GB2517523A

GB2517523A - Flywheel control scheme

Info

Publication number: GB2517523A
Application number: GB1401069.8A
Authority: GB
Inventors: Rick Townend
Original assignee: GKN Hybrid Power Ltd
Current assignee: GKN Hybrid Power Ltd
Priority date: 2013-07-19
Filing date: 2014-01-22
Publication date: 2015-02-25
Anticipated expiration: 2034-01-22
Also published as: BR112016017119A2; GB201401074D0; GB2517524B; CN106464076B; GB2517524A; JP2017506488A; GB201401070D0; GB2517523B; CN106464076A; GB2517808A; GB201401069D0; GB2517808B

Abstract

A control system for a flywheel assembly 12, comprising a rotor 19, includes a plurality of sensors 13 each arranged to detect rotational motion of the rotor and to provide a motion signal indicative of the rotational motion. A controller 14 is arranged to compare each motion signal to an expected signal; and if any of the motion signals are not equal to the expected signal, the controller arranged to disable the flywheel to prevent flywheel acceleration. The rotor may be a permanent magnet rotor. The flywheel assembly may be driven by inverter 11. The expected signal may be a square wave signal. The sensors may be optical sensors detecting dark and light sections of the flywheel. The flywheel rotor and sensors may be positioned within a vacuum. The sensors may detect dust particles from the rotor blinding at least one of the sensors.

Description

Flywheel Control Scheme This invention relates to a flywheel control scheme. It is particularly suitable for, but by no means limited to, kinetic energy recovery systems and methods for both mobile applications such as vehicles, ships and trains etc as well as static energy recovery systems such as railways, power smoothing, wind power systems and off grid applications.

Background

Flywheels are well known devices, used for storing energy in a rotating mass. The amount of energy stored in a flywheel is proportional to the square of its rotational speed. In general, energy is transferred to a flywheel for storage by the application of a torque to the flywheel, causing its rotational speed to increase. Conversely, energy can be released or recovered from a flywheel by the flywheel applying a torque to a load, as a result of which the flywheel's rotational speed decreases.

Many known flywheel assemblies include an electrical machine which can function either as a motor or a generator. When the electrical machine acts as a motor (i.e. when the flywheel assembly is in "motoring" or "recuperating" mode), electrical energy supplied to the machine is converted to kinetic energy and, as a result, the flywheel mass rotates more quickly. When the electrical machine acts as a generator (i.e. when the flywheel assembly is in "generating" or "boosting" mode), kinetic energy stored in the flywheel mass is converted to electrical energy and can be supplied onwards to another component with a system such as an electric motor.

In practice, the speed at which a flywheel mass can rotate, and hence the amount of energy which the corresponding flywheel assembly can store, will depend at least in part on the mechanical strength and strain capabilities of the flywheel assembly. For example, an important factor is its behaviour in response to the mechanical stresses which are experienced at high rotational speeds.

When a flywheel assembly comprises an electrical machine, another factor in how fast the flywheel mass can rotate is the manner in which the electrical machine can be controlled. It is desirable to implement a reliable and efficient control scheme, which works at a range of flywheel rotational speeds. It is also desirable to avoid and/or reduce losses, for example losses due to heat dissipation, as much as possible. And it is important that the flywheel assembly is as safe as possible, including being safe in the event that the flywheel assembly experiences a failure or breakage.

According to known methods, the manner in which the rotating parts of a flywheel assembly are to be controlled can have a significant influence on the manner in which the physical components of the flywheel assembly are constructed. For example, the magnetic rotating pad(s) of the electrical machine can be shaped in a particular manner in order to determine their magnetic properties, in preparation for application of electrical pulses to the flywheel control system when motoring, or the generation of current when generating.

It is preferable to form flywheels from materials which are of low density, to reduce the stresses on the wheel, and which are extremely strong. It has been found that composite materials including fibre filaments are particularly suitable for use in flywheel fabrication. To be of use in, for example, hybrid vehicles or uninterruptable power supplies, energy storage flywheels need to operate at extremely high speeds, greater than 10,000 rev/mm or even more than 50,000 rev/mm. Accordingly, there is considerable demand for composite flywheels that are able to operate at speeds of these orders.

Flywheels need to operate safely within specified parameters such as operating speeds, temperatures, loadings, impact forces and vibration.

Known flywheels running at rotational speeds in excess of their safe operating ranges may fly apart, with small and large parts of the flywheel being ejected at high speed and energy from the remnants of the rotating mass. The rotating remnants may undergo catastrophic failure such as collision with the housing oi break up of the support shafts and bearings. Such catastrophic failures could be extremely dangerous and result in significant damage. Such a failure is often referred to as a burst failure mode.

Accordingly, there is a need for flywheel assemblies exhibiting increased efficiency of energy transfer and increased safety whereby a catastrophic failure can be mitigated with minimum damage to surrounding systems.

GB 1312924.2 filed l9 July 2013 and GB1312927.5 filed 19th July 2013 are hereby incorporated by reference in their entirety.

Summary

According to a first aspect there is provided a method of controlling a flywheel as defined in Claim 1 of the appended claims.

Thus there is provided a method of controlling a flywheel comprising a rotor, the method comprising the steps of detecting rotational motion of the rotor at a plurality of sensors, providing, from each sensor, a motion signal indicative of the rotational motion, comparing each motion signal to an expected signal and if any of the motion signals are not equal to the expected signal, disabling the flywheel to prevent flywheel acceleration.

Optionally, the method wherein the expected signal comprises a square waveform.

Optionally, the method wherein the sensors comprise optical sensors.

Optionally, the method wherein the plurality of sensors detect dark and light sections of the rotor of the flywheel.

Optionally, the method further comprising, based on one or more motion signals, deriving a rotational speed of the rotor, and if the rotational speed is above a predetermined threshold, disabling the flywheel to prevent flywheel acceleration.

Optionally, the method wherein disabling the flywheel to prevent flywheel acceleration comprises disabling switching signals of an inverter of the flywheel.

Optionally, the method wherein the predetermined threshold is determined by stresses and strains that the flywheel assembly undergoes when in motion.

Optionally, the method wherein the sensors are arranged to detect dust particles from the rotor blinding at least one of the sensors.

According to a second aspect there is provided a flywheel control system as defined in claim 8. Accordingly there is provided a flywheel control system comprising a rotor, a plurality of sensors each arranged to detect rotational motion of the rotor and to provide a motion signal indicative of the rotational motion, a controller arranged to compare each motion signal to an expected signal, and if any of the motion signals are not equal to the expected signal, the controller arranged to disable the flywheel to prevent flywheel acceleration.

Optionally, the flywheel control system wherein the flywheel rotor and the sensors are arranged to be positioned within a vacuum.

Optionally, the system wherein the sensors are positioned at 200 intervals around an axis of rotation of the rotor.

Optionally, the system wherein the magnetic rotor comprises N-S pole pairs.

Optionally, the system wherein the N-S pole pairs are positioned at 300 intervals around an axis of rotation of the rotor.

Optionally, the system wherein the sensors comprise optical sensors.

Optionally, the system comprising 3 sensors.

Optionally, the flywheel control system is further arranged to carry out any method as disclosed herein.

In another aspect there is provided an apparatus comprising a processor arranged to perform the method as defined in claim 16.

In another aspect there is provided a computer program comprising computer program code that is executable in use to perform the method as defined in claim 17.

In another aspect there is provided a computer readable medium comprising computer program code that is executable in use to perform the method as defined in claim 18.

With all the aspects, preferable and optional features are defined in the dependent claims.

Brief Description of the Drawings

Embodiments will now be described, by way of example only, and with reference to the drawings in which: Figure 1 illustrates a system for controlling a flywheel comprising an inverter in a triple half H-bridge arrangement according to an embodiment; Figure 2 illustrates a flywheel and sensors according to an embodiment; Figure 3 illustrates an inverter of the flywheel system according to an embodiment; Figure 4 illustrates inverter switching waveforms for flywheel acceleration mode (motoring) in a first and second rpm range of operation; Figure 5 illustrates inverter switching waveforms for flywheel acceleration mode (motoring) in a third rpm range of operation; Figure 6 illustrates inverter switching waveforms for flywheel acceleration mode (motoring) in a fourth rpm range of operation; Figure 7 illustrates inverter switching waveforms for flywheel acceleration mode (motoring) in a normal rpm range of operation; Figure 8 illustrates inverter switching waveforms for flywheel deceleration mode (generating) in a normal rpm range of operation; Figure 9 illustrates sensor output waveforms; Figure 10 illustrates sensors and the controller of the flywheel system in detail; Figure 11 illustrates an arrangement of flywheel stator teeth winding; Figure 12 illustrates the relationship between mechanical rotation and electrical rotation on the flywheel rotor.

Figure 13 illustrates a method according to an embodiment; l0 Figure 14 illustrates method steps that may be performed in each of CPLD 232, 242, 252 in conjunction with CPLD 260 in relation to fail-safe shutdown; and Figure 15 illustrates method steps that may be performed in each of CPLD 232, 242, 252 in conjunction with CPLD 260 in relation to overspeed protection.

In the figures, like elements are indicated by like reference numerals throughout.

Overview A simplified scheme for controlling initialisation, and both generating and recovering modes of a permanent magnet flywheel is provided. The flywheel is a high-speed, high-power system that operates with modest duty cycle. Appropriate rating of an inverter is a non-trivial task because the rating of most commercial inverters is given for a continuous duty cycle. As a result, there is a tendency in known systems to over specify the flywheel inverter (11) rating in preliminary analyses which results in a more expensive system than is necessary.

As described in co-pending application GB1312924.2 filed on 19th July 2013, the make-up of the multi-layer composite structure of the rotor 19 comprises an outer portion 26 having carbon tow in resin only, and an inner annulus 20 having a multi-layer composite (MLC) of glass fiber tow with resin and dispersed magnetic particles.

As a result of this make-up of the multi-layer composite structure of the rotor, the energy lost in the rotor due to the production of eddy currents as the magnetic field of the rotor rotates past a coil of the stator 18 is negligible compared to a standard solid permanent magnet rotor of a conventional metal construction. This is because in known non-MLC systems, glue holds the magnets together that form the magnetic rotor. The structural joints of such rotors cause eddy currents to be induced as the magnetic flux changes with rotation as would be understood. These eddy currents in turn cause heat build-up and other lossy effects. With the present rotor 19, the magnetic particles of the inner annulus 20 are so small (less than the critical flaw size as associated with Griffith's Crack Theory) that the (effective) individual tiny magnets are insulated from one another by the multi-layer composite and hence no eddy currents are formed. Therefore the magnetic particles that form the permanent magnet of the inner annulus 20 do not heat up. This has the benefit of negating the heating effect of non-MLC systems.

This has the added benefit of less heat being produced in the system as a whole. As with other systems, the flywheel assembly is positioned in a vacuum 15 where the only heat escape mechanism is by way of radiation from the rotating mass to the casing of the vacuum 15 which contains the flywheel system 12 and sensors 13 (see Figure 1). Therefore, a system that produces less heat is beneficial as it enables a more simple control algorithm that need not be as concerned with heat build-up when storing and harvesting energy to/from the flywheel assembly.

Further, in known systems, three-phase sinusoidal current waveforms necessary for efficient power transfer to/from the flywheel are shaped by way of PWM of the switching signals of the inverter to cater for energy lost in the permanent solid magnets of those systems (to reduce heat build-up in the stator) and also to control heat build-up in the switches Si to S6 owing to the current being conducted therethrough. With the present system, this sinusoidal shaping is not required.

Schemes 5 and 6 at least, are only concerned with the timing of pulses for optimum torque transfer to the flywheel.

As the control algorithm is not required to perform waveform shaping to prevent eddy current build-up in the rotor, the algorithm may be a lot more simple comprising PWM pulses that are only tailored to limit the current that is conducted through the switches. Compared to known systems, the switching losses are reduced as the switches Si to S6 are physically switched a reduced number of times per inverter state (see Figures 7 and 8 for switching waveforms of normal operating rpm range) and therefore heat up less which also aids a reduced overall heat build-up. Further efficiencies are obtained as the electronics are not driven so hard therefore requiring less power.

The benefits of the system as a whole include negligible heating effect which in turn benefits the supporting electronics systems as heating of electronics may cause additional timing issues and reduced lifetime. Further, less processing power is required to calculate the waveforms involved and less cooling is required as the heat build-up is inherently reduced. The supporting electronics are less of a leach on the system which improves efficiency of the system as a whole.

As will be shown, the switching duty cycle can be controlled in order to maximise current flow through the switches without compromising switch performance or maximum rating. This has the additional benefit that the switches are able to conduct for longer and hence more torque can be imparted onto the flywheel and hence a faster increase in speed (acceleration) of rotation is obtainable. This leads to more energy being transferred into and out of the flywheel over a given period of time, and more energy being recoverable (greater utilisation of the flywheel).

Further, the use of optical sensors for the sensing of position and correct operation allows a fail-safe operation as will be explained in detail. This is because when dust internal to the vacuum obscures the sensor in any way there is an assumed failure situation and the system can be shut down as a precaution to avoid catastrophic failure.

Detailed Description

Figure 1 illustrates a system for controlling a flywheel comprising an inverter 11 in a triple half H-bridge arrangement.

The three phases of the inverter (1, 2, 3) are coupled to flywheel assembly 12 for the transfer of electrical energy into kinetic energy of a spinning rotor of the flywheel when in motoring mode, (accelerating the flywheel to store energy) and for the transfer of kinetic energy from the spinning rotor into electrical energy when in generating mode.

Flywheel assembly 12 comprises an 18 tooth stator 18 surrounded by a 12-pole permanent magnet rotor 19. The rotor is able to rotate around the stator. Preferably, the rotor and stator are coaxial and are annular in shape.

Other arrangements of stator and rotor could also be used. For example a 2, 4, 6, 8, or 12-pole rotor (1, 2, 3, 4, 5, 6 pole pairs). A greater number could be used if desired. For the number of poles, n, the formula for the number of teeth, t on the stator can be expressed as is t=n*1.5 i.e. 2 magnetic pole pairs = 4 pole pieces = 6 I 0 tooth stator.

For an increased number of pole pairs, greater transfer of energy to/from the flywheel is achieved but with increased complexity of the system as a whole. The limiting factor is the diameter of the rotor as the magnetic rotor pieces (the inner annulus magnetized MLC) should be a minimum size in order to meet manufacturing constraints. The current of the electromagnets used to magnetise the inner annulus MLC must be sufficient to create the permanent magnet in the MLC. This is determined by the minimum copper conductor cross sectional area, which results in the windings of the electromagnets being constrained by the capacity of copper to conduct as would be understood.

The magnetising device is described in more detail in another GB patent application being filed today in the name of Williams Hybrid Power Limited and so will not be discussed further herein.

DC link capacitor 10 provides voltage smoothing between the AC flywheel system (11, 12, 13, 14) and the DC system that supplies/takes electrical energy to/from flywheel assembly 12. The DC link capacitor 10 may be between lOUpE to 1 mE, and nominally 700 pP +/-10%.

Sensors 13 are positioned in a manner so as to be able to detect position of the rotor as will be explained herein below. Sensors 13 may be optical sensors, for example and not limited to the visible spectrum or the infra-red spectrum. Each sensor may comprise a source of light (for example an LED) and a sensor for sensing light originating from the source that has reflected off passing segments of the flywheel rotor 19.

Controller 14 is arranged to provide signals (si, s2, s3, s4, s5, s6) by way of optional gate drive module 9 to operate the switches of the inverter so that suitable waveforms and current flow are provided to coils of the stator 18 to impart torque and hence rotational motion to the rotor when motoring. The controller is also arranged to allow the conversion of rotational motion to electrical energy when generating.

Signals from the optical sensors may be provided to controller 14 to determine the speed and position of the rotor and also to determine correct operation of the flywheel apparatus.

Controller 14 may comprise a memory device 16 and processor 17 associated therewith. Memory device 16 may comprise one or more look-up tables of data associated with the controller for determining switch firing times, duty-cycles, phase-shifts in relation to flywheel speed and/or position and/or other variables.

A means 7 of providing a signal 6 indicative of current and/or voltage in the stator coils (typically standard PID control) is provided to controller 14. The flywheel assembly may be voltage controlled or current controlled. Voltage control has the advantage that it can be measured external to the machine, for example on the output (see connections 4) or at any other suitable point, whereas current control is measured in the stator (see connections 5 shown here on each of the inverter phases for convenience, for example current loop sensors) which is more difficult, however current control may additionally protect the stator windings and switches (see figure 3) of the inverter from overload.

Flywheel assembly 12 is shown and described in co-pending GB patent application number GB1312927.5 filed on 1gth July 2013, the entirety of which is incorporated by reference herein. Flywheel assembly 12 and associated components as shown in Figure 1 operate in the same manner as a permanent magnet synchronous motor comprising a rotor 19 and a stator 18. The rotor 19 of Flywheel assembly 12 comprises an inner annulus and an outer portion. The inner annulus is permanently magnetised. The stator 18 of flywheel assembly 12 (which is surrounded by rotor 19) comprises coils in a star formation as would be understood, however other coil formations such as delta or centre tap star could also be used. In general, the coils of each phase are arranged around the 18 teeth of the stator 18 as shown in Figure 11. As can be seen, the teeth are wound in a repeating pattern whereby every fourth tooth is wound by the same coil and each coil is wound around 6 teeth (for an 18 tooth stator). Each tooth may be wound a plurality of times.

In an embodiment, flywheel assembly 12 is contained within a cavity or vessel comprising a vacuum 15 to reduce air resistance and hence provide a more efficient system. Sensors 13 are also positioned within vacuum 15 as detection of flywheel rotation requires such an arrangement, and further to provide fail-safe operation as described herein. In other embodiments, the sensors may be positioned outside the vacuum however, they must be close enough for successful detection of flywheel rotation.

Once magnetised, the inner annulus 20 as illustrated in Figure 2 can be incorporated into the flywheel assembly 12 along with the outer portion 26. According to an embodiment, before being incorporated into the flywheel assembly 12, the inner annulus 20 and/or the outer portion 26 of the rotor can be painted. For example, the N and S poles 21, 22 or areas of the rotor in alignment with the poles may be painted black & white respectively or vice versa (27). Preferably, the outer portion 26 is painted as described. Optionally the areas may be painted dark and light.

According to an embodiment, six N-S pole pairs (21, 22) are formed in the inner annulus. Physically, the poles 21, 22 should be substantially equally sized.

Therefore each pole 21, 22 occupies an arc of approximately 30° (mechanical) around the inner annulus. The stator 18 of flywheel 12 may comprise 18 teeth (28) to match the six pole pairs and three phases as illustrated in Figure 2.

Sensors 13, preferably comprising three optical sensors 23, 24, 25 are positioned, for example, at 200 (mechanical) intervals around the annulus centre point as shown in Figure 2. The sensors, therefore, are offset from the N-S pole pairs and are positioned at an electrical separation of 120°. Figure 12 shows the relationship between mechanical rotation and electrical rotation on the flywheel rotor. With a 3-phase, 18 pole architecture, it only requires 60 mechanical degrees of rotation (the rotational distance between the three sensors) to complete 360 degrees of electrical rotation, that is to say that 1 mechanical revolution equals 6 electrical revolutions.

The visual distinction between the N and S poles and the offset positioning of sensors 13 (23, 24, 25) can be used, for example, to enable the sensors to detect and track rotation of the inner annulus 20, this data, in turn, may be used to monitor position and speed of the flywheel assembly. This data is provided to controller 14 for use in providing switch firing signals (Si, 52, 53, 54, 55, 56) to the inverter ii.

In other embodiments. fewer sensors may be used, or even no sensors which control occurring from measured outputs of the statorwindings (1,2, 3). In a preferred embodiment, three sensors are used for redundancy as would be understood.

Turning to inverter ii, which is illustrated in more detail in Figure 3 and comprises six power electronic switches which may comprise, for example, IGBTs or MOSFETs (51, 52, 53, 54, 55, 56) arranged as a triple half H-bridge between voltage rails Vt and V-. Phase signals 31, 32 and 33 are taken intermediate each switch pair as would be understood and are coupled to the 3 phases of the stator respectively (see 1,2, 3of figure 11 and of figure 1).

Coupled in parallel with each respective switch is an associated reverse biased diode (Dl, D2, 03, D4, 05, D6) to suppress back EMF as would also be understood.

The DC Link Capacitor 10 also forms part of a boost converter which also comprises the inductance of the windings and a respective switch of the inverter as would be understood.

The switching schemes used to generate the gate signals within the inverter are non-conventional. The unique electro-mechanical characteristics of the flywheel assembly 12 allow computationally simple switching schemes to be used. This provides reduced switching losses within the inverter and less processor power which aids the desire to provide a system of increased efficiency.

In an embodiment, the flywheel assembly 12 can spin to a nominal mechanical (rotational) speed of 36,000 rpm (600 rev/s) but this could be as high as 50,000rpm.

The 12 pole (six pole-pair) inner annulus (of rotor 19) and stator 18 operate in the same manner as a permanent magnet AC (PMAC) machine with a nominal electrical frequency of 3600 Hz (at 36,000 rpm = 600 rev/s x 6 pole-pairs).

The rotational speed of the flywheel rotor 19 and whether the flywheel is accelerating or decelerating determines the switching scheme that is applied to the inverter switches to control the PMAC machine. The following table details the switching scheme: Typical Speed Range Acceleration! Switching Switching (RPM) ____________ Deceleration Scheme Scheme Name From: To: Number 0 500 Acceleration 1 PWM Start 500 4000 Acceleration 2 PWM 4000 8000 Acceleration 3 8 Pulse PCM 8000 15000 Acceleration 4 4 Pulse PCM 15000 36000 Acceleration 5 2 Pulse PCM 36000 14000 Deceleration 6 1 Pulse PCM 0 36000 Neither 7 Freewheel Switching schemes 1 to 4 are typically used to accelerate the flywheel to nominal minimum operating speed of 15000rpm. Once this speed has been reached schemes ito 4 are not used further while the flywheel is operating above 15000rpm.

Motoring and generating takes place above 15000rpm. In other embodiments, the nominal operating speed may be higher or lower than 15000rpm. In these other embodiments, schemes equivalent of switching schemes 1 to 7 may be operational IS at different rpm ranges as would be predetermined. The nominal operating speed is preferably outside of the mechanical resonance of the flywheel assembly. In the following discussion of switching schemes, embodiments may make use of any or all of the switching schemes at different rpm ranges as determined for the flywheel assembly in use and in any order as determined by the properties of the flywheel assembly in use. Each flywheel assembly may be individually tested and tuned in order that frequency ranges and switch timing are optimised for torque transfer to/from the flywheel assembly.

In general, there are two conflicting and opposing effects which have to be balanced out by the switching schemes -back EMF and impedance in the rotating flywheel system. At low rotational speed of the flywheel (low back emf), the current in the coils of the stator is not limited by the impedance of the coils, therefore a PWM scheme of switching is utilised in order to limit current and avoid over-current (burning out of the coils) or heating effects in the switches of the inverter and coils.

As rotational speed of the flywheel increases, impedance (back emf) increases, and the number of switching events may be reduced to reduce parasitic losses and heating of the electronics.

Figure 13 shows a method comprising steps as described herein. The ordering of the steps may be changed as desired to cater for different properties of the flywheel in use.

Switching Scheme 1 (PWM Start) PWM Start is used to move a rotor 19 from a rest position. The three sensors 23, 24, provide signals to controller 14 according to the painted black and white segments of the annulus 20 (of rotor 19) of flywheel 12. For example, a high signal represents a black segment, and a low signal represents a white segment, or vice versa. The sensors may be active logic low, and present an open collector style output.

In this scheme, and at step 130 of Figure 13, processor 17 of controller 14 employs a polling algorithm to read the state of the three sensors and applies gate drive signals (si to s6) according to the table below. Processor 17 may directly output drive signals or may provide suitable signals to a PLD device 260 that in turn, provides the gate drive signals via gate drive module 9 (see Figure 10). In the table, a 1' signifies the respective switch is closed, and a 0' signifies that the respective switch is open.

The inverter state is cycled from A to F in sequence dependent on the signals from the three sensors. It takes a 60° rotation of the inner annulus 20 of the flywheel for states A to F to be cycled through in turn.

The below commutation algorithm, for example, may be provided in a look-up table which may reside in associated memory 17: Inverter Position Sensor ________ Gate drive signal _____ _____ _____ State 23 24 25 Si S2 53 S4 S5 S6 A H L L 1 0 0 0 0 1 B H H L 1 1 0 0 0 0 C L H L 0 1 1 0 0 0 D L H H 0 0 1 1 0 0 E L L H 0 0 0 1 1 0 F H L H 0 0 0 0 1 1 As shown in Figure 4, to regulate the current that is switched by switches Si to 56, the high side gate drive signals (Si, S3 and S5) are logically ANDed with an asynchronous PWM signal of frequency 8 or 10kHz and variable duty cycle by controller 14 (shown in figure 4 as dark areas). The duty cycle of the PWM signal may be selected from a lookup table (possibly contained in associated memory 17) of duty cycle vs. flywheel speed. A signal indicative of flywheel speed is provided by the output of one of the sensors. Optionally more than one of the sensors can be used for redundancy. For example, flywheel start up requires power from the associated machine, for example the vehicle within which the flywheel is installed, or some other external source, Often, such power is limited owing to how fast or hard (or heavy) the vehicle is breaking to use regenerative breaking power. Therefore, the duty cycle is changed to use as much power as is available to provide optimal torque transfer to the flywheel for start up, limited by the impedance of the coils of the stator 18 and the current limit of the switches of inverter ii. Preferably, low side gate drive signals (52, 54, S6) are held constant within each inverter state (A-F) to minimise switching losses in the inverter ii.

As shown in Figure 4, the electrical angle refers to the cycling of the inverter states as rotation of the flywheel annulus 20 occurs. A theoretical 3600 electrical rotation takes 6 inverter states (600 of actual mechanical rotation).

Switching Scheme 2 (PWM) When the rotor 19 is rotating at low speed (approximately 500 rpm in one embodiment), a transition is typically made to switching scheme 2 "PWM" at step 132 of Figure 13. In this scheme, gate drive signals Si to S6 remain the same as switching scheme i as shown in Figure 4 (logically ANDed with an asynchronous PWM signal of frequency 8 or 10kHz on high side gate drive signals, constant on for low side gate drive signals). However, the inverter state is determined in a different manner.

The inverter states remain commutating in a cyclical manner as for scheme 1 PWM start, however in this scheme, a transition from one inverter state to the next is based on a timing signal.

The timing signal is derived from a single sensor input (optical sensor 23) and is 6 times the electrical frequency of the rotor of the flywheel (which itself is 6 times the mechanical rotational frequency as discussed previously) and is synchronised to the flywheel. Within each electrical cycle all six inverter states are stepped through as shown.

To synchronise the timing signal to the flywheel motion, the timing signal may be phase-lock looped to the single sensor input with a frequency multiplication factor of 6.

Throughout the switching schemes 2 to 6, the timing of the gate drive signals Si to 36 is matched to the rotational speed and position of the rotor 19 in order to minimise the current required for maximum torque transfer to the flywheel by aligning the relative positions of the magnetic field induced in the coils of the stator and the field of the permanent magnet rotor. To achieve alignment, a variable phase shift between Q0 and 3600 electrical degrees may be implemented between optical sensor input and timing signal which has the effect of shifting the electrical signals and hence when the consequential magnetic field is setup in the stator coils. This phase shift may be set within software executed by processor 17 and interpolated from a lookup table of phase shift vs. flywheel speed which may be stored within associated memory 16.

The further addition of a fixed phase shift also aids tuning of individual flywheel assemblies to mitigate for any mechanical inaccuracies in the positioning of the black and white painted segments on rotor 19, such as the segments not being of a consistent size, or not being aligned up correctly with the actual N-S poles of the permanent magnet rotor 19. Tuning may be achieved by placing each flywheel assembly in a test jig and measuring the magnetic fields when rotating at a nominal working speed (for example 15000-36000rpm of scheme 5 in one embodiment). The point at which the maximum magnetic pulse occurs as the rotor 19 rotates is found and applied to a software look-up table which may be stored in associated memory 16 or processor 17 for each individual flywheel assembly.

The duty cycle for the PWM may be selected from a lookup table of duty cycle vs. flywheel speed in the same manner as switching scheme 1.

Synchronised switching to flywheel position as described in relation to scheme 2 is implemented in schemes 2 to 6, and is not described again in relation to those schemes.

The variable and fixed phase shift is implemented as above in schemes 2 to 6, and is not described again in relation to those schemes.

The duty cycle is implemented as above in schemes 2 to 4 and is not, therefore, described again in relation to those schemes.

Switching Scheme 3 (8 Pulse PCM) When the rotor 19 is rotating at approximately 4000rpm, typically following the PWM switching scheme, the 8 Pulse PCM switching scheme is entered at step 134 of Figure 13. In this mode the low side gate drive signals are held constant within each inverter state (A-F), but high side gate signals are synchronously modulated to provide 8 pulses. Figure 5 illustrates the scheme 3 gate drive signals for a duty cycle of 50%.

The width of each pulse is calculated as: 1 1 Tpuise = d Xy where: Tpuise is the pulse width d is the duty cycle f is the electrical frequency Each pulse train (which comprises 8 pulses) is centre aligned.

For example, at lower speeds of rotation (schemes 2 to 4), the inductance in the coils of the stator 18 is not the over-riding consideration and hence to control the flow of current through switches Si to 36, a reduced number of pulses may be used. If switches Si to 36 were kept switched on for too long, to much current would flow through them which could exceed the maximum rated operating conditions for the switches. As the rotor speed increases (for example schemes 4 and 5) the number of switches (i.e. the number of switching pulses applied) may be limited to reduce the duty cycle of the switches as the back emf increases and the inductance increases which has the effect of reduced current flow therefore the switches can be maintained in a conductive position for longer without exceeding the current limit of either the switches or the cols of the stator 18. The aim is to limit the duty cycle of switches Si to S6 as much as possible and use the inductance of the PMAC machine to limit the current in the switches Si to S6 and the coil of the stator 18.

Typical switching speed of this scheme may be in the range 3.2 to 6.4 kHz.

Switching Scheme 4 (4 Pulse PCM) The 4 pulse PCM switching scheme is similar to the 8 pulse scheme. When the rotor 19 is rotating at approximately 8000rpm, typically following the 8 pulse scheme, the 4 Pulse PCM switching scheme is entered at step 136 of Figure 13.

In this mode the low side gate drive signals are held constant within each inverter state (A-F), but high side gate signals are synchronously modulated to provide 4 pulses. Figure 6 illustrates the scheme 4 gate drive signals for a duty cycle of 50%.

The width of each pulse is calculated as: 1 1 Tpuise = d Xy where: Tpuise is the pulse width ci is the duty cycle f is the electrical frequency Each pulse train (which consists of 4 pulses) is centre aligned. Typical switching speed of this scheme may be in the range 3.2 to 6 kHz.

Switching Scheme 5 (2 Pulse PCM) When the rotor 19 is rotating at approximately 15000rpm, typically following the 4 pulse PWM switching scheme, the 2 Pulse PCM swit-ching scheme is entered at step 138 of Figure 13.

The 2 Pulse PCM switching scheme is the scheme used to accelerate the flywheel (rotor) in its normal operational range (in one embodiment 15000 to 36000rpm).

For each inveiter state (A-F) both high and low side gate signals are synchronously modulated to provide 2 pulses. Figure 7 illustrates the scheme 5 gate drive signals for a duty cycle of 50%.The width of each pulse is calculated as: 1 1 d x where: Tputse is the pulse width d is the duty cycle f is the electrical frequency The duty cycle is controlled via a PID control loop (7) that regulates the DC link voltage or current 10 of the inverter 11 by way of signal 6 supplied to controller 14. If the set DC link voltage is higher than the actual DC link voltage the duty cycle is reduced (lower power transfer from DC link to flywheel 12). If the set DC link voltage is lower than the actual DC link voltage the duty cycle is increased (higher power transfer from DC link to flywheel 12). The duty cycle is set so as to limit at a pre- determined maximum output current through switches Si to 56 and a pre-determined maximum speed of the rotor 19. Typical switching speed of this scheme may be in the range 3 to 7.2 kHz.

Switching Scheme 6 (1 Pulse PCM) The 1 Pulse PCM switching scheme is the main scheme used to decelerate the flywheel in its normal operational range (in one embodiment 36000 to 15000rpm).

(step 140 of Figure 13).

For each inveiter state (A-F) both high and low side gate signals are synchronously modulated to provide 1 pulse. Figure 8 illustrates the scheme 6 gate drive signals for a duty cycle of 50%.The width of the pulse is calculated as: 1.

Tpuise = d where: Tpuise is the pulse width d is the duty cycle f is the electrical frequency Typical switching speed of this scheme may be in the range 3.6 down to 1.5 kHz.

The duty cycle is controlled via a PID control loop (7) that regulates the DC link voltage or current 10 of the inverter 11 by way of signal 6 supplied to controller 14. If the set DC link voltage is higher than the actual DC link voltage the duty cycle is increased, but with an additional 180° phase shift (which provides higher power transfer from the flywheel assembly 12 to the DC link 10). If the set DC link voltage is lower than the actual DC link voltage the duty cycle is reduced, but with an additional 180° phase shift (which provides lower power transfer from the flywheel assembly 12 to the DC link 10).

The duty cycle may be additionally limited at a pre-determined maximum output current through switches Si to S6 and a pre-determined maximum speed of rotor 19.

In this scheme, the switches are used to switch the windings of the stator 18 on' and off' as a form of a boost converter. With a respective switch turned on' (closed), the inductances formed by the stator windings charges as the permanent magnet rotor rotates past the coils of the stator. When that switch is then turned off' (open), the current built in the inductances flows through the reverse biased diodes, charging the DC link capacitor 10, which may be utilised for driving any DC load, like an electric

motor for example.

During normal operation, switching schemes 5 and 6 are typically cycled between according to whether the system is carrying out motoring (step 138) or generating (step 140).

Switching Scheme 7 (Freewheel) In the freewheel scheme, all gate drive signals are turned off (step 142). This scheme may be used, for example, when the system that in which the flywheel assembly is installed is in a state where neither motoring mode nor generating mode is required at that time. From the freewheel scheme, any of schemes ito 6 may be entered dependent on the rotational speed of the rotor at the time when motoring or generating is required again.

The switching scheme is operable to control the switches at up to at least 6000Hz IS (flywheel mechanical speed of 60,000 rpm). Switching scheme 7 may also be entered into if a shutdown of the flywheel system is required, for example, due to a fault.

Sensors Sensors 13 (23, 24, 25) provide measurement of rotor position and correct rotor operation to controller 14. Rotor speed is calculated in the CPLD 260 or CPU 16 When in motion, the black and white segments of the rotor cause sensor square wave outputs as shown in the table of section switching scheme 1 (PWM start)'.

This is also shown in figure 9. As will be appreciated, with the 30 degree black and white segments of the rotor and the 20 degree spacing of the sensors, there is a 10 degree rotation of the rotor before each inverter state (60° mechanical, 360° electrical) degrees for each full cycle through inverter states A-F). When the rotational speed of the rotor speeds up or slows down, the timing of these inverter state changes will also speed up or slow down in a corresponding manner.

As discussed in relation to the switching schemes, only one sensor output need be monitored by controller 14 to provide speed information of the rotor. This is because one complete revolution of the rotor requires 6 high-low pairs of output signal from any one sensor. The measurement of all three sensors for speed would not provide any other useful information when determining the speed of the rotor. The monitoring of one sensor output therefore reduces required controller resources for speed measurement.

Fail-Safe Shutdown The three sensor arrangement may provide fail-safe shut down of the flywheel apparatus. One or more devices may monitor the sensor outputs to detect the waveforms that are shown in Figure 9. The sensor outputs may be checked against expected outputs as would normally be caused by the black-white sections moving past each sensor and producing either a high or a low output dependent on a black or white segment being present at the sensor.

Figure 10 shows sensors 13, controller 14 in more detail than Figure 1. In a hardwired embodiment, each sensor 23, 24, 25 outputs to a respective CPLD (complex programmable logic device) 232, 242, 252. Any programmable logic device may be used such as an FPGA, flash programmable device or other suitable device. Each device 232, 242, 252 associated with each sensor has an output to a further programmable device 260. As would be understood, any one sensor provides a signal comprising square waves in relation to the passing black and white segments of the rotor. Six square waves are produced per full rotation of the rotor (six N-S pole pairs), and from this a speed reading may be derived by any of the programmable logic devices (232, 242, 252) associated with each sensor.

In addition to speed data, the three sensor system provides shutdown in case of failure. A failure could be caused, for example, by physical failure of the flywheel arrangement, or failure of one or more sensors. Such faults, when detected by the sensors in conjunction with associated components cause a shutdown of the flywheel system to avoid a potential catastrophic failure.

The three sensors 23, 24, 25 provide outputs to the three CPLD's 232, 242 and 252 respectively. In an embodiment, the three CFLD's are arranged to provide a signal indicative of correct operation of the respective sensor to CPLD 260.

Figure 14 shows method steps that may be performed in each of CPLD 232, 242, 252 in conjunction with CPLD 260 in relation to fail-safe shutdown. With this arrangement, programmable device 252 may constantly self-monitor (see step 141 of Figure 14). If the expected square wave waveform (as shown in Figure 9) is detected from sensor 25 by programmable device 252, an output is provided to device 260 of a signal indicative of correct operation of sensor 25 (see branch 143 of Figure 14). Conversely, if an unexpected waveform is provided by sensor 25, which may include no waveform at all, then a signal indicative of incorrect operation (which could be caused by either or both of a failed sensor or a failed flywheel assembly) is provided to programmable device 260 (see branch 144 of Figure 14).

Device 242 may self-monitor in the same manner in relation to sensor 24 and device 232 may self-monitor in the same manner in relation to sensor 23. Therefore, it can be seen that only if all three sensors are providing expected waveform outputs will a signal indicative of correct operation of all three sensors be provided to programmable device 260. Accordingly, device 260 is arranged to stop gate drive module 9 (step 148) from providing the gate switching signals Si to S6 upon deducing (step 146) that any of the sensors is providing an incorrect waveform. No stop signal is produced if all three sensors are providing expected waveform outputs (step 150). This way, if any one of sensors 23, 24 or 25 is not producing a correct output, any additional energy is not transferred to the flywheel and operation of the flywheel will immediately stop. As would be understood, steps 141 and 146 continuously take place in order to provide fail-safe shutdown operation.

For example a partial failure (such as a crack) of the rotor may cause fine particles of carbon dust to be liberated within the enclosed vacuum housing 15 of the flywheel and sensors. The carbon dust particles effectively blind one or more of the optical sensors by coating the sensor to prevent either the illuminating light or the sensing of the light or both. As a result, the incorrect waveform detected will cause the failure mode to shut down the system by preventing operation of gate switching signals Si toS6.

In the hardwired embodiment as described, there is no requirement for the software executed by either processor 17 or programmable device 260 to have knowledge of the need to shutdown upon a sensor failure. The monitoring of the sensors in relation to expected sensor operation is taken care of by programmable devices 252, 242, 232. This avoids the need for safety critical testing of the software and reduces software development and maintenance time.

In another embodiment, the programmable devices may be omitted. In this embodiment, individual sensor outputs that are input to programmable device 260 would therefore form part of the safety shutdown system. In such an embodiment, the software would also form pad of the safety critical system in which case robust software testing and maintenance of this functionality would be necessary.

Optionally, in an embodiment, the outputs from sensors 23, 24 and 25 may be fed to the Control Logic CPU 17, where additional control software may be provided to control the flywheel system based on the sensor outputs. In this embodiment, CPU 17 is arranged to provide an output to CPLD 260 to instruct CPLD 260 to drive gate drive unit 9 (or not if incorrect sensor operation is detected).

Overspeed Detection Each physical flywheel assembly has a preferred speed range as discussed previously, and a maximum speed for reliable operation. The maximum speed is determinable from the stresses and strains that the component pads of the flywheel assembly 12 undergo when the rotor 19 is in motion.

Figure 15 shows method steps that may be performed in each of CPLD 232, 242, 252 in conjunction with CPLD 260. Each sensor 23, 24 or 25 is able to monitor for overspeed detection. Programmable devices 232, 242 and 252 may monitor the waveform from each sensor (step 152). By counting either a high or low edge of square waves produced by each sensor from the passing black and white segments of rotor 19 when it is in motion, a logic signal can be output to programmable device 260 if too high a speed has been reached. In one embodiment, 18 cycles (3 rotations) are counted. The time taken for the 18 cycles to propagate though the counting circuitry determines the running speed. Programmable devices 232, 242 and 252 may each comprise an inbuilt clock or may be provided with a clock signal (not shown) in order to count pulses in relation to time.

Should the speed be determined to be above a predetermined maximum, gate drive signals Si to 56 may be disabled (by way of an overspeed signal from a respective CPLD at step 154). No overspeed signal is produced if the speed is below the pre-determined maximum (step 156). CPLD 260, in turn, monitors (step 158) for the disable signal and may provide a further signal to disable the gate drive signals (step 162) in order that the flywheel does not increase in rotational speed any further. No stop signal is produced if no overspeed signal is received by CPLD 260 (step 160).

For example, if sensor 23 is considered to be indicating an over speed scenario, switch signals Si to 56 to the inverter are subsequently shutoff, thereby not allowing the inverter 11 to commutate the flywheel system. This provides a safe shutdown of the flywheel system.

If appropriate, to slow the flywheel do'ni rapidly energy may be extracted from the flywheel at this time which has the effect of slowing the flywheel down for example by shorting the windings through an appropriate resister or heat sink. This caters for the situation where the reason for the shutdown is a supporting electronics failure as the electronics are not used to slow the flywheel. If they were used to slow the flywheel, the fault could be such that the flywheel would be re-energised and a catastrophic failure could inadvertently occur.

The approaches and methods described herein may be embodied on a computer readable medium, which may be a non-transitory computer readable medium. The computer readable medium carrying computer readable instructions arranged for execution upon a processor so as to make the processor carry out any or all of the methods described herein.

The term "computer readable medium" as used herein refers to any medium that stores data and/or instructions that cause a processor to operation in a specific manner. Such storage medium may comprise non-volatile media and/or volatile media. Non-volatile media may include, for example, optical or magnetic disks.

Volatile media may include dynamic memory. Common forms of storage medium include, for example, a floppy disk, hard disk, solid state drive, magnetic tape, or any other magnetic data storage medium, a CD-ROM, any other optical data storage medium, any physical medium with patterns of holes, a RAM, a PROM, and EPROM, a FLASH-EPROM, NVRAM, any other memory chip or cartridge.

Accordingly, the improved method and system described herein can be used to implement a flywheel assembly in a range of different applications, for example in high speed motor vehicles and/or in public vehicles including buses, trams, and other devices such as cranes, lifts and aircraft for example. l0

The flywheel assembly provides a kinetic energy recovery device whereby electrical energy can be recovered rather than being lost to the environment as would occur if no recovery system were provided. For example, rather than slowing down a vehicle by the use of friction brakes and dissipating excess energy as heat, the energy may be recovered as the vehicle is slowing down.

Claims

CLAIMS1. A method of controlling a flywheel comprising a rotor, the method comprising the steps of: detecting rotational motion of the rotor at a plurality of sensors; providing, from each sensor, a motion signal indicative of the rotational motion; comparing each motion signal to an expected signal; and if any of the motion signals are not equal to the expected signal, disabling the flywheel to prevent flywheel acceleration.
2. The method as claimed in claim 1 wherein the expected signal comprises a square waveform.3. The method as claimed in claim 1 or 2 wherein the sensors comprise optical sensors.
3. The method as claimed in any preceding claim wherein the plurality of sensors detect dark and light sections of the rotor of the flywheel.
4. The method as claimed in claim 1 further comprising, based on one or more motion signals, deriving a rotational speed of the rotor, and if the rotational speed is above a predetermined threshold, disabling the flywheel to prevent flywheel acceleration.
5. The method of any previous claim wherein providing a signal to disable the flywheel to prevent flywheel acceleration comprises disabling switching signals of an inverter of the flywheel.
6. The method of claim 4 or 5 wherein the predetermined threshold is determined by stresses and strains that the flywheel assembly undergoes when in motion.
7. The method according to any previous claim wherein the sensors are arranged to detect dust particles from the rotor blinding at least one of the sensors.
8. A flywheel control system comprising: a rotor; a plurality of sensors each arranged to detect rotational motion of the rotor and to provide a motion signal indicative of the rotational motion; a controller arranged to compare each motion signal to an expected signal; and if any of the motion signals are not equal to the expected signal, the controller arranged to disable the flywheel to prevent flywheel acceleration.
9. The flywheel control system according to claim 8 wherein the flywheel rotor and the sensors are arranged to be positioned within a vacuum.
10. The system of claim 8 or 9 wherein the sensors are positioned at 200 intervals around an axis of rotation of the rotor.
11. The system according to any of claims 8 to 10 wherein the magnetic rotor comprises N-S pole pairs.
12. The system according to claim 11 wherein the N-S pole pairs are positioned at 30° intervals around an axis of rotation of the rotor.
13. The system according to any of claims 8 to 12 wherein the sensors comprise optical sensors.
14. The system according to any of claims 8 to 13 comprising 3 sensors.
15. A flywheel control system according to any of claims claim 8 to 14 further arranged to carry out the method of any of claims ito 7.
16. Apparatus comprising a processor arranged to perform the method of any of claims ito claim 7.
17. A computer program comprising computer program code that is executable in use to perform the method of any of claims 1 to claim 7.
18. A computer readable medium comprising computer program code that is executable in use to perform the method of any of claims 1 to claim 7.
19. A method substantially as herein described with reference to and as illustrated in any combination of the accompanying drawings.
20. A flywheel control system substantially as herein described with reference to and as illustrated in any combination of the accompanying drawings.