Improving energy efficiency in public transport ... - IEEE Xplore

IEEE Vehicle Power and Propulsion Conference (VPPC), September 3-5, 2008, Harbin, China

Improving energy efficiency in public transport: stationary supercapacitor based energy storage systems for a metro network Ricardo Barrero*, Xavier Tackoen**, Joeri Van Mierlo* *

Vrije Universiteit Brussel, IR-ETEC, Pleinlaan 2, B-1050 Elsene, Belgium. Email: [email protected] **Université Libre de Bruxelles, ATM, Avenue F.D. Roosevelt, 50, 1050 Brussels, Belgium. Email: [email protected] * Vrije Universiteit Brussel, IR-ETEC, Pleinlaan 2, B-1050 Elsene, Belgium. Email: [email protected]

Abstract— This article will assess the installation of stationary super capacitor based Energy Storage Systems (ESS) along a metro line for energy savings purposes. The influence of the ESS size and distribution along the line will be studied taking into account different traffic conditions. The ESSs will be configured with regards to energy content, voltage variation, maximum current and power losses. To carry out the study, an ‘effect-cause’ or ‘backwards looking’ model of the light rail vehicles and the electric network has been developed in Matlab/Simulink. A power flow controller to handle the energy flow in function of the network voltage and ESS state of charge will be proposed. Keywords — Supercapacitor, Stationary Energy Storage, Simulation, Rail Vehicles, Energy Efficiency.

I. INTRODUCTION To reduce emissions, electric powered light rail vehicles are in use in urban areas. Although these mass transit vehicles enable large reductions in terms of emissions, their energy efficiency could be significantly improved. This improvement can be reached by the hybridization of their power system, with the inclusion of an energy storage system (ESS) for energy recovery purposes [1,2,3]. Modern rail vehicles can feed back to the network up to 40 % of the energy supplied to them [4]. Since most of the substations can not feed this energy back to the mains, it should be either utilized by other vehicles or burnt in the braking resistors. The rate of energy that can be fed back to other vehicles will depend on the traffic density. The more vehicles circulate nearby, the higher the chances to utilize this energy. Recent studies estimate that for a high density metro network, the energy fed back to the network could achieve 20% of the supplied energy in rush hours while the remaining braking energy is lost in the resistors [5]. Introducing an ESS, on-board or stationary, at substation level, would increase the global efficiency of the system by capturing the, otherwise lost, braking energy. Due to the high current circulating on the metro line, Electrochemical Double Layer Capacitors (EDLC), commonly known as supercapacitors or ultracapacitors, are very convenient for this purpose. Although their C 2008 IEEE. 978-1-4244-1849-7/08/$25.00○

energy density is limited compared to that of batteries, they have optimal power characteristics and can cope with the braking power peaks [6,7]. Another added value is that, unlike batteries, which require complex algorithms to estimate the state of charge (SoC) [8], the determination of supercapacitor SoC is easily obtained by measuring their terminal voltage. Benefits such as voltage stabilization [ 9 , 10 ], peak power shaving and reduced losses on the line can also be achieved with the ESS. However, the systems proposed in this paper will be controlled to optimize the energy efficiency. The following articles [11,12] focus on the benefits of hybridizing electric powered light rail vehicles with onboard supercapacitor modules. However the influence of the ESS energy capacity is not discussed. Reference [13] discusses about the benefits of different ESSs for both on-board and stationary applications on subways, but, sizing, positioning and control algorithm are not detailed for stationary applications. A description of a real supercapacitor based ESS and measurements of prototype behaviour are given in ref. [4]. The proposed paper will describe the benefits of using stationary ESSs on the line and will present a power flow controller to handle the energy flow. The influence of the ESS size and positioning along the line will be evaluated under different traffic conditions II. METHODOLOGY A simulation program based on the ‘effect-cause’ method [14,15] has been developed in Matlab/Simulink. It can model the power flow in light rail vehicles and the feeding electric network. Starting from a given speed cycle it calculates the power requested by the light rail vehicle from the feeding network. With this power request, the voltage and current at every substation are determined. A complete description of the program can be found at [16]. III. CASE STUDY The following study will be applied to line 2 of the Brussels Metro network. The total length of the route is around 8 km, with 14 stops, and it is fed by 9 unidirectional substations. The study will cover three different scenarios: peak time, off-peak time and night and


weekend periods. The metro trains studied can be made up of 3 to 5 cars depending on the time scheduled. Each car has a tare weight of 30400 kg and a capacity for 223 passengers (counting on 4 persons per square meter), that makes a total weight of 45100 kg when it is fully loaded.

80

80

60

60

40

40

20

20

0 0

100

200

300

400

500

600

700

800

900

Altitude [m]

v [km/h]

A. Assumptions -The vehicle auxiliaries’ consumption is set at the average measured value of 20 kW/car. -The driving cycle used for simulations is based on the observations of the real route measurements. It follows the theoretical directives of accelerating up to 70 km/h with an acceleration of 1.1 m/s2 whenever it is possible. Line altitude is shown by Figure 1.

0 1000

time(s)

Figure 1. Built-in driving cycle used for simulations

-The parameters shown at Table 1 are assumed for the simulation of different scenarios mentioned in the previous section. Peak Time Off-Peak Night&WE

Cars per metro train 5 5 4

Occupancy rate [p/m2] 4 2 Only Seats

Time delay between trains [min] 3 4 10

Table 1. Simulation scenarios

- Figure 2 represents the traffic density during the peak period. It is modeled in the way that the total distance of the route is covered by a vehicle (vehicle 6, in green) that starts from the first stop (d=0) at t=0. At that moment, there are already 5 vehicles running ahead with a time span of 180s. The distance between them depends on the driving cycle part. Every 180s after t=0, one vehicle starts the cycle from the first station. Thus, the network is populated with vehicles while the vehicle 6 is covering the route. The off-peak and night/week-end scenarios are based on the same simulation principle but with fewer vehicles on the line at the same time with a time span respectively of 240s and 600s.

energy exchange is possible between vehicles of opposite directions through the substations. This is why the measured values for energy recovery are higher in the real measurements at peak time (around 34 % of the total traction energy) than in the simulation results obtained (25%) that will be shown in section IV. -Substations no load voltage is constant and set to V0 B. Parameters Vehicle Rolling resistance coefficient: Aerodynamic drag coefficient: Gearbox efficiency: Motors efficiency: Motor drive efficiency: Network Third rail resistance: Rail electric resistance: Substation internal Resistance: Distance between substations: No load substation voltage V0: DC/DC converter (SC) efficiency:

0,005 0,6 93 % 90 % 91 % 24,6 mΩ/km 17,2 mΩ/km 13 mΩ 1000 m 876V 91 %

C. Stationary Energy Storage Systems Stationary ESSs have some advantages and drawbacks when compared to on-board ESSs. On the positive side, they are installed at ground level, where weight and space is not a big handicap and they can also store and feed energy from/to different vehicles, which results in a more active system. On the negative side, the losses on the line will not be reduced as much as with on-board systems and the storage capabilities are reduced with the increase of the distance of the vehicles to the ESS. Voltage levels are limiting the energy transfer.

Figure 3. Example of vehicle sending energy to the ESS

The vehicles can send energy back to the network as long as their voltage (Vvehicle) does not go over a predefined limit, VREGEN_MAX. When this value is exceeded, the current being sent to the network is reduced. For the example of Figure 3, with only one braking vehicle and one ESS on the line, the maximum current that the vehicle will send, due to voltage limitations, follows Eq 1 Figure 2. Detail of peak period traffic density model

-The developed program models the behaviour of one direction of the metro line. In practice, the third rails of each direction are not connected directly but the substations feed both directions and, in some cases,

IV
VCHARGE? Yes

No

P=0

V< VDISCHARGE?

No

Yes

SoC0.25?

SoC0.5?


Voltage SC [V]

Figure 6 gives a relative overview of the voltage limits set. The shaded green area represents the possible charging region while the brown one shows the discharging region. Most of the charging occurs around the voltage value VCHARGE. The discharge happens at values around VDISCHARGE and lower. 920 900 880 860

SoC

Power SC [MW]

State variable

840 2

A part of this potential energy in Eq 5 is converted in kinetic energy as the vehicle reaches the top speed. Observing that the acceleration phase distance is, in this case, around one third of the distance between 2 stops and assuming a constant slope, the remaining potential energy once the vehicle is running at top speed will be around 2/3⋅EP_MAX. The energy that the vehicle can give to the network (Ev) will be given by Eq 6 and Eq 7.

1 0 1 0

-1 1

0.5 180

In theory the maximum energy that a vehicle can store corresponds to the sum of the kinetic and potential energy (EKIN and EP). Considering the vehicle weight, a maximum speed of 72km/h and a maximum altitude difference between 2 consecutive stop of 20m, then the maximum values for the kinetic and potential energy stored will be as stated by Eq 4 and Eq 5 . 1 2 Eq 4 E KIN _ MAX = ⋅ M MAX ⋅ v MAX = 12.5kWh 2 E P _ MAX = M MAX ⋅ g ⋅ hMAX = 12.25kWh Eq 5

190

200

210

220 time [s]

230

240

250

260

Figure 7. Example of controller behaviour

Figure 7 shows the behaviour of one stationary ESS at off-peak period. The top graph shows the ESS voltage at the grid connection point; the second graph shows the current state variable (1 for charge state, 2 for discharge state and 0 for neutral state); the third depicts the ESS power (positive for charge and negative for discharge) and the fourth the ESS state of charge. At t=180s the state is discharge, but the ESS is completely full, i.e. SoC=1; therefore the power is 0. At t=187s, the voltage of the ESS rises over the charge activation threshold (VUP_NEUTRA) as a consequence of a vehicle braking nearby, and the charge state is activated. Then, a charging power is requested from the ESS taking into account that the charging voltage should not be lower than VCHARGE (891 V in this case). At t=200s there is a voltage disturbance due to end of the acceleration phase of another vehicle. This provokes an initial voltage rise that is responded by an increase of the charging power of the ESS until the voltage is stabilized. Then at t=204s the SoC of the ESS is 1 and the charging is stopped. The neutral state is turned into discharge state when at t=230s, the voltage goes below the discharge activation threshold VDOWN_NEUTRAL. The ESS provides power to the network until it is empty again. E. ESS sizing Determining the optimal size of a stationary ESS is not a straight forward matter. For on-board ESS, a common principle to size it, in terms of energy, consists on the evaluation of the maximum kinetic (and potential) energy that the vehicle can accumulate. Nevertheless, when assessing the optimal size of a stationary energy storage system, some other considerations have to be taken: -Vehicles energy regeneration rate: the amount of energy subject to be stored in the stationary ESS decreases in networks with high energy regeneration rate. -Energy transfer limitation due to voltage limits as shown in Eq 1.

Ev = (EP + EKIN − EROL − EA − EBR_ MECH) ⋅ηV − EAUX

Eq 6

ηV = η gearbox ⋅η motor ⋅η drive

Eq 7

Where EP is the potential energy, EKIN the kinetic energy, EA aerodynamic drag energy, EROL the rolling resistance energy, EBR_MECH the energy dissipated by the mechanical brakes, EAUX the energy used to power the vehicle auxiliaries load and ηv the vehicle component efficiencies. Assuming that all the braking will be electrical, EBR_MECH =0, from simulations, it is obtained that the recoverable energy at pantograph level is around 0.45 to 0.65 times the sum of kinetic and potential energy, depending on the vehicle weight and auxiliaries consumption. The worst case scenario is as follows Ev_MAX ≈ 0.65⋅(EP + EKIN)= 0.65⋅[(2/3)⋅EP_MAX + EKIN_MAX] Ev_MAX ≈ 13.4 kWh The energy that can be stored in one single charge in the stationary ESS will depend on other factors. Some of them, such as the line losses and the DC/DC converter efficiency, can be estimated. However, the restrictions due to the voltage thresholds, distances, traffic conditions, energy exchange between vehicles and distribution of ESSs modules along the metro line, represented in Eq 8 by F(V,d,nv,nESS), entail a very complicated analytic solution suitable for several scenarios.

Esc = E v ⋅η L ⋅η DC / DC ⋅ F (V , d , nV , n ESS )

Eq 8

Due to all this dependencies, several simulations have been done to study the effect of the sizing and distribution of ESSs modules along the metro line. The systems subject of study will have a usable energy capacity of 2,26 kWh (small); 4,53 kWh (medium); 6,79 kWh (large) and 9,06 (extra large). F. ESS configuration Due to supercapacitor cells low voltage and energy density; it is needed to arrange them in series, to achieve a


working voltage; and to parallel several strings to obtain the required energy capacity and power capabilities. 1) Design criteria -The voltage variation of the SC will be kept between 100% and 50% of its maximum voltage. Thus, the available energy of the SC will be 75 % of the total energy stored according to Eq 9:

1 2 2 ETotal = ⋅ CTotal ⋅ (VTOTAL max − VTOTAL min ) 2

Eq 9

-The current of the SC cells will not go over 400A. Thus, its discharge efficiency, according to the manufacturer internal resistance specifications of 0.3mΩ, will be over 93%. In practice, if the SC internal resistance is higher, the maximum current should be reduced to have a higher efficiency. -Maximum ESS voltage will be lower than network voltage (876 V at no load) to ease the DC/DC converter design. 2) Proposed modules Small Cells: C=1500F, Vmax= 2.7 V. Configuration: 10 strings x 232 cells in series Usable energy: 2,26 kWh Max Voltage: 580 V Cells weight: 742 kg

Maxwell© 1500F cell

Medium Cells: C=3000F, Vmax= 2.7V. Configuration: 10 strings x 232 cells in series Usable energy: 4,53 kWh Max Voltage: 580 V Cells weight: 1275 kg


Large Cells: C=3000F, Vmax= 2.7V. Configuration: 15 strings x 232 cells in series Usable energy: 6,79 kWh Max Voltage: 580 V Cells weight: 1914 kg


Extra large Cells: C=3000F, Vmax= 2.7V. Configuration: 20 strings x 232 cells in series Usable energy: 9,06 kWh Max Voltage: 580 V Cells weight: 2552 kg


IV. SIMULATION RESULTS In order to evaluate the potential energy savings in the metro line by using stationary ESSs and to help finding its optimal size and positioning, several simulations have been run in different scenarios. First, the conventional line with several vehicles running simultaneously was simulated to obtain and record the needed data from the vehicles and substations, i.e. voltages, currents, energy consumed, energy

regenerated, etc. Next, the line was simulated with different ESS installed along; the new data were recorded and the results compared. A. Peak Time The traffic conditions, where vehicles time span is 3 minutes, are defined in Figure 2. The results for a conventional metro network are given in Table 2. Substations delivered energy [kWh] 1080 Traction energy (vehicles) [kWh] 1333 Braking energy regenerated (vehicles) [kWh] 336 Line losses [kWh] 82 Max. available braking energy (vehicles) [kWh] 615 Energy recuperation (E regen./ E traction) [%] 25 Table 2. Energy flow for unidirectional conventional line at peak time

Based on this results, Figure 8 shows graphically some of the results obtained when ESS are installed on the line. The cases with 1, 2, 4, 6, and 8 ESSs modules on the line are studied for different energy content modules. The total energy savings achieved (represented by the green line on Figure 8) grow with the number of ESSs and with the energy content of the modules, which seems logical. In this scenario, at peak time, the savings vary from 1.6 % with only 1 small ESS on the line to almost 12 % when 8 large ESS are installed. Another way to measure the influence and benefits of the ESS is to analyze what is the amount of energy saved per module every hour. This is represented by the yellow bars on Figure 8. The energy saved per module increases with its size for small capacities, but it does not significantly increase when many ESS modules are installed on the line. Besides, it seems that the increase of the number of the ESS on the line decreases the amount of energy saved by each module. With one ESS on the line, the module performance goes from 63 to 133 kWh/h saved. On the other hand, when 8 ESS are on the line, the energy saved per module varies from 43 to 58 kWh/h. The reasons of these differences are several. First, the more ESSs on the line, the more the braking energy is, somehow, split among them. Second, the amount of energy saved by an ESS module is much higher when it is installed at the end of line (which is the case of the ESS installed at 7000m, only 1000 m away from the end of the line) than in the middle of it, due to the smaller chances of the vehicles approaching the end of line to send this energy to other accelerating vehicles, since there will not be any vehicle ahead. To evaluate the global performance or ‘efficiency’ of the system, the concept of energy saved per ESS capacity installed at the line is introduced on the bottom right corner graph of Figure 8. The tendency is clear, the modules performance decreases with the capacity installed. However, it is noticed that the global energy savings increase with this parameter. A compromise between these 2 values should be found, high energy savings with acceptable module performance. Finally, when 8 large ESS are installed on the line, the vehicles energy recuperation rate goes up to 39% of their traction energy (compared to the 25% achieved with the conventional line, see last row of Table 2).


5%

0

0%

2,27

Module capacity (kWh)

Energy saved/module (kWh/h)

5%

0

[kWh/h]

10% 50

Total E savings [%]

[kWh/h]

15%

100

0% 9,06

Module capacity (kWh) Energy saved/module (kWh/h)

Energy savings

4,53

6,80

9,06

[kWh/h]


Energy savings


Energy savings

Energy saved / ESS capacity installed

150

20%

6,80

0% 2,27

9,06

8 ESS (every 1000m)

150

4,53

6,80

5%

Module capacity (kWh) Energy savings

6 ESS (every 1500m)

2,27

4,53

10% 50

0

0%

0

9,06


5%

15%

100

20%

35

1 ESS at 7000

15%

100

10% 50

5%

0

[(kWh/h)/kWh]

6,80

10% 50

Total E savings [%]

4,53

15%

100

Total E savings [%]

10% 50

[kWh/h]

15%

100

20%

150

20%

2 ESS at 2000 & 6000

25

4 ESS every 2000m

15

6 ESS every 1500m

0% 2,27

4,53

6,80

9,06


Total E savings [%]

150

20% Total E savings [%]

[kWh/h]

150

2,27

4 ESS (every 2000m)

2 ESS at d = 2000 & 6000m

1 ESS at d=7000m

8 ESS every 1000m

5 2,27 4,53 6,80 9,06

Energy savings


Figure 8. Results for simulations of different scenarios at metro peak periods.

B. Off-peak time In this case, the vehicles time delay is 4 minutes. Results for a conventional network are given by Table 3. Substation delivered energy [kWh] 744 Traction energy (vehicles) [kWh] 859 Braking energy regenerated (vehicles) [kWh] 159 Line losses [kWh] 43 Max. available braking power (vehicles) [kWh] 392 Energy recuperation (E regen./ E traction) [%] 18 Table 3. Energy flow at off-peak time

At a first sight, one remarkable fact when comparing peak and off-peak periods: the vehicles energy recuperation rate goes down to 18% due to the lower number of vehicles and their higher distance. This energy regeneration difference is subject to be stored in the ESS. Results of energy savings with ESS on the line are given by Figure 9. The tendencies are very similar to the previous scenario (peak time). The energy savings increase with the energy capacity installed. They vary from 3 %, with one small ESS module, to more than 18 % with 8 large modules on the line. This savings are higher than those obtained at peak time scenario due to the smaller regeneration rate achieved by vehicles at off-peak periods. The energy saved per module [kWh/h] is also a bit higher, an average of 10 kWh/h higher savings than those at peak time. With 1 ESS on the line, the modules savings vary from 62 to 171 kWh/h while they go from 50 to 63kWh/h with 8 ESS. On the one hand, it could be thought that the lower traffic density could drive to lower module performance,

but, on the other hand, the higher amount of available braking energy, due to the lower regeneration rate among vehicles, compensates for it. Moreover, the proportion of line losses over total energy provided is lower in this scenario, around 5,7 %, than that at peak time around 7,5 %. This is due to the lower current on the line. It is clear that the energy consumption reduction is given thanks to the higher receptivity of the network. The vehicle energy regeneration rate grows from 18% (conventional line) to almost 41 % with 8 ESS modules. In the peak time period this difference was smaller: from 25% for the conventional line to 39% with 8 ESSs. C. Night and weekend periods The vehicles time delay in this case is 10 minutes. Results for a conventional network are shown in Table 4. Substation delivered energy [kWh] 235 Traction energy (vehicles) [kWh] 246 Braking energy regenerated (vehicles) [kWh] 18 Line losses [kWh] 7 Max. available braking power (vehicles) [kWh] 115 Energy recuperation (E regen./ E traction) [%] 7 Table 4. Energy flow at night and weekend periods

Results of savings with ESS on the line are depicted in Figure 10. It is observed that the energy savings grow fast in function of the modules capacity when up to 4 ESS are present on the line. When more than 4 ESS are installed, the energy savings grow slowly while the modules behaviour gets worse. On the contrary to what happened in the previous scenarios, where the modules have a better performance at the end of lines, now, with low traffic density, they have a


2 ESS at d= 2000 & 6000m

100

10%

50

5%

0

10%

50

5% 0% 2,27

9,06

6 ESS (every 1500m)

100

10%

50

5%

0

[kWh/h]

15%

Total E savings [%]

[kWh/h]

100

10%

50

5%

0

0% 2,27

Energy savings


15%

100

10%

50

5%

1 ESS at 7000

35

2 ESS at 2000 & 6000

25

4 ESS every 2000m 6 ESS every 1500m

15

0% 2,27

4,53

6,80

9,06

8 ESS every 1000m

5




Energy savings

9,06

20%

0

0%

6,80


150

9,06

4,53


8 ESS (every 1000m)

150


15%

Energy savings

200

20%

6,80

150

9,06


Energy savings

200

4,53

6,80

20%



2,27

4,53

[(kWh/h)/kWh]

6,80

100

Total E savings [%]

4,53

15%

0

0% 2,27

150

200

Total E savings [%]

15%

20%

[kWh/h]

150

4 ESS (every 2000m)

200

Total E savings [%]

20%

[kWh/h]

200

Total E savings [%]

[kWh/h]

1 ESS at d=7000m

2,27 4,53 6,80 9,06 Energy savings


Figure 9. Simulation results of different scenarios at off-peak periods.

5%

5% 2,27

9,06

20% 15%

50

10% 5%

0

0% 6,80

9,06


20%

100

15% 10%

50

5% 0

30

4,53

6,80

9,06

Module capacity (kWh) Energy savings


Energy savings

1 ESS at 7000 2 ESS at 2000 & 6000

20

4 ESS every 2000m

10

6 ESS every 1500m

0% 2,27

9,06


30% 25%

[kWh/h]

100

Total E savings [%]

[kWh/h]

25%

6,80


150

30%

0% 4,53


Energy savings

8 ESS (every 1000m)

6 ESS (every 1500m)

10%

2,27

9,06


Energy savings

150

4,53

6,80



2,27

4,53

15% 50

5%

[(kWh/h)/kWh]

6,80

20%

100

0

0%

Total E savings [%]

4,53

10%

50

0

0%

0

15%

25% [kWh/h]

10%

50

20%

100

Total E savings [%]

15%

[kWh/h]

20%

100

25%

30%

150

30%

150 Total E savings [%]

[kWh/h]

25%

2,27

4 ESS (every 2000m)

2 ESS at d= 2000 & 6000m 30%

Total E savings [%]

1 ESS at d=7000m 150

8 ESS every 1000m

0 2,27 4,53 6,80 9,06

Energy savings


Figure 10. Results for simulations of different scenarios at metro night and weekend periods.

better performance when installed in centered positions of the line due to the easier access to the few trains circulating in both sides, ahead and back. The performance of the modules is worse than in the previous scenarios despite the little regeneration rate of

conventional vehicles. The problem here is due to the low traffic density. Although the savings achieved are high in proportion to the energy delivered, the energy flow is not as high as in the other scenarios. That is why the modules


savings are lower than 30 kWh/h when 8 ESS are installed on the line. In this case the vehicles energy recuperation rates goes from 7% in a conventional line (see Table 4) to 44% with 8 ESS on the line. V. CONCLUSIONS The results obtained show the benefits, in terms of energy savings, of using stationary ESS on a metro line. On the positive side, savings between 11% (at peak time) and 26% (at night and weekends) can be achieved depending on the scenario. The energy saved per module depends on the traffic conditions, its size and positioning. Their savings can go from 150kWh/h when only a large size module is positioned at the end of the line at off-peak times, to 30kWh/h saved by one small module when 8 ESS are on the line at night time. It is not straightforward to choose the best option in all circumstances, but considering that saving energy is the primary goal, it can be concluded that at least 1 ESS every 2000 m is required. The distribution of 1 ESS every 1500 m also seems fair. Regarding the module size, if the ESSs modules are spread every 2000 m or 1500 m, the small (2.27 kWh) or medium size (4,53 kWh) modules are the best compromise considering the results obtained and the amount of cells installed. Evaluating the modules ‘efficiency’ in terms of energy saved/energy installed, the small size module would be the best option. However the higher savings achieved by the medium size module will lead to better results in a long term. Since not only technical reasons are used to evaluate the best option, an economic analysis of the energy saved versus the cost of the system will be helpful in the decision process. This analysis will be further studied in the course of this project. A control strategy to optimize the energy savings has been presented. In order to obtain a good performance of the stationary ESS under low traffic conditions, it is important that the ESS is able to rapidly discharge itself as soon as possible once it is full. The difficulty of that is detecting the accelerating vehicles. The voltage drop caused by a distant accelerating vehicle at the ESS module is rather small. An ‘aggressive’ controller, in the sense that it tries to discharge when little voltage drops occur, such as the one presented, will help improving the energy savings. When optimizing the stationary ESS for energy savings, the line voltage is not markedly improved and the line losses remain almost the same, unlike the case of on-board ESS, where line losses are reduced and voltage stability is improved. ACKNOWLEDGMENT This research is financed by the Institute for the Encouragement of the Scientific Research and Innovation in Brussels (IWOIB / IRSIB). REFERENCES

[1]

[2]

[3]

[4] [5]

[6]

[7] [8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

Van Mierlo J, Maggetto G, et al., “How to Define Clean Vehicles? Environmental Impact Rating of Vehicles”, International Journal of Automotive Technology (IJAT), KSAE,SAE, ISBN 1229-9138, Vol 4, Nr 2, Pg 77-86, 2003 Van Mierlo J, Timmermans J.-M. et al., “Environmental Rating of vehicles with different alternative fuels and drive trains: a comparison of two approaches,” in Transportation Research Part D, vol. 9, 2004, pp. 387-399, 2004. Chan C.C., ‘The State of the Art of Electric, Hybrid, and Fuel Cell Vehicles’, Proceedings of the IEEE, Volume 95, Issue 4, April 2007, Page(s):704 – 718. Gunselmann W, ‘Technologies for Increased Energy Efficiency in Railway Systems’, in proceedings of EPE 2005, Dresden. ISBN: 90-75815-08-5. Steiner M., Scholten J., Khlor M.,‘Energy Storage on Board of Railway Vehicles’, in Proceedings of ESSCAP’2006. Lausanne, Switzerland. November 2006. Burke A, ‘Ultracapacitors: why, how, and where is the technology’, Journal of power sources. Vol 91 (1), pg 37-50. NOV 2000. Auer J, Sartorelli G, Miller J.M., ‘Ultracapacitors – improving energy storage for hybrid vehicles’, in Proceedings of EET 2007, Brussels, Belgium. Shen WX, Chan CC, Lo EWC, Chau KT, ‘Estimation of battery available capacity under variable discharge currents’, JOURNAL OF POWER SOURCES, Volume: 103, Issue: 2, Pages: 180-187, Published: 2002. Rufer A, Hotelier D, Barrade P, ‘A supercapacitor-based energy storage substation for voltage compensation in weak transportation networks’, IEEE Transactions on power delivery, vol. 19 (2), pg 629-636. APR 2004 Rufer A, Barrade P, Hotellier D, Hauser S, ‘Sequential supply for electrical transportation vehicles: Properties of the fast energy transfer between supercapacitive tanks’, Journal of circuits systems and computers. vol 13 (4), pg 941-955. AUG 2004. Steiner M, Klhor M, Pagiela S., ‘Energy Storage System with UltraCaps on Board of Railway Vehicles’, in Proceedings of EPE 2007, Aalborg, Denmark. Destraz B., Barrade P., Rufer A., M. Klohr, ‘Study and Simulation of the Energy Balance of an Urban Transportation Network’, in proceedings of EPE 2007 Aalborg Foiadelli F., Roscia M., Zaninelli D., ‘Optimization of storage devices for regenerative braking energy in subway systems’, Power Engineering Society General Meeting, 2006. IEEE Volume , Issue , 18-22 June 2006 Page(s): 6 pp. Van Mierlo J, Maggeto G, ‘Innovative iteration algorithm for a vehicle simulation program’, IEEE transactions on vehicular technology. vol 53 (2). pg 401-412. MAR 2004 Van Mierlo J, Van den Bossche P, Maggetto G, ‘Models of energy sources for EV and HEV: fuel cells, batteries, ultracapacitors, flywheels and engine-generators’, Journal of power sources. Vol. 128 (1), pg. 76-89, MAR 29 2004 Barrero R., Tackoen X., Van Mierlo J., ‘Quasi-static simulation method for evaluation of energy consumption in hybrid light rail vehicles’, Accepted for proceedings of VPPC 2008. 3-5 September, Harbin (China). Barrero R., Tackoen X., Van Mierlo J., ‘Supercapacitors onboard light rail vehicles: enhanced energy storage systems for improved vehicle efficiency’, in Proceedings of IEEEASME Joint Rail Conference 2008. Wilmington (USA).