1

The control of switched reluctance drives and their use for flywheel energy storage
K. De Brabandere1, J. Driesen, and R. Belmans

Abstract-- This paper consists of two main parts. In the first part, the main aspects of the topology and control of a switched reluctance drive are considered and compared with the topology and control of conventional AC drives. The focus herein lies on the control of switched reluctance drives . In the second part, the switched reluctance drive is considered for a flywheel energy storage system for use in for instance an UPS. It is shown that for this application, switched reluctance drives have some essential advantages compared to conventional AC drives. Index TermsSwitched reluctance machines, generator, flywheel storage, drive control.

such as induction or permanent magnet synchronous machines. A. Torque development Torque production in reluctance machines is achieved by the tendency of the rotor to move to a position where the inductance, and hence the magnetic field energy, of the excited winding is minimized [1]. An example of a 6/4 switched reluctance machine is shown in fig. 1. This threephase machine is denoted 6/4 because it has six stator and four rotor poles. In the position shown, the resulting torque tends to rotate the rotor in clockwise direction towards the aligned position. If the rotor continues past the aligned position, due to its inertia, the attractive force between the poles produces a braking torque. To eliminate this negative torque, the current must be switched off while the poles are separating. The ideal current waveform is therefore a series of pulses synchronized with aligning rotor poles and stator poles. By exciting the stator windings of the different phases each time two rotor poles approach the appropriate stator poles, a continuous torque production and rotation of the rotor is achieved. In generator mode, windings are excited during the separation of the respective poles, yielding a braking attraction.

I. INTRODUCTION Switched reluctance machine drives have gained considerable attention among researchers due toseveral reasons. First of all, the rotor construction is simple and robust, as it consists of pure (laminated) steel with neither permanent magnets, nor electrical windings, while the stator can be easily manufactured as the windings are not distributed but concentrated around the salient poles. Secondly, the machine is cost-effective in comparison with induction machines, while it is claimed to have a comparable or even higher efficiency and power to volume ratio [17]-[20]. Moreover, it is an inherently four-quadrant variable speed drive [10]. In this paper, the main aspects of principal operation, inverter topology and control of switched reluctance machines are reviewed. Also, the use of the switched reluctance machine as a generator is considered and for this purpose a short review of publications in this matter is performed. Finally, the application in a flywheel energy storage system is considered and it is shown that this application permits to exploit most of its specific advantages.

II. PRINCIPLE OPERATION OF SWITCHED RELUCTANCE

Although switched reluctance machines are in fact synchronous machines, their operation and control differs fundamentally from conventional rotating-field AC machines,
Fig. 1 6/4 switched reluctance machine with one phase excited.
1 Karel De Brabandere is the corresponding author (karel.debrabandere@esat.kuleuven.ac.be). The authors belong to the research group ESAT-ELECTA, K.U.LEUVEN, Kasteelpark Arenberg 10, B-3001 Heverlee, Belgium, tel. +32/16/32.10.20, fax: +32/16/32.19.85. The groups website is http://www.esat.kuleuven.ac.be/electa .

The requirement for an optimized current shape makes it impossible to run such a machine directly from the grid. A power electronic inverter with a control based on feedback of the rotor position is a necessity.

2 that the produced torque is independent of the direction of the current, since i2 is always positive. With the machine driven in saturation, although (5) being no longer valid, this conclusions remain true. III. INVERTER TOPOLOGY As torque is independent of the direction of the current, the flux-linkage and the current, as well as the topology of the inverter circuit, may be unipolar. This mode of operation is preferred because it permits a simpler form of controller. However, to be able to switch off the current while the poles are aligned (in motoring mode), a negative voltage must be applied to enforce this fast current change in the relatively large inductance of the pole winding. The half-bridge phase leg circuit shown in fig. 2 can supply current in only one direction, while it can supply a positive, negative (reversed) or zero voltage at the phase terminals. This circuit is capable of operating the machine as a motor and generator. Several other circuits are possible besides the one shown in fig. 2, using less power electronic switches and employ various means to produce the reversed voltage [2]-[4]. However, in general these circuits either compromise the performance of the machine or require additional components, while the fault-tolerant operation, more in particular the impossibility to have a shoot-through, of the drive is destroyed. For this and other reasons the circuit shown in fig. 2 is by far the most common.

B. Circuit equations The general voltage equation of an electrical coil with a resistance R and with (t ) representing the instantaneous flux coupled with this coil, can be written as:

d (t ) u = R i (t ) + dt

(1)

while, in case of a switched reluctance machine the inductance L is a complex function of position and (through the non-linear ferromagnetic hysteresis phenomenon) current:

(t ) = L(i (t ), (t )) i (t ) ,
where

d = m dt

(2)

is the mechanical angle of the rotor position and

is the rotational speed.

Using (1) and (2), the dynamic phase voltage equation of a switched reluctance machine with the rotor position and the rotational speed be stated as follows:

m , neglecting mutual coupling effects, can

(3)

e = m i

Fig. 2 Half-bridge phase leg for switched reluctance machine.

IV. CONTROL OF THE SWITCHED RELUCTANCE DRIVE The control of electrical machines is typically achieved through several control loops. Fig. 3 e.g. shows the basic control scheme of a servo drive [1]. The outer control loops are generally similar for all drive types, while the inner torque loop differs in switched reluctance machines fundamentally from other electrical machines. Much of the classical theory of torque control in electric drives is based on the DC machine, in which torque is proportional to flux multiplied by current. The flux and current are controlled independently, be it through

i 2 L(i, ) T= 2

(5)

When L > 0 the torque is positive and electrical power is converted into mechanical output (motoring), while when L < 0 the torque is negative and mechanical power is converted into electrical power (generating). Note

Hence, the induced voltage contains information about the rotor position. This property can be exploited for position feedback without a shaft sensor. With constant current, (4) is linked to both increase in magnetic field energy and produced mechanical power. In unsaturated conditions, both terms equal eachother, and torque can be expressed as:

L(i, )

(4)

The last term is the back EMF e of the switched reluctance machine:

  

L(i, ) di u ph = R i (t ) + L(t ) + i (t ) i dt L(i, ) + i (t ) m

3 mathematical transformations in AC field-oriented control. Generally speaking, in classical DC and AC machines, the flux is maintained constant while the current is varied in response to the torque demand. In switched reluctance machines, unfortunately, there is no straightforward equivalent of field-oriented control. Torque in the switched reluctance machine is produced by pulses of phase current, synchronized with the rotor position. The flux in each phase must usually be built up from zero and returned to zero each stroke. The inner torque loop consists of controlling timing and regulation of these pulses. Usually there is no torque sensor and therefore the torque control loop is not closed. Because the switched reluctance motor drive cannot be characterized by a simple torque constant kT (torque/ampere), smooth torque control must be achieved through feed-forward torque control algorithms, incorporating some kind of motor model. The fact that the machine inductance is not only dependent on the rotor position, but also on the excitation current, complicates the development of control strategies for switched reluctance machine drive systems. For all other electrical machines, control strategies are derived based on the machine parameters being constant for most of the excitation range [1], [4]. V. SWITCHED RELUCTANCE MACHINE AS A GENERATOR A. Current control Current control of the switched reluctance machine in generating is identical to motoring mode. The only difference is timing of current pulses. While in motoring mode the windings are excited during rising inductance (approaching of poles), in generating mode each winding is excited while the poles separate from each other. The objective of the generator control is normally to keep the DC-link voltage at the desired value. Unlike field-oriented AC machines, power generation is not continuous but pulsewise. A DC-link capacitor must be retained to provide excitation power during the fluxing interval, more in particular during the first part of each stroke. The ratio between excitation energy and generated energy is called the excitation penalty [9]:

Pin,elec Pout ,elec

(6)

Fig. 3 Servo drive control system

A. Low-speed control: chopping At lower speeds, the current is regulated either by PWMvoltage or direct instantaneous current regulation (e.g. hysteresis band current control). The phase voltage is switched between UDC and +UDC (hard-chopping) or between zero and +UDC (soft-chopping) (fig. 2). As speed increases also the firing or commutation angles (turn-on and turn-off angles) of the commutation cycle are advanced. The turn-on angle must be advanced to be able to build up the flux from zero to the desired value before the inductance starts rising (approaching of poles). Similarly, the turn-off angle must be advanced to be able to reduce the flux to zero before the inductance starts diminishing (separating of poles). B. High-speed control: single-pulse operation As speed increases, the back-EMF e increases according to (4). The speed for which e equals UDC is called base speed [3]. At speeds higher than this base speed, there is insufficient voltage available to control the current. The current can then only be controlled by timing of current pulses. This control mode is called single-pulse mode.

w v y x wv v u

b a`

V XYTW U

The DC-link voltage decreases during the fluxing interval, and increase during the de-fluxing interval. The variation or ripple in the DC-link voltage depends on the energy needed to build up the flux each stroke and the DC-link capacitance. Ref. [8], [15] show that the switched reluctance generator is unstable in open-loop when operated with fixed commutation angles. The essence of its instability is the proportionality of the excitation power to the back-EMF. If the load is too small, the DC-link voltage rises exponentially towards infinity, and if the load is too large, the DC-link voltage falls towards zero. For this reason it is necessary to feedback the DC-link voltage or to control the average output current independently of the DC-link voltage. The controller must maintain the average DC-link voltage constant in much the same way as it must maintain a constant average torque in motoring mode. Therefore, in principle, the architecture of a generator controller is similar to that of a motor controller. However, the DC-link capacitance has an integrating or smoothing effect such that it requires a lower bandwidth to control the DC voltage than to control the torque [1]. Although relatively little has been reported on the design of the switched reluctance generator controller, there is a broad variety in the proposed controller schemes [1]. At lower speeds, the current pulses are controlled by chopping the voltage between zero and UDC [5], [7] or between +UDC and -UDC [6] with fixed commutation angles. The control is similar to the current chopping control in motoring mode. At higher speeds the generator is operated in single-pulse mode [8]-[12]. The current is regulated by adapting the commutation angles. As the speed increases, also the amplitude of the negative back EMF increases and can become larger than the negative supply voltage. In this case an uncontrollable rise of the phase current occurs. To avoid over-current conditions and torque pulsations, the control angles have to be retarded as a function of the DC-link voltage as well as of the speed. Measuring the

t s rp i h g q

SRGQ P I HG G F

E 87 D C 7 B A @ 98 7 6

fe

' 15( 4 ( 3 012 ) '( & %

$ # !  " !     

4 variation of the DC-link voltage is necessary to avoid unstable operating conditions. Ref. [13] discusses the occurrence of over-currents in switched reluctance generators more in detail. Ref. [12] describes a single-pulse method, trying to maximize the energy conversion effectiveness at high rotor speeds. It is shown that maximum energy conversion is obtained when back-EMF e and UDC are in balance. This balance is achieved through adaptation of the firing angles and variation of the DC-bus voltage in proportion to the speed. The drawback of this method is the necessity of a step-down DC/DC converter in order to achieve a constant output voltage. B. Start-up Since the switched reluctance machine is singly excited, like the induction machine, a DC source is necessary during startup from cold, i.e. with the machine not magnetized and no voltage on the DC-link capacitor. Once started, the system is self-sustaining in its reactive power and the DC source can be disconnected. The efficiency of the switched reluctance generator is very similar to the efficiency of the switched reluctance motor. The main difference is that the losses in the converter can be reduced as in the generator, the bulk of the current flows through the diodes having in general less conducting losses than the switching devices [14]. Switched reluctance machine rotors are constructed entirely from high-strength low-cost materials.

A comparison between induction machines and switched reluctance machines is less simple: Because of the absence of currents flowing in the rotor, there are no Joule-losses in the rotor, on the other hand eddy current and hysteresis losses might be higher. Efficiency, speed-range, power density and cost are comparable, but often in slight favor of the switched reluctance machine [17]-[20]. The inverter topology of a switched reluctance drive is inherently more fault-tolerant because of the absence of shoot-through possibility. Disadvantages of the switched reluctance machine drive are its rather complex control, often higher sound and vibration levels and its relatively recent development. However, the simple and yet robust construction of the switched reluctance machine rotor, together with its high efficiency, high power density and wide speed-range makes this machine a serious competitor for flywheel energy storage. Moreover, the simple layout of the rotor makes it possible to produce a high-inertia rotor with a multipole cross-section. By applying this topology, the requirement for an additional flywheel inertia may be omitted. Finally, its implicit fourquadrant operation allows for a direct response in case of a sudden energy demand, e.g. when a voltage dip occurs.

VI. APPLICATION IN FLYWHEEL ENERGY STORAGE In flywheel energy storage, electrical energy is stored in the form of kinetic energy. As the amount of stored energy is proportional to the square of the rotation speed, a high-speed, low-loss operation is desired. In view of its application in for intstance an uninterruptible power supply system (UPS) or a dynamic voltage restorer (DVR), reliability and immediate and fast energy transfer are key. Hence, primary requirements for the motor/generator of a flywheel energy storage system can be summarized as [16]: high output power capability, fast response on sudden energy demand (e.g. when a voltage dip occurs), ability of the rotor to withstand high speeds, ability of the rotor to operate in vacuum with adequate thermal management (i.e., little or no heat generation in rotor), negligible zero-torque spinning losses, high reliability, high efficiency and low cost. Switched reluctance machines have the following advantages over permanent magnet synchronous machines: There is no concern with demagnetization, hence switched reluctance machines are inherently more reliable than permanent magnet machines. There need not be an excitation field at zero torque, thus eliminating electromagnetic spinning losses.

VII. CONCLUSIONS The basic operation of a switched reluctance machine is reviewed for the motor as well the generator operation, Thereby illustrating its advantages over other types of standard AC drives. The four-quadrant operation is appropriate for the operation as a flywheel energy storage. It is especially very useful for high-speed operation due to its mechanical properties and low losses.

VIII. REFERENCES
[1] [2] T. J. E. Miller, Electronic Control of Switched Reluctance Machines. Oxford: Newnes, 2001. S. Vukosavic and V. R. Stefanovic, SRM Inverter Topologies: A Comparative Evaluation, IEEE Trans. on Industry Applications, vol. 27, no. 6, pp. 1034-47, Nov./Dec. 1991. T. J. E. Miller, Switched Reluctance Motors and their control. Lebanon, Ohio: Magna Physics Publishing, 1993. R. Krishnan, Switched Reluctance Motor Drives. Boca Raton: CRC Press, 2001. C. M. Stephens, and A. V. Radun, Current chopping strategy for generating action in switched reluctance machines, U.S. Patent 5166591, Nov. 24, 1992. C. A. Ferreira, S. R. Jones, W. S. Heglund, and W. D. Jones, Detailed Design of a 30-kW Switched Reluctance Starter/Generator System for a Gas Turbine Engine Application, IEEE Trans. on Industry Applications, vol. 31, no. 3, pp. 553-561, May/June 1995.

[3] [4] [5]

[6]

5
[7] R. Inderka, M. Menne, and R. W. De Doncker, Generator Operation of a Switched Reluctance Machine Drive for Electric Vehicles, in Proc. 1999 European Power Electronics and Applications Conf., Lausanne. D. E. Cameron, and J. H. Lang, The Control of High-Speed VariableReluctance Generators in Electric Power Systems, IEEE Trans. on Industry Applications, vol. 29, no. 6, pp. 1106-1109, Nov./Dec. 1993. P. C. Kjaer, C. Cossar, J. J. Gribble, Y. Li, and T. J. E. Miller, Switched Reluctance Generator Control using an Inverse Machine Model, in Proc. 1994 International Conf. on Electrical Machines, pp. 380-385. W. F. Ray, and M. T. Ebrahim, A novel high speed switched reluctance generator, in Proc. 1995 European Power Electronics and applications Conf., volume 3, pp. 811-816. G. Suresh, B. Fahimi, K. M. Rahman, M. Ehsani, and I. Panahi, Fourquadrant Sensorless SRM Drive with High Accuracy at All Speeds, in IEEE 1999 Applied Power Electronics Conf., vol. 2, pp. 1226-1231. T. Sawata, P. C. Kjaer, C. Cossar, T. J. E. Miller, A Control Strategy for the Switched Reluctance Generator, in Proc. 1998 International Conf. on Electrical Machines, vol. 3, pp. 2131-2136. M. Menne, R. B. Inderka, and R. W. De Doncker, Critical States in Generating Mode of Switched Reluctance Machines, in Proc. 2000 IEEE Power Electronics Specialists Conf., pp. 1544-1550. R. Cardenas, W. F. Ray, and G. M. Asher, Transputer-based control of a switched reluctance generator for wind energy applications, in Proc. 1995 European Power Electronics and Applications Conf., volume 3, pp. 69-74. A. V. Radun, Generating with the switched reluctance motor, in Proc. IEEE 1994 Applied Power Electronics Conf., pp. 41-46. H. F. Hofmann, "High-Speed Synchronous Reluctance Machine for Flywheel Applications," Ph.D. dissertation, Univ. of California, Berkeley, 1998. M. R. Harris and T. J. E. Miller, Comparison of design and performance parameters in switched reluctance and induction motors, in IEE 1989 Electrical Machines and Drives Conf., London, UK, 1989. D. A. Staton, R. P. Deodhar, W. L. Soong and T. J. E. Miller, Torque Prediction Using the Flux-MMF Diagram in AC, DC and Reluctance Motors, IEEE Trans. on Industry Applications, vol. 32, no. 1, pp.180188, Jan./Feb. 1996. P. J. Lawrenson, Switched Reluctance Drives: a perspective, in Proc. 1992 International Conf. on Electrical Machines, Manchester, UK, 1992, pp. 12-21. B. K. Bose, Power Electronics and Variable Frequency Drives. Piscataway, New York: IEEE Press, 1997, chapter 2, G. R. Slemon.

IX. BIOGRAPHIES
Karel De Brabandere graduated in 2000 as electrotechnical engineer from the Katholieke Universiteit Leuven (K.U.Leuven), where he is currently working as research assistant and preparing a PhD on the subject of the control of inverter-driven applications. His interests include power quality in distributed grids, machine drives and power electronics.

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15] [16]

[17]

Johan Driesen (S93-M97) graduated in 1996 as Electrotechnical Engineer from the Katholieke Universiteit Leuven (KULeuven). Since 1996 he was a research assistant of the Fund for Scientific Research of Flanders (F.W.O.-Vl.). He received the 1996 R&D-award of the Belgian Royal Society of Electrotechnical Engineers (KBVE) for his Master Thesis on power quality problems. He received the Ph.D. degree in Electrical Engineering at K.U.Leuven in 2000 on the finite element solution of coupled thermal-electromagnetic problems and related applications in electrical machines and drives, microsystems and power quality issues. Currently he is a postdoctoral research fellow of the Belgian Fonds voor Wetenschappelijk Onderzoek - Vlaanderen. In 2000-2001 he was a visiting researcher in the Imperial College of Science, Technology and Medicine, London, UK. Ronnie J. M. Belmans (S77-M84-SM89) received the M.S. degree in electrical engineering in 1979 and the Ph.D. in 1984, both from the Katholieke Universiteit Leuven, Belgium, the special Doctorate in 1989 and the Habilitierung in 1993, both from the RWTH Aachen, Germany. Currently, he is a full professor with the K.U. Leuven, teaching electrical machines, power electronics and variable speed drives. His research interests include power quality, electrical energy systems, electrical machine design and vibrations and audible noises in electrical machines. He was with the Laboratory for Electrical Machines of the RWTH Aachen, Germany, as a Von Humboldt Fellow (October 1988-September 1989). From October 1989 to September 1990, he was visiting professor at the McMaster University, Hamilton, ON., Canada. He obtained the chair of the AngloBelgian Society at the London University for the year 1995-1996. Currently he is a visiting professor at Imperial College in London U.K. Dr. Belmans is a fellow of the IEE (U.K.), the International Compumag Society and the Koninklijke Vlaamse Ingenieursvereniging (KVIV).

[18]

[19]

[20]