A Kinetic Energy Storage System for Railways Applications

1 I.J. Iglesias. Presenting Author, 1J.C. Martinez, 1C. Tobajas, 2M. Lafoz, 2L. García-Tabarés and 3C. Vazquez. Coauthors

1ADIF. Railways Infrastructure Manager, Madrid, Spain. 2CEDEX Studies and Experimentation Centre of Public Works and Transportation (Fomento) Ministry, Madrid, Spain. 3CIEMAT Centre for Energy,

Environment and Technology Research, Madrid, Spain

Abstract

This paper shows the design, development and tests of a Kinetic Energy Storage System (KESS) developed jointly by ADIF and CEDEX to be applied in a rail electrical substation. The basic behavior of such a system is to store the braking energy of trains in a rotating flywheel and to give energy back once it is needed to give traction power to the same or other trains fed by the same catenary. CEDEX work is focused on the design and development of the KESS unit itself, and ADIF is working on analyzing the most appropriate place to install the system from the point of view of optimizing the energy saving, as well as the auxiliary systems needed to install the KESS unit it in a rail substation.

1.- Introduction

The electric consumption in rail electric substations is characterised by the existence of high peaks in some traffic conditions. This conduct to the oversizing of some specific equipments such transformers, power converters, protection devices etc.., and in some cases to possible penalties of the supplying electric companies due to the overdimensioning of the line supplying energy to the rail system. (figure 1) This phenomena appears not only in the electric substations feeding the high speed lines but also in those feeding conventional lines. In both cases is due to the difficulty of predicting the behaviour of loads varying both in time and space and caused as a consequence of this variable energy demand.

Figure 1: Typical energy consumption at an electric railways substation.

In order to improve this situation, ADIF as the Spanish Rail Network Manager, is working in collaboration with CEDEX and other Spanish institutions (CIEMAT, Seville University, Green Power Technology, Elytt Energy and Metro de Madrid) on the possibility of installing in some specific substations a Kinetic Energy Storage System able to supply energy during these consumption peaks and smoothing the load profile, as it is shown in figure 2.

- 6 ,0

- 4 ,0

- 2 ,0

0 ,0

2 ,0

4 ,0

6 ,0

8 ,0

10 ,0

12 ,0

14 ,0

POTE

NC

IA (M

W)

POTENCIA DE DISEÑO

- 6 ,0

- 4 ,0

- 2 ,0

0 ,0

2 ,0

4 ,0

6 ,0

8 ,0

10 ,0

12 ,0

14 ,0

POTE

NC

IA (M

W)

POTENCIA DE DISEÑOPOTENCIA DE DISEÑOPOWER PEAKS

Figure 2. Left side: Normal Consumption. Right side: Peak power elimination by means of installing an energy storage unit.

2.- Energy Storage coming from the regenerative braking of trains

When electric regenerative braking is implemented in the trains, the energy coming from this braking process can not always be used for another trains and therefore it have to be eliminated in electric resistances (reostatic braking). Some studies performed at Spanish railways have shown that, in some lines this braking energy can reach an important value, and therefore, the possibility of recovering it become more important. In big cities commuter lines, characterized by a high traffic density as well as a high number of stops, which produces a frequent number of braking and accelerating process, the fact of recovering the braking energy becomes very important. These commuter lines seem to be the more appropriate place to install energy storage units. These units not only will improve the power consumption curve but also will store the energy generated by the train braking. Regenerative braking sends the energy to the catenary, and there are three possibilities: a) The energy is consumed by another train b) The energy is given back to the electric network if the substation is bidirectional and c) The energy is stored in anyway waiting to be consumed by other train. Possibility a) is the best one but it depends on the traffic; possibility b) needs a bidirectional substation (not normally available) or an special power electronics unit able to give back energy to the network and finally if options a or b are not possible the only possibility to not loose the energy is to store it. Figures 3 a and 3b show the two last possibilities:

Figure 3a. Trains equipped with regenerative braking make posible to give back energy to the electric network.

0 1 2 3 4 5 6 7 8 9

x 104

0

1

2

3

4

5

6x 104

TIEMPO(S)

PO

TENC

IA(k

W)

0 1 2 3 4 5 6 7 8 9

x 104

0

1

2

3

4

5

6x 104

TIEMPO(S)

PO

TENCIA

(kW

)

Figure 3b. To avoid the loss of the braking energy, when it can not give back to the network or

consumed by other train, a Kinetic Energy Storage System (KESS) will be installed in the electric substation.

The working principle of the KESS unit is the following: if there is no any rail vehicle in the area when a train is braking, the energy given back by this train will cause an increase of the catenary voltage. This level of voltage will cause that the flywheel converter will increase its input current causing the acceleration of the flywheel. Once the catenary voltage will give back to its nominal value, or in case of reaching its maximum speed, the flywheel converter will decrease its input current up to zero and the energy will remain stored in flywheel rotating kinetic energy. The system will remain at this stage up to some demand of energy will be done by an incoming train (or the train stopped will start moving). In this case the KESS converter will supply to the catenary the energy needed by the train and it will reduce its speed up to reaching the minimum value. In this moment if more energy is demanded by the catenary, the KESS unit will be disconnected and the energy will be supplied by the substation. The KESS unit control is therefore very simple and it is based on the catenary voltage level, the minimum and maximum speed of the flywheel and the energy demanded or generated by the trains.

3.- Application of KESS in ADIF network

From the analysis performed by ADIF in different rail lines, it can be concluded that the amount of energy given back to the catenary is very important, because at the time being all the electrical vehicles incorporate regenerative braking ability. The demanding longitudinal profile of the Spanish rail network where high slopes are usual, increase the possibility of regenerating energy from the train braking. In urban commuter lines, characterised by the high traffic density and a high number of commercial stops-and therefore very often braking and accelerating processes- the previous fact becomes even more important, being these lines the most appropriate place to install energy storage systems. In this case the two main targets of these systems- load curve smoothing and energy saving-can be

achieved. In this case as the KESS has to be prepared both to store or supply energy, the nominal value of the energy stored is 50% of the maximum value, which means a nominal speed of 70% the maximum value. Another possible operation mode of KESS can be to support the catenary voltage in places far from the electric substations, where the voltage can drop below the permitted value. In this case the nominal state of the KESS is the maximum speed, storing the maximum amount of energy because the interaction with the catenary is only in the sense of injecting energy when the voltage can not be maintained by the distant electric substation. The KESS unit supports the catenary voltage up to reaching its minimum permitted speed. Once the catenary voltage has been recovered, the KESS starts accelerating again up to reaching its maximum speed and being prepared for a next possible intervention. Apart from the previous advantages-storage and latter use of the trains braking energy, catenary voltage support and possible reduction of electrical substation dimension-the installation of a KESS unit can give the following advantages: 1) Economical negotiation with electric suppliers by reducing the unbalance caused by rail electrical substations. 2) Modular concept of KESS allows a progressive implementation as well as a progressive increase of substation power without replacing the existing components. 3) To be used as an UPS in substations and signaling systems. 4.- SA2VE (Advanced Energy Storage System) Project In 2003, ADIF in collaboration with CEDEX, an Institute belonging to the Public Works Ministry, launched a first project (named ACE2) for developing a conventional flywheel energy storage system for power management in high speed lines substations with the aims of power consumption levelling and braking energy recovery. At the time being, the first two prototypes of Energy storage Units are finished and have been tested at CEDEX facilities. Prior to these tests, two previous and smaller prototypes have been developed and tested in the same facilities. The system is based on the use of a metallic flywheel, a switched reluctance motor/generator (SRM) and the corresponding power electronics, including a machine side converter and a grid side converter. 200 MJ (55.6kWh) of energy can be stored in a high strength steel flywheel, rotating over 7.000 rpm, which is magnetically levitated and guided with conventional high speed bearings. 350kW of power can be exchanged in the electrical machine which is a switched reluctance, 6/4 poles, motor-generator, working in either single pulse or hysteresis band, depending on its speed. The machine is driven with a full H converter, bidirectional in voltage and current thus reducing magnetic losses in the rotor of the SRM. Finally there is a grid side converter connecting the grid of the substation to a dc-link previous to the machine converter. It is a conventional single-phase H converter, controlled by means of a double hysteresis-band current strategy. A specific and very robust digital control platform has also been developed for this application to manage the complete system. A second and more ambitious project has started in 2006 to substitute the conventional bearings with magnetic bearings and a more efficient SRM control. The main objective of this second project (named SA2VE, financed by the Education and Science Ministry of Spain) is he application of the kinetic energy storage system (KESS) to both smoothing the consumption of a DC substation (3kV for urban trains or 1.5kV for metro) and storing the energy back from the train braking. The three main parts of the KESS, i.e. a) the set electric machine+flywheel, b) The intermediate transformer+power converter and c) The control unit are shown in figure 4. Once this prototype will be tested successfully the third and last part of this ambitious project would be the installation of 16 units similar to the one developed in a rail substation. This system would be able to store an energy of 3.200 MJ at a rate of 5.600 Kw. This system would be able to supply energy to a 5.6 Mw train during 9.5 minutes, having therefore an extremely high capacity of smoothing the electric substation power peaks.

Figure 4. Schematic of the kinetic energy storage system and it implementation in a real track.

5.- The Kinetic Energy Storage System (KESS) Description

The part of the system where the energy is physically stored is the flywheel which is made of a high strength steel alloy, forged and machined to an almost-cylindrical shape avoiding stress concentration at any point and reducing them to the maximum theoretical value which is located in the axis of the cylinder. The shaft of the flywheel is machined from the same initial piece. Its lower part holds the rotor of the electrical machine and the bearings are placed at each end of this shaft. The overall weight of the flywheel is 6 Tons and in order to reduce the load of the bearings, it is magnetically levitated using a hybrid system of permanent magnets and one electromagnet which adjust the net weight to the required optimum value. Guidance of the flywheel is achieved by conventional high speed ceramic bearings. The rotating mass is working in a vacuum in order to reduce frictional losses. To rotate the flywheel a 6/4 Switched Reluctance Machine is used [4]. Although other electric machines suit the KESS application, for example PM synchronous and induction machines, a switched reluctance machine was selected mainly because of its low cost, robustness and low rotor losses. Its rotor is shrank fitted in the shaft of the flywheel and its stator in the housing of the system. The machine has been designed taking into consideration not only electromagnetic parameters to achieve the required power, but also mechanical aspects to guarantee a level of stresses in the rotor compatible with those of the flywheel. The final prototype is provided with a 350 kW machine and a 200 MJ flywheel. A low pressure atmosphere inside of the machine considerably reduces the aerodynamic losses of the flywheel. On the other hand, more complex machine design is required to guarantee a pressure isolation. Extra isolation being required at the windings and the feedthroughs. Breakdown voltage is reduced with the pressure as the Paschen’s law [5] demonstrates.

V = f(p,d) , where ’p’ is the pressure and ‘d’ is the distance.

The theoretical relationship for the direct current breakdown voltage of two parallel-plane electrodes immersed in a gas, as a function of the gas pressure and electrode separation, is shown by the Paschen curve. This relationship predicts the occurrence of a minimum breakdown voltage for a certain product of the pressure times the separation. Therefore the electric isolation required is higher as the pressure is reduced inside the machine. As for the bearings, since no commercial bearings are available to simultaneously support the speed and axial load, a magnetic levitation system has been provided to reduce the axial load, leaving only 500 Kg so as to ensure the mechanical stability of the system. Two levitation systems based on permanent magnets and electric coil are in charge of reducing the load on the bearings.

ELE

ELECTRIC MACHINE(SRM) + FLYWHEEL

Electric

CONTROL UNIT

POWER CONVERTER

NETWORKNETWORK

6. - The Power Converters

A double electronic converter is used to drive the KESS, composed of a MSC and a GSC sharing a common DC link. The MSC is in charge of driving the switched reluctance machine [6][7][8][9][10] and the GSC is in charge of controlling the DC link voltage and connecting the system to the network. Figure 3 shows the power electronic circuit topology used to drive the machine. The DC link voltage varies from 650 V to 1000 V depending on the machine speed in order to achieve the single pulse situation at the machine phases and reducing the commutation losses. Average commutation frequency at the MSC is 600 Hz. Full-bridge IGBTs topology has been chosen for the MSC to get bidirectional current, maintaining the magnetic orientation at the rotor and minimizing the rotor losses [11]. Moreover, soft-switching techniques and a balanced commutation between the IGBTs have been implemented to maintain the same stress level in all the switches. Since only single-phase connection is permitted at the railway substation, GSC topology also acts as a full-bridge and it is operated by a double-hysteresis band current control strategy. Charging and discharging energy cycles is important for the thermal design of the power electronics devices. Output voltage is 400 V because the electric isolation requirements at the machine, which is increased to 25 kV at the substation by means of a power transformer. Figure 5 shows the power electronics converters topology as used to drive the KESS.

Figure 5. Power Electronics Converters topology used to drive a KESS connected to a single-

phase grid. From 0 to 4500 rpm the machine is supplied with 650 V and the current level, due to the reduced electromagnetic force, is controlled by hysteresis band strategy. Equation 1 presents the current evolution in a saturated machine. The usual operation range is between 4500 and 6600 rpm whilst the DC voltage is controlled so as to achieve a single pulse current in the machine phases. As a result, the commutation losses at the converter and the hysteresis losses in the machine are being reduced to their maximum.

dtd

ddwiRV

dtdi dc

λθλ

⋅−⋅−= (Eq.1)

Figure 6 shows the different commutation modes for the MSC, hysteresis-band and single-pulse. Current and voltage are shown in the schemes with IGBTs status so as to share the electrical stress. During hysteresis band operation, +Vdc, -Vdc and 0 volts are applied to the machine phases to control the current. During single-pulse operation, just +Vdc and –Vdc is sufficient to control the current whilst the commutation frequency is reduced to the value of w(rpm)/30.

U =400V

Network-Side Converter

+Vdc

-Vdc

Phase 1 Phase 2 Phase 3

Machine-Side ConverterCTOTAL = 6,7 mF

L = 1 mH

Topology:

U =400V

Network-Side Converter

+Vdc

-Vdc

Phase 1 Phase 2 Phase 3

Machine-Side ConverterCTOTAL = 6,7 mF

L = 1 mH

Topology:

(a) hysteresis-band mode (b) single-pulse mode

Figure 6. Machine-side converter commutation modes and experimental results.

Figure 7 presents some experimental results obtained when the KESS was charged and discharged in different operation modes. Machine current and network voltage and current are displayed. Bidirectional machine current can be identified at the illustration.

Figure 7.a. SRM charging operation (motor mode) at different speeds.

Figure 7.b. SRM discharging operation (generator mode).

7. - The Control System

Full-bridge converters provide each stator winding with independent driving circuits. The proposed control architecture reproduces this idea to achieve a totally independent operation (electric and electronic) for each bridge. In this way, a decentralised control architecture based on microcontroller boards connected through a 1Mbps CAN network is used. The on-board basic circuit is a 16-bit single-chip Renesas H8s2623 microcontroller working at clock frequency of 20 MHz. Modularity and decentralisation allows to distribute tasks in order to achieve very short cycle times, which are needed to deal with the high di/dt managed at the machine-side (fsample>50kHz). The flexibility of this design makes it suitable for prototype systems, as it is easily adaptable to different machine topologies and/or changes in the control strategies. Besides, this scheme allows the controller to grow larger in a loose manner, adding additional boards for controlling or monitoring other possible subsystems. Figure 8 shows the control modules scheme for one KESS unit and the control required for a complete Energy Storage Plant.

Figure 8. KESS control scheme based on a distributed architecture. Control scheme for a complete ES Plant.

The SRM operation requires a very precise position detection. The control system receives three position signals from the SRM originating from three laser sensors, in order to determine the machine phase activation sequence and the speed estimation as presented in Figure 9. A sensorless strategy is being developed obtaining the position and speed though the current measurements. It will substitute the laser sensors, thus reducing the total cost.

Figure 9. Control boards and laser sensors to estimate position and speed.

The current and voltage measurements received by the control boards are firstly filtered to avoid electromagnetic noise. The control boards and rack are designed so as to achieve a good EMI protection. Moreover, the control desk is located 20 m from the power converters and machine and the IGBT’s switching signals are sent though optic fibre.

8. – Laboratory Testing Facilities

CEDEX has invested in some testing facilities for Kinetic Energy Storage Systems in order to acquire the appropriate mounting and transportation tools, power supply levels and safety measures. The final prototype has been tested for the application of energy storage in railway systems but not with the maximum power. It is planned to test it at rated power in a few weeks. On the other hand, some transportation companies related to railway applications like subway and intercity trains are interested in Energy Storage Systems and a new project has been launched to continue developing these technologies. Figure 10 shows the Energy Storage Laboratory in CEDEX. The KEES prototypes are located inside of the pit. Power electronics converters are also presented at the picture.

Figure 10. Energy Storage Laboratory at CEDEX.

9. - Conclusions

A complete Kinetic Energy Storage System (KESS) has been developed to smooth the power consumption profile in a railways substation, always very irregular due to trains coincidence, starting, braking, etc. The system has been designed to be robust, and maintenance free, using well-known technologies which have been pushed to reasonable limits. The development has been confirmed with the fabrication and tests of intermediate prototypes and it is now starting the first running tests at a pilot facility. A second prototype will soon be concluded and taken to a substation for real tests under traffic conditions. A sensorless strategy will substitute the laser sensors to estimate the rotor position and speed. A second generation of KESS is under development in the framework of a new R&D program which includes magnetic bearings, advanced control and new applications for these systems. The installation of KESS in a real railways electric substation will be completed in 2009 in two steps: First step would be the demonstration of the goodness of the developed prototype in saving energy and smoothing electric load consumption. Reliability of the system, although is currently tested at lab, would be strongly tested during this first step. Once the first step would be completed the second one

will consist on the installation of a real size energy storage plant, storing 3.600 MJ at a rate of 5.6MW, able to supply energy to a real train during almost ten minutes.

Acknowledgements

Projects ACE2 and SA2ve have been funded by the Spanish Science and Education Ministry. References [1] Hentschel, F.; Müller, K.; Steiner, M.; “Energy Storage on Urban Railway Vehicles”. Proceedings of the UIC Energy Efficiency

Conference, Paris, 2000. [2] C.Vázquez, M.Lafoz, D. Ugena, L.G. Tabarés, Control System for the Switched Reluctance Drive of a Flywheel Energy-Storage Module,

European Transactions on Electrical Power, 2007. [3] C.Vázquez, Desarrollo del Sistema de Control de un Almacenador Cinético de Energía con Máquina de Reluctancia Conmutada, Ph.D.

Dissertation, Universidad Complutense de Madrid, 2007. [4] T. J. E. Miller, Electronic Control of Switched Reluctance Machines. Oxford: Newnes, 2001. [5] Paschen, F.; “Uber die zum Funkenübergang in Luft, Wasserstoff und Kohlensaüre bei verschiedenen Drucken erforderliche

Potentialdifferen”z. Ann. Phys. Chem, Ser. 37, 69-96. 1889. [6] Vukosavic, S.; Stefanovic, V.R.; SRM Inverter Topologies: A Comparative Evaluation. IEEE Transactions on Industry Applications.

Vol. 27, Nº 6 , pp. 1034-1047, Nov./Dec. 1991. [7] Mir. S.; “Classification of SRM Converter Topologies for Automotive Applications”. SAE Technical Paper Series 2000-01-0133. SAE

2000 World Congress. March 2000. [8] Barnes, M.; Pollock, Ch.; “Power Electronics Converters for Switched Reluctance Drives”. IEEE Transactions on Power Electronics,

Vol. 13, Nº 6, pp. 1100-1111, November 1998. [9] Miller, T.J.E.; “Switched Reluctance Motors and their Control”. Magna Physics Publishing and Clarendon Press. Oxford 1993. [10] Krishnan, R.; “Switched Reluctance Motor Drives: Modelling, Simulation, Analysis, Design and Applications”, CRC Press LLC, New

York, 2001. [11] Mthombeni L.T., Pillay P. “Core Losses in Motor Laminations Exposed to High Frequency or Non-sinusoidal Excitation”. IEEE

Transactions on Industry Applications, vol 40. Nº5. September/October 2004. p. 1325-1332.