Control of urban rail transit equipped with ground-based supercapacitor for energy saving and reduction of power peak demand

Electrical Power and Energy Systems 67 (2015) 439–447

Contents lists available at ScienceDirect

Electrical Power and Energy Systems journal homepage: www.elsevier.com/locate/ijepes

Control of urban rail transit equipped with ground-based supercapacitor for energy saving and reduction of power peak demand ZongYu Gao a,b,⇑, JianJun Fang a, YiNong Zhang a, Lan Jiang a, Di Sun a, Wenrong Guo a a b

College of Automation Beijing Union University Beijing, 100101, China School of Electrical Engineering Beijing Jiaotong University Beijing, 100144, China

a r t i c l e

i n f o

Article history: Received 1 July 2013 Received in revised form 13 October 2014 Accepted 21 November 2014 Available online 24 December 2014 Keywords: Control strategy Supercapacitor Energy storage system Metro trains Regenerative braking

a b s t r a c t An energy storage system based on Supercapacitor (SC) for metro network regenerative braking energy is investigated. The control strategy according to the various power requirements in metro line and differing characteristics of these storage devices are proposed to manage the energy and optimize the power supply system performance. In order to estimate the required energy storage system (ESS), line 5 of Beijing metro network is modeled through a novel approach, in different running interval conditions based on the real data obtained from Beijing metro office. A useful method is proposed to predict the instantaneous regenerative energy which is delivered to each substation before applying ESS and based on that the ESS configuration for each substation is determined. A simplified mathematical model of the whole metro network has been developed and the main features of the control strategy have been developed. Numerical simulations show the efficacy of suggested control and the energy saving obtained for metro trains. Ó 2014 Elsevier Ltd. All rights reserved.

Introduction Urban rail transit has some advantages such as a large capacity, timing, safety, environmental protection and energy saving i.e. when environmental protection was advocated in the world, to prevent the greenhouse effect today, the advantages of energy-saving, environmental protection of urban rail transit more and more get the attention of people. For the city railway train, regenerative braking mode is the main way of braking. Train regenerative braking energy can supply other train traction state using the same power supply area, so as to reduce the energy consumption of train. But when the train regenerative braking reduces the reliability and security of the urban rail transit, there may be renewable failure situation. In recent years, flywheel, battery, super capacitor energy storage device, such as solution regeneration failure of the practical application of growing, how to use energy storage device to solve this problem is becoming more and more urgent. SC energy storage compared to other energy storage method has long cycle life, high power density and no pollution to the environment, etc., has gradually been widely used. The time of charge and discharge of SC is short, the urban rail transit operation is frequent start–stop and voltage peak obvious fluctuate, and this is a ⇑ Corresponding author at: College of Automation Beijing Union University Beijing, 100101, China. Tel: +86 13426444302; fax: +86 10 64900513. E-mail address: [email protected] (Z. Gao). http://dx.doi.org/10.1016/j.ijepes.2014.11.019 0142-0615/Ó 2014 Elsevier Ltd. All rights reserved.

very good fit SC and therefore SC is an important choice for energy storage components in the area of urban rail transit. SC is a further popularized application. Paper [1] urban rail power supply on Pscad environment, network model, regenerative current value on each substation simulation has carried on the super capacitor capacity in accordance with the configuration and the economic evaluation; Paper [2] in the Matlab environment for urban rail traction power supply simulation, and respectively under different control strategies for vehicular and ground type SC energy storage system is analyzed in capacity configuration and energy saving. In [3] and [4], electrical trains have been considered as a useful public transportation that their efficiencies can be improved by applying the ESS; however, ESS sizing and network modeling have not been discussed in these references. In [5], different mechanical and electrical techniques have been overviewed in order to improve the energy efficiency in electrical railway systems. Some investigations have been done about the advantages of the onboard ESS in both electrical [6–11] and diesel trains [12–14]. Advantages of using different ESS for both onboard and stationary systems have been represented in [15], but the control algorithm, optimal positioning and ESS sizing have not been discussed. Quasi-static backwards looking method has been used for simulation of energy consumption of the vehicles in [16–21]. In [22–24], stationary ESS has been applied to save the regenerative energy. Stationary ESS has been proposed for voltage

Z. Gao et al. / Electrical Power and Energy Systems 67 (2015) 439–447

regulation of weak points in [25]. But, the metro network has not been modeled and the algorithm of ESS sizing has not been presented. Maximum regenerative energy of each station depends on the energy consumption of the auxiliary equipment, resistive forces, traffic conditions, and energy exchange between trains. Because of these difficulties, references [24–28] have considered several ESS with different capacitances and made several trialand-error simulations to find the best configuration with the highest energy saving capability. In [22–24], the utilized power flow controller for ESSs is just based on the voltage variation of the supply line. In this paper, at first the regenerative current is analyzed, then the power flow controller is designed based on the voltage variation of the line, the current variation of electric line, and the maximum permissible charging current of the ESS. Since the current of ESS is under control, it is more reliable. In this paper, the metro supply network and metro trains are modeled using real data obtained from Beijing line-5 metro office. The model shows the behavior of the metro line, trains, stationary ESS, and irreversible substations. The network model is simulated in the digital simulation environment of matlab software. In comparison with previous modeling methods presented in Refs. [13–16], the proposed approach presents a good physical insight into the network model. Moreover, it can be extended easily. Unlike [13–16] which use trial and error method to find the best ESS configuration with the highest energy saving, in this paper, an effective method is proposed to calculate the maximum instantaneous regenerative energy of each station analytically. Then, appropriate ESS configuration is suggested for each station.

Traction substation Pnet

Power Mains

ESS

SC

SC

Charge Discharge Udis

Uchar

Usub

DC/DC

DC/DC

Charge

Discharge

Braking

Traction

ESS: Energy Storage System Fig. 1. Vehicle network traction characteristic.

substation

0.02 m

R 1km

substation

Grid voltage fluctuation

L 0.15mh 1km

Station1 train

DC grid

Station2

DC grid voltage of traction DC grid voltage of brake

A

0km

B

C

2.17km

:substation

Characteristics of network and control strategy

D

5.85km 7. 25 km

:Train(powering)

:Train(coasting)

bs ta tio n su

bs ta tio n tra in

su

bs ta tio n su

sta t tra ion in

upline

su bs ta tio n tra in

Fig. 2. Voltage fluctuation of DC system.

System arrangement This study is applied to the line-5 of Beijing metro network. The total length of the line is about 42 km. It connects the south west of Beijing to the north east of it with 24 stations.

Traction substation

Power Mains

ESS

sta t tra ion in

440

E

9. 39 km

F

downline

11.68km

:train(braking)

Fig. 3. Sketch of a domestic metro line.

Fig. 1 shows the constitute of railway vehicle with power supply, including ESS, power mains, discharge and charge. From this figure, we know the flow direction of current with charge and discharger, and know the character of SC control curve. Motor control states can be divided into three modes. When powering, the traction motor absorbs energy from the feeder line, leading to the voltage drop; when braking, the motor as generator feeds energy to the feeder, leading to the voltage rise. The voltage fluctuation significantly affects the characteristic of the train running [20]. In this paper, the data used for the study are based on the real measurements of Beijing metro line 5 and are demonstrated in Fig. 2. The figure shows the details of driving cycle between two stations. Fig. 3 shows the running character of the whole line.

The maximum speed is 80 km/h during acceleration and the maximum acceleration is 1 m/s2. The railway transit network model includes trains, unidirectional substations, ESS, and connecting lines, as shown in Fig. 4. Substations are modeled as ideal DC voltage sources. The connecting lines are modeled as electric resistances. Since the trains are moving between these stations, the resistance among the train, the starting station and the next station is time variant. Therefore, for each time point, these values were calculated as

R0 ¼ k  xðtÞ=1000

ð1Þ

R00 ¼ k  ðd  xðtÞÞ=1000

ð2Þ

Fig. 4. Modeling of the metro network.

441

Z. Gao et al. / Electrical Power and Energy Systems 67 (2015) 439–447

where R0 is the line resistance between the train and the last station (X/km); R00 is the grid resistance between the train and the next station (X/km); k is the resistive coefficient, the value is set to 0.019 O/km; d is the distance between the start and next stations. x(t) is the distance between the train and start station. In simulation environment, the value of resistors can change during the simulation. The network model includes all 24 stations of the line 5 as a sample. Driving cycle used for this study is based on the real measurements and is demonstrated in Fig. 2. This figure shows the details of driving cycle between two stations. Fig. 3 shows the running character of whole line. The maximum speed is 80 km/h during acceleration and the maximum acceleration is 1 m/s2. Line parameters and the efficiency of different components used in the simulations are given in Tables 1 and 2 respectively. Traction computing model (TPS) This model includes route condition, train parameter and run map message. This route condition includes gradient and curve i.e. train parameter is the load of train, auxiliary power, traction and braking curve, run map provided station distribute message, run and stop times of the train. Through this model, the speed,

the distance and the power can be obtained. This flowchart is defined in Fig. 5. Train and substation model A train is modeled by a controlled current source which draws active power at the accelerating time or delivers active power at the regenerating time. In this model, set the regenerative current limits, it can limit the regenerative current of pantograph. To compute the train power, the resistive forces are calculated by the formula proposed in Ref. [18]. Fig. 6 is the train model, the line resistance is decided by the distance of two components. Rf is filter resistance of train, Lf is filter inductance of train, Ufc is filter capacitor of train, Pedlc is discharge power or charge power of SC, Paux is auxiliary supply power, P is the power of train, and is determined by the following equation:

  Iin ðU fc  U in Þ  RðI þ Iin Þ  Rf I  L dt

dI 1 ¼ dt L þ Lf

dU fc 1 ¼ ðI  Iinv  Pedlc =U fc  Paux =U fc Þ Cf dt U out ¼ U in þ RIout þ L

Table 1 Line parameters.

ð3Þ

ð4Þ

dIout dt

ð5Þ

Iout ¼ Iin þ I

Substation dc bus voltage (V) Rail electric resistance (mX/km) Train gand-up Load operating mode The control mode Drive motor Train datum drag

1500 33 3M3T AW2:274.44t VVVF inverter 1C2M mode Asynchronous motor (rated 180 kW) ⁄ 4 R = (331.6 + 3.65V + 0.067V2)  9.8 (N) R: train resistance, V: train speed (km/h)

Table 2 Component efficiency. Gear transmission efficiency (%) Motor efficiency (%) Motor drive efficiency (%) DC/DC converter efficiency (%) Substation floating voltage Steel rail resistance Substation equivalent internal resistance Pantograph resistance Contact system resistance Run mode

93 90 90 91 825 V 0.009 X/km 0.07 X 0.015 X 0.003 X/km Traction–cruise–coasting– braking

ð6Þ

Traction substation is uncontrolled rectification, current is unidirectional. Fig. 7 is the model, switch S1 is closed when substation output current is in forward direction. Switch S1 is disconnected when substation output current is negative. In order to model SC charge state, substation parallel connection a braking resistor, braking resistor series connection a switch S2 for control voltage. Switch S2 closed during Vsub P Uchar, analog SC charge. Switch S2 disconnected during Vsub 6 Uchar, analog SC stopping charge. Through control the closed and open of S2, the remaining energy flowing is controlled between SC with Dc line network. Line network voltage is uprised when direct current (Dc) line network Uout

Iin

Uin Iout

Line Rs Rf

0

Pedlc /Ufc Paux /Ufc Iinv

U2 Umax Ufc

Ufc

Lf

Ufc

U1

-I max Regenerative current limitator

Cf

P/Ufc

Iref

Fig. 6. The train model.

SC

Idin +Iuin

Uout

Vsub

Isub

Uuin Iuout

Line Rs Z Voltage (V)

Rs + Us _

substation Current (A)

Fig. 5. TPS model.

Fig. 7. Traction substation model with SC.

442

Z. Gao et al. / Electrical Power and Energy Systems 67 (2015) 439–447

emerged renewable brake energy, the energy is absorbed by adjacent train when line network voltage is under Uchar. Part of renewable energy is absorbed by adjacent train when line network voltage exceeded Uchar, the remaining renewable energy is absorbed by SC. Through measurement Vsub and Isub, Ir is renewable current when Isub is negative, the remaining renewable power of substation is Pr = Uout  Ire, remaining renewable energy is R Er = Prdt, Pr and Er is important reason for the SC configuration. Control strategy In the wayside SC energy storage system (SESS) control, the charging and discharging currents are given according to the

Fig. 8. Power distribution control strategy.

changes of the feeder line voltage. When grid network is high voltage, SC absorbs current; when low voltage, SC releases energy to DC link. Not to consider the weight and size restrictions of SC, the large capacity for the SC can be chosen to stabilize the whole line network voltage. Energy management of wayside SC system gives more consideration to network. When the train is accelerating, the energy system outputs energy to suppress the voltage fluctuate of the pantograph voltage fall and improve the acceleration characteristics. When train braking, the system absorbs braking energy to prevent the voltage fluctuate of the pantograph and at the same time preventing the regeneration failure, and improve the electrical braking performance. Based on the power distribution curve between the SC and the substation, this paper proposes a control strategy for the energy storage system, as shown in Fig. 8. As shown in Fig. 8, Udc⁄ is given through the line current. In detail Udc⁄ is given through comparing the line current with a reference line current. Through think about the logical conditions, SC energy storage system can take actions according to the energy management strategy designed. A current loop is added and many benefits are gained. Firstly, it can avoid unwanted inflow and outflow between the SC and the feeder line. Secondly, it can decide the voltage dynamic value and the static value. Thirdly, it can decide the power distribution of the power between the feeder line and the SC. The control algorithm flowchart is shown in Fig. 9. In the control strategy, when SOC < 0.9, charge through constant high power; when SOC = 0.9, reduce the charging power; when SOC = 1, stop charging during discharging; when SOC > 0.5, discharge through constant high power; when 0.5 > SOC > 0.25, reduce the discharging power; when SOC = 0.25, stop discharging. Through SOC loop control, SC SOC value is kept in the range from 0.25 to 1. Mathematical simulations In this part, several simulations are carried out for the study described in previous parts. The matlab and dSPACE software is used to perform the investigation. Simulation parameters For test the feasibility of the control strategy, a model of substation installation SC is built by Matlab/Simulink software. Through this model, the whole working conditions of the system are considered. The model included powering, coasting and braking conditions. Table 3 is the parameters of simulation platform and Table 4 is the parameters of the SC. Simulation platform The new railway transit network with SC included three parts: urban rail vehicles; traction substation; SC energy storage system (SESS); SESS model structure is shown in Fig. 10. SESS includes three parts: input, output and simulation operation. Simulation operation includes train traction computing, power flow calculation and SC energy storage models. Table 3 Parameters of simulation platform.

Fig. 9. Control algorithm diagram of wayside SC storage system.

Rate power DC-link voltage Powering voltage action range Braking voltage action range Train gand-up Control mode Steel rail resistance Pantograph resistance

2000 kW 1500 V 1100–1300 V 1600–1800 V 3M3T VVVF inverter 1C2M 0.009 X/km 0.015 X

443

Z. Gao et al. / Electrical Power and Energy Systems 67 (2015) 439–447 Table 4 Parameters of SC.

Voltage (V) Current (A)

Cell Total capacity Internal resistance Energy Maximum voltage

Braking action point voltage Braking action voltage renge

Iv Vehicle current

Vmax = 2.5 V 48 F 0.2 X 5.1 kW h 1000 V

This traction computing model includes route condition, train parameter and run map message. The route condition includes gradient and curve, i.e., substation parameters are the load of train, auxiliary power, traction and braking curve, run map provided station distribute message, run and stop times of the train. Through the model, the speed, the distance and the power can been obtained. Simulation results The variation of line currents and line voltage during the whole operation is shown in Fig. 11. When the train is accelerating, the DC line voltage drops. When the line current exceeds the limit volume and the voltage drops below the action value 1300 V, the SC releases energy to the line network to reduce the line current and contain the voltage drop at the pantograph. Finally, the DC link voltage maintains at 1200 V. When the vehicle is coasting, SC energy storage system is in standby mode. When the vehicle is braking, the DC line voltage rises. When the braking current exceeds the limit volume and the line voltage rises above the action value 1700 V, the SC absorbs regenerative braking current to contain the voltage rise at the pantograph and prevent the regeneration failure.

2000

Il Line network current

1000 Traction action point voltage Traction action voltage range Udc DC network voltage

0 Is bidirection DC/DC Port voltage

-1000 0

10

30

40

50

60

70

80

Fig. 11. Line currents and line voltages of substation.

realized by a PWM converter, whose current is fed back to the grid through LCL filter. PWM converter uses grid voltage oriented control method. The reference value Id is calculated according to the characteristics of the load. When the Id > 0, the PWM converter operates in rectifier state; when Id < 0, the PWM converter operates in inverter state. The SC bank is the product of Beijing Supreme Power Systems Co. Ltd. Parameters are as follows: rate voltage 320 V, capacity 1.5 F and internal resistance 2.75 X. The picture of experimental platform is shown in Fig. 13. Simulation results Analyzing the brake energy and SOC The instantaneous value of regenerative braking power can be calculated based on the characteristic curve equation of train regenerative braking [26] and equation P = Be  V, C1, C2 is constant. Suppose CD is nature characteristic area

Result of experiment

PCD ¼ BeCD  V ¼ Experimental platform

20

C2 V

ð7Þ

Suppose AC is constant torque area

Based on the laboratory conditions, a 3 kW SC experimental platform is built. Fig. 12 shows the circuit diagram of the experimental platform. The platform consists of substation simulation system, load simulation system, and SC energy storage system. The substation simulation system converts 210 V AC into 300 V DC through the diode rectifier. 210 V AC is turned into 380 V AC through an auto-regulator and a three-phase isolation transformer (ensure the vehicle simulation system and substation systems connecting to the grid at the same time). Load simulation system is

Traction / Braking curve

Line conditions Limit speed

Input parameter Initial voltage

SC control strategy SESS

ð8Þ

Train decelerated motion (constant deceleration) from initial velocity V0 beginning

V 2  V 20 ¼ 2aS

ð9Þ

where V is the instantaneous velocity of train, V0 is the initial velocity (5 km/h) of train, a is the acceleration of train, S is the displacement of train. Differentiating to (9) two side

Traction grid voltage constrains renewable current a-t s-t TPS p-t

Running time

SOC Limit

PAC ¼ BeAC  V ¼ C 1  V

Pantograph voltage Pantograph current Substation output power

DC - RLS

Pc

Remaining renewable power/ energy

Simulation output

……

Substation position Substation characteristic

Charge/discharge renewable power/energy SESS : SC Engery Storage Simulatior

TPS : Train Performance Simulatior DC - RLS: DC - Railway Loadflow Simulator Fig. 10. Simulation model.

444

Z. Gao et al. / Electrical Power and Energy Systems 67 (2015) 439–447

Line impedance L1

R1

YD11

Il

Id

L2

IS

LB

LA C1

Isolation AC power transformer AC210V B1

R3

DC 300V C2

T3

R2

CB

Rd rectifier

Converter

L3

Substation simulation system

T1

C3

I SC

Lsc T2

Usc

r 270V EDLC

Bidirectional DC-DC transducer SC ESS Fig. 12. Block diagram of 3 kW SC platform.

The capacitor state of charge (SOC) is a function as follows:

SOC ¼

ESC ESCmax

1 CU 2 ¼ 1 2 2 SC ¼ CU SCmax 2



U SC U SCmax

2 ð17Þ

During traction, SC provided energy is

  1 1  C  U 2SC1  U 2SC2 ¼  6:6  ð1882  1562 Þ 2 2 ¼ 36:3kJ

W SC-rel ¼ Fig. 13. Prototype of 3 kW SC platform.

During braking, SC absorbed energy is

V dV a

W SC-abs ¼

dS dt

ð11Þ

V Vdt ¼ dV a

ð12Þ

V¼

In different characteristic area of regenerative braking curve, the regenerative energy of train producing is defined by the following equation:

EAC ¼

Z

dEAC ¼

Z

Z

PAC dt ¼

Z

V2

PAB dt ¼

V1

¼

Z

V2

V1

ECD ¼

Z

BeV dV ¼ a1

dECD ¼

Z

¼

V3

V2

V2

BeVdt

V1

  C 1 V 22  V 21

ð13Þ

PCD dt ¼

Z

Z

V3

PCD dt ¼

C 2 dV C 2 ln ¼ a2 a2 V

BeVdt

V3 V2

Ee ¼ EAC þ ECD

ð15Þ

The total energy is defined by the following equation during train braking:

E¼

dEAD ¼

Z

PAD dt ¼

Z 0

After added ESS, the energy consumption of train running is as follows:

ð21Þ

So, the energy conservation of train running is

save energy 69:1  54 ¼ ¼ 21:8% energy dissipation of no SC 69:1

ð22Þ

ð14Þ

where V1 is snapback speed from the constant torque range to the constant power range, V2 is snapback speed from the constant power range to the characteristic range, V3 is top speed of train. a1 is the instantaneous deceleration of constant torque area, a2 is constant power area. The regenerative total energy of every station is defined by the following equation during train braking:

Z

ð20Þ

Through the calculation, we obtained that the energy consumption of train running is dropping to 21.8%.

V2

 

W no-SC ¼ 75:5  6:4 ¼ 69:1 kJ

g¼

V3

ð19Þ

Through the later simulation known Wtrac = 75.5 kJ, grid provided energy 39.2 kJ. The feedback energy of regenerative brake is Wbrake = 21.5 kJ, grid absorbed energy 6.4 kJ. During no ESS, the energy consumption of train running is as follows:

W add-SC ¼ 75:5  21:5 ¼ 54 kJ

2a1

V2

Z

  1 1  C  U 2SC4  U 2SC3 ¼  6:6  ð1702  1562 Þ 2 2 ¼ 15:1kJ

ð10Þ

V3

F b Vdt ¼

Z 0

V3

FbV dV a

Grid voltage (V)/Grid current (A)

dS ¼

ð18Þ

2000

Grid current(A)

1500

Grid voltage(V)

a

b

Exp renewable current(A)

1000 500 0 -500 -1000 -1500

Remaining renewable current

400

500

600

700

800

900

1000

Time (S)

ð16Þ Fig. 14. The voltage and current in pantograph of train.

1100

445

2000 1500 1000 500 0 -500 -1000 -1500

Grid current Grid voltage Exp renewable current

300

400

500

600

700

800

900

Remaining renewable energy(kWh)/power(kw)

2000 1500 1000 500 0 -500 -1000 -1500

1000

1000

Pscmax

0 -1000 1 Escmax

0 -1

300

400

500

600

700

800

900

1000

Time (S)

Time (S)

(a) train pantograph voltage and current of upline

(b) layoutplan of upline remaining power and energy

Grid current Grid voltage Exp renewable current

300

400

500

600

700

800

900

Remaining renewable energy(kWh)/power(kw)

Grid voltage (V)/Grid current (A)

Grid voltage (V)/Grid current (A)

Z. Gao et al. / Electrical Power and Energy Systems 67 (2015) 439–447

1000 Pscmax

0 -1000 2 0

Escmax

-2

1000

300

400

500

600

700

800

900

1000

Time (S)

Time (S)

(c) train pantograph voltage and current of downline

(d) layoutplan of downline remaining power and energy

2000

Grid current

1500

Grid voltage Exp renewable current

1000 500 0 -500 -1000 -1500

400

500

600

700

800

900

1000

Remaining renewable energy(kWh)/power(kw)

Grid voltage (V)/Grid current (A)

Fig. 15. Multi-trains running simulation results when headway 270 s.

1100

1500 1000 500 0 -500 4 3 2 1 0 -1

2000

Grid current

1500

Grid voltage Exp renewable current

1000 500 0 -500 -1000 -1500

400

500

600

700

800

900

1000

1100

Escmax

400

500

600

1500 1000 500 0 -500

800

900

1000

1100

6 4 2 0 -2

Pscmax

Escmax

400

500

600

Time (S) (c) train pantograph voltage and current of downline

700

Time (S) (b) layoutplan of upline remaining power and energy Remaining renewable energy(kWh)/power(kw)

Grid voltage (V)/Grid current (A)

Time (S) (a) train pantograph voltage and current of upline

Pscmax

700

800

900

1000

1100

Time (S) (d) layoutplan of downline remaining power and energy

Fig. 16. Multi-trains running simulation results when headway 360 s.

Fig. 14 is the simulation result of pantograph current and voltage for train, through the ‘‘a’’ range of Fig. 13 we obtained that the pantograph voltage is restrained in the value of 1750 V during renewable braking, the remaining renewable current of SC absorbing is the D-value of expecting renewable current (red1 line) and grid current (green line). The ‘‘b’’ shows that the train is in the end 1 For interpretation of color in Fig. 13, the reader is referred to the web version of this article.

area of traction and be entering into coasting. Right now, only the auxiliary power part got energy through Dc supply network and energy consumption is low. But other train is in the area of renewable brake and the catenary voltage would rise because of the shortage of renewable load. Figs. 15–17 are the simulation result of different departure intervals for multiple trains, these figures shown the result for the current and voltage of train, pantograph, and remaining renewable energy and power when the Dc Supply network is running.

Grid current

1500

Grid voltage Exp renewable current

1000 500 0 -500 -1000 -1500

500

600

700

800

900

1000

1100

1200

Pscmax

10 5 Escmax

0 -5

500

600

700

800

900

1000

1100 1200

Time (S)

(a) train pantograph voltage and current of upline

(b) layoutplan of upline remaining power and energy

2000

Grid current

1500

Grid voltage Exp renewable current

1000 500 0 -500 -1000 -1500

1500 1000 500 0 -500

Time (S)

500

600

700

800

900

1000

1100

1200

Remaining renewable energy (kWh)/power (kw)

Grid voltage (V)/Grid current (A)

2000

Remaining renewable energy (kWh)/power (kw)

Z. Gao et al. / Electrical Power and Energy Systems 67 (2015) 439–447

Grid voltage (V)/Grid current (A)

446

2000 Pscmax

1000 0 -1000 20 10 0 -10

Escmax

500

600

700

800

900

1000

1100

1200

Time (S)

Time (S)

(c) train pantograph voltage and current of downline

(d) layoutplan of downline remaining power and energy

Fig. 17. Multi-trains running simulation results when headway 450 s.

Because the design goal of SC is to absorb remaining renewable energy, so the capacity of SC is decided by the biggest remaining renewable energy and power is decided by the biggest remaining power. Analyzing the train remaining renewable energy and power Through previous simulation result, the remaining renewable energy and power can be obtained for different departure intervals. Figs. 18 and 19 show that SC needs to absorb the smaller energy and power when the departure intervals are shorter. At the same time, capacity configuration not only considers the highest speed but also needs to consider multiple factors.

Proposed ESS configuration Fig. 18. The surplus regenerative power of train in different indeparture intervals.

The remaining renewable energy and power can be obtained through simulation result, using BMOD0063P125 configuration capacity by series-parallel combination. Two constraint conditions need to consider when configuration SC banks: (1) Because SC self-characteristics that larger changes internal resistance and voltage of charge and discharge, this would impacted the work efficiency of SC. The available energy of SC is 25% of former when the minimum discharge depth of SC is half of the rated voltage. (2) Because SC energy storage device and the supply network of Dc 1500V in parallel, the DC/DC convertor work in Boost state when SC discharge to Dc supply network, the Boost circuit step-up ratio is ‘‘3’’ based on references, and this time the circuit is stable, so the minimum working voltage is 500 V (1500/3).

Fig. 19. The surplus regenerative energy of train in different indeparture intervals.

Through previous two constraint conditions, the series and parallel configuration results of different departure intervals are shown in Table 5.

447

Z. Gao et al. / Electrical Power and Energy Systems 67 (2015) 439–447 Table 5 The initial capacity setting in different departure intervals.

Compound mode Working voltage (V) Weight (kg) Volume (m3) Storage

Braking to stop for one time

Headway 270 s

Headway 360 s

Headway 450 s

12 series, 4 parallel and 2 banks 750–1500 5566 6.444 9

7 series and 2 banks 500–875 847 0.98 1.04

8 series, 2 parallel and 2 banks 500–1000 1936 2.24 2.6

10 series, 2 parallel and 2 banks 625–1250 2904 3.36 3.5

Conclusion In this paper, the Beijing line-5 metro supply network and trains were modeled. For the study, real data of metro line and trains were obtained from metro office. An efficient algorithm was proposed to predict the maximum instantaneous regenerative energy of each station. To save the maximum instantaneous regenerative energy in each station, 3000 (F), 2.7 V SC was used. Appropriate ESS configurations were proposed for different departure intervals. The investigation shows that during the peak period, the regenerative energy saves about 500 MW h/year for different time intervals. During the off-peak period, it saves about 300 MW h/year for different time intervals. The maximum energy saving is around 21.8% at headway 270 s and 18.6% at headway 360 s. In this study, a noteworthy benefit/cost analysis was applied to whole line and it is found that the suggested ESS will save 1500 MW h energy per year. The Daily energy saving is 12%. Acknowledgment The authors would like to thank Beijing Education Commission Project, ‘‘train system design boundary technology research’’ Year: 2014–2015 (Grant Number: 12213994701/008). References [1] Teymourfar Reza, Asaei Behzad. Stationary super-capacitor energy storage system to save regenerative braking energy in a metro line. Energy Convers Manage 2012;56:206–14. [2] Iannuzzi D, Murolo F, Tricoli P. A sample application of supercapacitor storage system for suburban transit. In: Proc. int. conf. elect. syst. for aircraft, railway and ship propulsion. Bologna (Italy); Oct. 19–21, 2010. p. 1–7. [3] Smith RA. Railways: how they may contribute to a sustainable future. Proc Inst Mech Eng Part F: J Rail Rapid Transit 2003;217(2):243–8. [4] Kamata S. JR East takes up the challenge of ‘searching for a railway that is kinder to the Earth’. Proc Inst Mech Eng Part F: J Rail Rapid Transit 2000;214(2):117–22. [5] Gunselmann W. Technologies for increased energy efficiency in railway systems. In: Proceedings of the EPE; 2005. [6] Destraz B, Barrade P, Rufer A, Klohr M. Study and simulation of the energy balance of an urban transportation network. In: Proceedings of the EPE; 2007. [7] Steiner M, Scholten J. Energy storage on board of railway vehicles. In: Proceedings of the EPE; 2005. [8] Steiner M, Scholten J, Khlor M. Energy storage on board of railway vehicles. In: Proceedings of the ESSCAP; 2006.

[9] Roasto I, Lehtla T, Möllerand T, Rosin A. Control of ultracapacitors energy exchange. In: Proceedings of the EPE-PEMC; 2006. [10] Chymera M, Renfrew AC, Barnes M. Analyzing the potential of energy storage on electrified transit systems. In: Proceedings of the 8th world congress of railway research; 2008. [11] Montanié T. Electric energy storage evaluation for urban rail vehicles. In: Proceedings of the EPE; 2003. [12] Hillmansen S, Roberts C. Energy storage devices in hybrid railway vehicles: a kinematic analysis. Proc Inst Mech Eng Part F: J Rail Rapid Transit 2007;221(1):135–43. [13] Akli1 CR, Roboam X, Sareni B, Jeunesse A. Energy management and sizing of a hybrid locomotive. In: Proceedings of the EPE; 2007. [14] Barrade P, Destraz B, Rufer A. Hybrid vehicle in railways applications: supercapacitive energy storage for diesel-electric locomotives. In: Proceedings of the IEEE – vehicular power and propulsion symposium; 2004. [15] Foiadelli F, Roscia M, Zaninelli D. Optimization of storage devices for regenerative braking energy in subway systems. In: Proceedings of the power eng society general meeting; 2006. p. 18–22. [16] Van Mierlo J, Maggetto G. Vehicle simulation program: a tool to evaluate hybrid power management strategies based on an innovative iteration algorithm. Proc Inst Mech Eng Part D: J Automob Eng 2001;215(9):1043–52. [17] Van Mierlo J, Maggetto G. Innovative iteration algorithm for a vehicle simulation programme. IEEE Trans Veh Technol 2004;53(2):401–12. [18] Mi C, Gao, Emadi A. Modeling and simulation of electric and hybrid vehicles. Proc IEEE 2007;95(4):729–45. [19] Gazella L, Sciarretta A. Vehicle propulsion systems introduction to modeling and optimization. Springer-Verlag Berlin Heidelberg; 2005. [20] Wipke KB, Cuddy MR, Burch SD. ADVISOR 2.1: a userfriendly advanced power train simulation using a combined backward/forward approach. IEEE Trans Veh Technol 1999;48(5):1751–61. [21] Markel T, Brooker A, Hendricks T, Johnson V, Kelly K, Kramer B, et al. ADVISOR: a systems analysis tool for advanced vehicle modeling. J Power Sources 2002;110(2):255–66. [22] Barrero R, Tackoen X, Van Mierlo J. Quasistatic simulation method for evaluation of energy consumption in hybrid light rail vehicles. In: Proceedings of the IEEE vehicle power and propulsion conference; 2008. [23] Barrero R, Tackoen X, Van Mierlo J. Energy savings in public transport. IEEE Veh Technol Mag 2008;3(3):26–36. [24] Barrero R, Tackoen X, Van Mierlo J. Improving energy efficiency in public transport: stationary supercapacitor based energy storage systems for a metro network. In: Vehicle power and propulsion conference; 2008. [25] Zhang Yi-cheng, Wu Lu-lu, Wang Bo, Liang Hai-quan. Modeling and simulation of the 1500 V metro supply network and vehicles. In: Vehicle power and propulsion conference; 2008. [26] Takeshi Konishi, Akinobu Okui, Yoshinobu Nakamichi, Hidetaka Nara, Uemura. Fundamental study on energy storage system for DC electric railway system. In: Power conversion conference. PCC-Osaka; 2002. p. 1456–1459. [27] Narimani Mahmood, Hosseinian Seyed Hossein, Vahidi Behrooz. A modified methodology in electricity tracing problems based on Bialek’s method. Int J Elect Power Energy Syst 2014;2:74–81. [28] Liu Ying, Liu Nian, Xie Chi. Analysis of transient electric field distribution inside the large generator under strong surge voltage. Int J Elect Power Energy Syst 2012;2:121–5.