawing 3.

As opposing magnets pass each other, the repulsive force varies as the upper and lower magnets rotate past each other. If not corrected this would lead to a vertical vibration being set-up. To minimise this effect, a prime number of concentric rings, #18, each comprising of a large and different prime number of magnets in each opposing ring is constructed. For example the top ring of magnets could be a number such as 283 and lower opposing ring could have same 283 magnets. This also rninimises any cogging effect due to the selection of prime numbers and their attributes.

Thus it is necessary to have a prime number of concentric rings of magnets, such as 11, with the "start" position, #8 & #9, or location of exact matching of a pair of magnets on each ring, displaced by 360 Degrees divided by the prime number of rings, such as 11 already suggested, shown by angle #12.

Note: For simplicity only 2 concentric rings of magnets are shown on Drawing 3.

More concentric magnet rings of differing prime numbers assist with stability and minimises notching or cogging.

#s 8, 9, 10 & 11 represent fixed or lower magnets in 4 locations only. #13 shows the illustrated two concentric rings of rotating magnets on the top array. Dotted lines, #54, represent the ID & OD of the lower fixed magnetic array.

Wobble will be best avoided if the same prime number of magnets are used top and bottom of each ring, and to let vertical plane vibration be controlled by prime number of concentric rings instead, (Say 11, 13, 17, 19 or even 23) offset from each other in terms of start position.

Page 7 Horizontal plane: In this plane the frustrated cone magnet mounts act as a correction agent. If the vertical axis shifts, Drawing I, #14, as shown by arrows, then the air gap, #15 on one side will decrease and on the opposite side, #16 will increase. This immediately will generate correcting repulsive forces as these forces are an exponential function of the air gap. Hence the magnets will be self correcting in this aspect.

Wobble movement: Since there is only constraint magnetically at the bottom of the flywheel, the ratio of height of the hollow cylinder and the maximum diameter will have an effect. For example if a ratio of 1:1 was used then this effect will be much more pronounced and difficult to control as the centre of gravity is elevated. A ratio of < 0.5 would be desirable. However the change of air gap in a "wobble" situation due to the large diameters of the elevating magnet rings and their placement across the width of the flywheel section, means that even in a wobble situation due to magnets in a frustum conical layout a "wobble" will produce a change in air gap across the elevating magnets on both sides, leading to corrective forces being applied.

Flywheel expansion due to centrifugal forces: As the flywheel speeds up it expands and if the same size I surface area of magnets were used in top and bottom magnet ring arrays, then a change in repulsive forces would result as only partial overlap would occur at full speed. (Assuming alignment at no speed.). This is overcome by using a rectangular magnet on fixed bottom array and a square one of the same or similar width on the rotating top array, such that the top magnet is always in full repulsive mode with bottom array of fixed magnets. See Drawing3,#8,9, l0&ll.

It would also improve any "cogging" effect if the upper magnet were round; so that as magnets passed each other, the elevating forces are smoothed as forces build smoothly and not suddenly in a notched format. Unfortunately this lowers the repulsive forces.

Notes on stability in general: In repulsive mode the primary instability emanates from two primary directions, due to design.

Radial -My top magnet must remain radially within the confines of the surface area of the fixed bottom magnet. This constrains the repulsive forces to being primarily between the magnet faces with no radial force segment. To achieve this, the fixed or lower magnets should always overhang the rotating magnets to maintain radial equilibrium in the event of excessive radial movement.

Circumferential -Instability is taking place as magnets pass each other, ameliorated by the multiple offset prime numbers of other magnet concentric arrays. The more of these will better smooth rotational motion and reduce starting torque and cogging.

This defines the elevating magnet effect and now a description of the drive arrangement which acts as a secondary mechanical damping to the magnetic "vibrations" is appropriate. PageS

Drive arrangement: The drive arrangement as well as damping has also to accomplish the following: The flat rectangular spokes from the flywheel to the centre shaft have to expand radially, Drawing 4, #15, as the flywheel runs up to its running speed. Due to the elevating magnets being in a frustrated cone they also have to accommodate vertical movement, #36 as well, as radial expansion takes place. l'his vertical movement will be a function of the cone angle, #16, which in turn is a function of the centrifugal forces acting on the top magnet support arrays and magnet repulsion horizontal forces opposing these centrifugal forces. See also Drawing 2.

It is proposed that these spokes shall be designed like single leaf springs, #20, to give vertical movement and with a concertina section, #19, to give radial expansion and some vertical expansion as well, with damping. It is also important that each spoke is carefully manufactured, probably, but not necessarily, in carbon fibre / epoxy composite to precise dimensions and physical properties such that all will act identically and have the same mechanical properties and mass. Wobble and horizontal displacements are damped out by effect of the spokes at 180 degrees. As one is say stretched the opposite is compressed to restore flywheel in a stable fashion around the spindle. Number of spokes can be a non prime number since system runs in a vacuum.

The spokes are firmly fixed to 1111) of flywheel flanges #56 and #57 by special lightweight bolts through flange holes #23, and to the shaft coupler, #21. For increased stability there are two sections of spokes, one near top and one near bottom of flywheel ID. The hollow shaft coupler, #21, having a central rubber damping section, #22, which is attached to the hollow shaft, #17, gives further damping to lateral vibrations on the flywheel, as well as assisting the spoke leaf spring #19 & 20, properties in the vertical plane. This vertical damping can only be affected if there is a thrust bearing on the shaft assembly.

Since the vertical deflection, which is proportional to the radial expansion and thus subject to an exponential law, then there may be an advantage in preloading the leaf spring at lower speeds and having least load (Thrust bearing load.) at higher speeds.

This will assist in bearing life.

The hollow rotating drive shaft, #17 is mounted, by means of bearings, on a fixed spindle, #55, which is fixed both at foundation level, #38, and at the top to the containment / vacuum chamber, #37. The key factor in the design of the spindle is the loading from the spokes and appropriately sized ceramic bearings to cope with mechanical loads and most importantly rotational speed.

Alternatively only ceramic roller bearings could be used which will give no thrust capability and float vertically. Magnetic thrust capability may be possible to design into hermetically sealed drive transmission system.

The fixed spindle, #55, being long and slender to accommodate the bearing sizes suitable, may need to be stiffer, it is affixed to the lid of the vacuum / containment chamber and will be subjected to buckling from end loading under vacuum and side loads imposed from flywheel. It thus may be necessary to taper upwards the lower set of spokes, #24. A fixed shaft system, which has a larger diameter central section and a means to fit a smaller bearing housing to the base is also possible.

Page 9 Manufacturing of magnet arrays: Drawing 5 shows the fixed elevating magnet bearing array segment in plan and section. #26 is precision aligned, centralised, levelled / shimmed and grouted to foundation #28 and bolted, #25 via 3 or more extensions, #27, part of #26 into the foundation. #25 should be Rotabolts or similar, so that initial bolt stretching is easily achieved and easily checked at maintenance intervals.

The magnet arrays are made up of segments. Due to the large prime numbers of magnets used, #30, there is no need for these to be a prime number of segments. The number should be such that the frustrated cone surface is formed from a series of flats.

The magnets and their magnetic surrounds are set in machined sockets and glued I fastened in place. The sockets will vary in depth to maintain a near true cone frustrum with the exception only of the flats across the magnets themselves. For example #34 will be a pocket deeper than pocket #35. This assists in keeping air gaps unifonn.

These sockets being a prime number means that each segment, composed of #26 & #29 has to be uniquely identified, machined and located. Segments should abut so that magnets can abridge, #32, between the segments and maintaining uniform spacing to match the prime numbers of magnets chosen. (This will necessitate careful machine tool programming and tracking.) Whilst the pockets are illustrated in a straight line across the segment, they will actually follow a curved path, #33, based on centre of rotation.

To prevent the possibility of eddy currents, the segmented arrays should be made of non-conducting material, at least in the area around the magnets. The magnets could be mounted on a machined fibre I epoxy composite, #29, or other material blends of a similar nature such as dough moulding compound. For machining purposes, a fibre such as Kelvar should be used. This could form the complete mounting or perhaps only the top of the mounting, #29 to insulate the conductive part magnetically. The purpose is to eliminate eddy currents and maximise system efficiency. #26 could even be cast iron or other heavier metal casting or fabrication since it is stationary.

Any combination of materials, #26 & 29, would have to be carefully joined for a permanent and robust solution.

Drawing 6 shows how the rotating, #39, and fixed, #26, magnet arrays abut each other in vertical plane. #26 is shown in dotted outline only.

#39 would preferably be aluminium or other lightweight metal, to minimise the centrifugal forces to be constrained by the rotating magnet arrays. This too has an insulating surface section, #41, to eliminate eddy currents. The magnets, #46, being either square or round are more widely spaced apart than the rectangular magnets, #30 in Drawing 5 to permit the centrifugal radial expansion. Magnets #30 and #46 by repulsive magnetic forces assist in the retention of the magnets in their pockets.

#39 and #46 could be manufactured in a uniform non electrically conductive body to maintain similar thermal expansion properties with flywheel main body, #18. Slight concentric curved protrusions or indents #47, on the interface between magnet array, #29 and flywheel, #18, will also assist in keeping that surface bonded in operation.

The flywheel, #18, is wound on site to inner spool, #44. Magnet array #39 is fixed to spool, #44, via a flange #42 on #39 and for example a bolted assembly, #43. Further enhancements against centrifugal force can be added in the form of reinforcements, #49.

Page 10 At zero speed, the magnetic forces still thrust radially inward on the top magnet arrays and must be constrained by fastening arrays circumferentially, or abutting each array to arrays on either side or by the constraining effect spool / mandrel, where fastening takes place and assisted by the curved protrusions or indents #47.

The drive spokes, #19 & #20 are bolted / fastened at location #23.

Spool #44. can be reinforced by ribs, #40, which can be tapered as shown or straight connecting the spool, #40, top and bottom flanges, #56 & #57.

The mechanical properties of #18, the flywheel and #39, the rotating magnet array, should be similar to minimise their potential separation.

Drawing 7 illustrates that each magnet, #58, has to be inserted into a tray, #52, made of electrical laminated steel or other magnetic material, such as iron powder, resin and fibre reinforcement in a cast pocket or from an open top box made of magnetic slot wedge material or any other soft magnetic material such as manganese zinc ferrite or nickel zinc ferrite. This simply provides the flux path, #53 a return route so that magnet operates at optimum performance. The size and cross section of #52 will have to be below a chosen flux density for this purpose. Tray, #52, may be laminated or moulded / cast, in iron powder / epoxy to prevent eddy currents or other soft magnetic material and resin.

Claims (22)

1. Bearingjtnd Drive Sy$em for large flisrwheel Cairns: 1. The large mass of the flywheel is suspended in air, vacuum or suitable gas by a prime number of concentric rows of large prime numbers of permanent magnets in repulsive mode, arranged in a segmented frustrated cone, giving pseudo-levitation, centralising properties and containing the rotating top array of permanent magnets against the forces of centrifugal motion, all the while keeping any frictional losses to a minimum, vertical bounce and wobble to a minimum by the use of flexible spokes attached to a central hollow drive shaft by means of a circular elastomer damped drive flange and in turn this central hollow shaft drives a magnet coupling / gearbox, which provides a hermetical seal against vacuum leakage, to external shaft.
2. 2. As in Claim 1, the angle of the frustrated cone is determined by the centralising forces needed and the resistance needed against centrifugal forces acting upon the top magnet array.
3. 3. As in Claim 1, the spokes are manufactured preferably, but not exclusively, in carbon fibre / epoxy, being capable of stretching vertically and radially to adjust to expansion of flywheel due to rotational forces as it rises up the frustrated cone ramp. They also act as motion dampers to assist in stabilising the flywheel.
4. 4. As in Claim 3, the spokes are fastened to a flange, which is fastened to the hollow shaft and this flange has within, an elastomer element, which assists with the damping.
5. 5. As in Claim 3, the spokes may have a preload imposed on them whilst stationary, so that as flywheel speeds up and rises up the ramp, the eventual thrust load on the spindle bearing is reduced at full speed, increasing bearing life.
6. 6. As in Claim 3, there may be one or two sets of spokes, arranged horizontally, or angled to facilitate the centre shaft and bearing arrangement.
7. 7. The magnets as in Claim 1 are arranged in a prime number of concentric circles.
8. 8. As in Claim 7, the greater the prime number of concentric circles, the more stable the flywheel becomes.
9. 9. The start positions of these magnet circles in Claim 7 are offset from each other by 360 Degrees divided by the prime number of magnet circles in total to prevent any notching effect and lower "cogging" torque to almost zero.
10. 10. By carrying out the actions in Claim 9, the vertical bounce effect is smoothed out.
11. 11. As in Claim 3, when the flywheel expands under centrifugal forces, the lower fixed magnets must be rectangular or elongated with semi circle ends, such that the upper round or square magnets are always in full contact with same repulsive forces, from zero to full speed.
12. 12. As in Claim 11, the upper rotating magnets must always be radially within the lower fixed magnets, inner and outer diameters, otherwise the repulsive forces will become unstable.
13. 13. As in Claim 11, the upper rotating magnets can be round so that engagement of repulsive forces is smoother and less cogging.
14. 14. As in Claim 1, whilst the magnet support arrays are flat on top, the magnet pocket depths differ across the flat, being deeper in the centre, so as to maintain the frustrated cone outline better and the effective flats only occur across the magnet widths.
15. 15. As in Claim 1, the magnet assemblies may bridge between the segments of the frustrated cone.
16. 16. As in Claim 1, the ratio of outer diameter to height has an important influence on stability. It preferably should be much less than 1:1.
17. 17. As in Claim 1, the bearing designs, magnetic and mechanical, the effect of precession torque from the earth's rotation must be taken into account. It may be advantageous to design centre fixed spindle to auto correct the effects of precession torque and keep magnetic air gaps constant.
18. 18. As in Claim 1, the magnet arrays can be wholly or partially manufactured from a non conducting material around the magnet I tray assemblies to eliminate eddy currents and improve efficiency.
19. 19. As in Claim 18, these trays can be made from an electrical steel laminated assembly, a moulded / resin component with high iron powder content or with other soft magnetic powder or from solid magnetic slot wedge material.
20. 20. As in Claim!, the final drive takes place by a magnetic coupling and or magnetic gearbox if' a ratio change is desirable. This drive is incorporated into the lid of the flywheel containment / vacuum housing and provides a hermetic seal.
21. 21. As in Claim 20, the final drive arrangement may be magnetically, axial or radial in layout. Radial is the preferred embodiment to best resist vacuum induced stresses.
22. 22. As per Claim 1, a bearing design which does not need lubrication and can run in a vacuum.