Professional Documents
Culture Documents
discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/260711741
CITATION
READS
25
1 author:
Vitaly Gelman
vg controls
9 PUBLICATIONS 17 CITATIONS
SEE PROFILE
stockbyte
Vitaly Gelman
70 |||
here has been substantial interest in the traction community for using
wayside energy storage systems (ESSs) to better use train braking
energy, thus reducing energy costs and peak power as well as voltage
stabilization. An alternative solution to reach the same goals is using
recuperating (also called reversible) traction power substations such as
reversible thyristor controlled rectifiers (RTCRs). This article compares
advantages and disadvantages of ESSs of three most common types
flywheels, batteries, and supercapacitorsand recuperating substations. The
analysis takes into account size, capital cost, the round trip energy losses,
and energy savings for heavy rail applications.
Heavy rail trains accumulate substantial mechanical energy during
acceleration. There are two ways to use this energy: 1) using recuperating
traction substations and 2) using ESSs.
General
All modern trains have regenerating capability and can
supply mechanical energy to the dc bus during braking. For a typical ten-car train with M7 cars, traveling at
60 mi/h, the mechanical energy is 234 MJ [1] (1 kWh =
3.6 MJ). Using this energy could provide substantial
economic benefits, and it is possible to get up to 80%
of mechanical e
nergy converted back to dc power at
train terminals. For an M7 car train from the example
above, up to 165 MJ (46 kWh) can be recovered back to
the dc bus.
If there is another train accelerating nearby, it can
absorb the braking energy and reduce the overall system power consumption. Theoretically, we could capture almost all recuperated energy if we could always
have another train accelerating at the same location
and at the same time that our train is braking. However,
this is impossible to achieve in practice. Even if there
is a train that could absorb regenerated energy somewhere in the system, the amount of network-absorbed
power is limited by the rails resistance, proportional to
a distance between the source (braking train) and the
receiver (accelerating) train. Therefore, time and space
distances between breaking and accelerating trains
determine how much braking energy the dc network
can absorb.
With short headway of 2.53 min during peak
hours, there are many trains running close to each
other. Consequently, the braking energy is transferred
efficiently, and the lost energy is small, only 810% of
the total energy.
With long headway of 30 min during off hours, the lost
energy could be up to 60% of the traction power supplied
||| 71
Acceleration (mi/h/s)
Speed (mi/h/10)
Train Power (MW)
Distance (km)
1.5
mi/h/s (km)
mi/h/10 (MW)
22
20
18
16
14
12
10
8
6
4
2
0.5
5 10 15 20 25 30 35 40 45 50 55 60
t
22
18
mi/h/10 (MW)
16
14
1.5
12
10
8
6
0.5
4
2
0
5 10 15 20 25 30 35 40 45 50 55 60
t
mi/h/s (km)
Acceleration (mi/h/s)
Speed (mi/h/10)
Train Power (MW)
Distance (km)
20
Batteries
The oldest energy storage technology is a battery. Batteries provide very high energy density but have limited
absorbed power density and a limited number of
achieving RTE improvement (no losses in the dc/dc converter) and no additional electromagnetic interference.
To avoid overcharging with direct connection, we need
to select fully charged battery voltage corresponding to
+5% incoming ac line voltage; then, under nominal voltage condition, we get an SOC of below 10%. Another option would be monitoring the SOC and disconnecting the
BESS when overcharged.
The Kawasaki Gigacell has been used recently in a
test in New York for wayside rail application. Unfortunately, [4] does not include data on losses, battery
life time, SOC under typical and worst-case conditions,
battery monitoring system, train voltage improvement
under different conditions, and energy savings over a
period of time.
The calendar life of Ni-MH batteries is normally assumed
to be ten years. The battery service life (number of cycles)
increases with decreasing depth of discharge (DOD). By selecting the number of parallel batteries in a BESS for a given
BESS current, we can change the DOD of each battery to
give us a service life close to its calendar life.
With our sizing parameters of absorbed energy (75 MJ)
and power (3 MW), the equivalent charging time for ESSs
is 75/3 = 25 s. With a 25-s pulse and 2-C current, the DOD
is 1.4% (25*2/3,600), and with 5 C current, the DOD is 3.5%
(25*5/3,600).
Kawasaki [7] shows the number of cycles only for
0.33% DOD, but if we use the formula for Ni-MH battery life cycles from [17, p. 16], we get 390,000 cycles
at 3.5% DOD, 1.6 million cycles at 1.4% DOD and 14.7
million cycles at 0.33% DOD. Kawasaki showed 2 million cycles for 0.33% DOD [7], seven times lower than
calculated by the formula. If the ratio (seven) is the
same for 3.5% DOD, the number of cycles will be 55,000
at 5 C current.
The 390,000 cycles might be sufficient for ten years
because the majority of energy during peak hours is
transferred between the trains. Therefore, the charge
discharge cycles are shallow with smaller DOD. If we
need to increase the number of cycles, we should connect more batteries in parallel. This way, at the same
BESS current, we decrease each battery current and its
DOD, thus increasing the number of cycles.
Let us size the BESS using a 30K4 Gigacell battery [7]
with 36 V and 141 Ah as a building block. To determine
the number of parallel connected batteries we assume
the battery current during the charge/discharge cycle
equals the maximum recommended current of 5 C to arrive at a calculated number of 390,000 cycles.
||| 73
battery life. Lower temperature reduces power and energy capabilities of the batteries. Because of the substantial power losses in the batteries, an active cooling
system is often necessary, either air or liquid. Generally,
low temperature is not a problem since the losses in the
battery will increase its temperature, but in cold climates, it might be necessary to provide heating to avoid
the battery freezing.
Supercapacitors
Supercapacitor ESSs (SESSs) from a system standpoint
are similar to BESSs. Their advantages are as follows: a
high number of chargedischarge cycles, a high current
both for charge and discharge, and better RTE. The disadvantage is low energy density, i.e., high cost per MJ
of energy storage. The cost per MJ decreased significantly during recent years, but it appears to have flattened out.
We must allow the supercapacitor voltage to change
during the chargedischarge cycle to use energy storage. To regulate power flow while the supercapacitor
voltage changes, we need a dc/dc converter between
the traction dc bus and the supercapacitor. The bigger
the voltage change, the higher the ratio between the capacitor current and dc bus current and, consequently,
the cost of the dc/dc converter. We need to provide
voltage equalization between series connected supercapacitors as well.
A good compromise is to allow the supercapacitor
voltage to change between 50% and 100% of its rated
voltage. Thus we will use 75% of its energy. To provide
capacitor current control with standard chopper, the
capacitor voltage should be below minimum bus voltage.
As a first approximation, let us assume the minimum dc
bus voltage is 500 Vdc, the train regeneration voltage at the
SESS is 800 Vdc, the regeneration power (charging power
for SESS) is constant 3 MW (800 V, 3.75 kA), and headway
is 5 min. At lower headway, the duty cycle is higher, but
the network absorbs the higher portion of the regeneration
current. Let us select the maximum supercapacitor voltage
of 500 Vdc and the minimum voltage of 250 Vdc, assume
also the supercapacitor charging time is 25 s.
Under those conditions during charging time the
supercapacitor current is (see Supercapacitor RMS
Current Calculation)
I CRMS - chrg = 1.36
Preg
3, 000 kW
= 1.36
= 8.16 kA.
VC max
500 V
2 $ 25 = 8.16 kA
300
1 = 3.33 kA ac.
6
CV 2C max
CV 2C
=
+ Preg t,
2
2#4
where C is supercapacitor capacitance and t is time.
Knowing the capacitor voltage as a function of time we can
find the current I C and calculate the supercapacitor RMS current during the charging process
IC =
1
,
1 + at
4
#
0
Preg 2
m
I 2C (t) dt = 1 c
T VC max
#
0
1
4
1 dt =
+ at
Preg 2
Preg 2
m # du = 4 In (4) c
m
= 1 c
3
aT VC max 1 u
VC max
4
Preg
I CRMS = 1.36
.
VC max
or 3.1%.
Assuming a dc/dc converter efficiency of 98%, we
get a total efficiency of 95% and RTE of 90%. At lower
power levels during peak hours when the current drops
because of network receptivity, the losses will decrease both due to current reduction and an increase
Preg
1
=
VC max
1 + 2Preg t
4 CV 2C max
Preg
Vd I d
=
VC
VC max
||| 75
The supercapacitors, being chemical devices, are sensitive to temperature. Low temperature is not a problemthe supercapacitors are operational to 40 C.
The calendar life of supercapacitors is ten years at
25 C, but it drops to 7.5 years at 40 C and further
drops to just 1,500 h at 65 C [19]. Active cooling systems of either air or liquid are necessary because of
the losses in the supercapacitors and dc/dc converter.
In hot climates, an air conditioner is required to keep
the supercapacitors temperature below 40 C.
A voltage equalization circuit is necessary to provide
equal voltage sharing between series connected modules to prevent overvoltage and subsequent module
destruction due to the capacitance difference
between the modules.
Module voltage and temperature monitoring is needed
to assure SESS reliable service.
Flywheel
A flywheel ESS (FESS) consists of a rotor (made either of
composite material or steel) rotating with a high speed
motor and dc /ac converter (actually dc to variable
frequency converter). To eliminate the energy losses
due to air friction, the rotor is contained in the vacuum
vessel, to reduce friction losses FESS uses magnetic and/
or hydrodynamic bearings. To prevent damage due to
rotor destruction, we need to provide a containment
vessel to keep debris inside. FESS is a complex system
consisting of a dc to variablefrequency converter
(VFC), motor controller, vacuum system, containment
vessel, etc.
To get 20 years of service in traction application with a
large number of cycles, the manufacturer needs to assure
both mechanical and electrical endurance. For mechanical parts, the stress level during cycling must not cause
a fatigue in all mechanical parts, especially in a flywheel.
The cycle number should be verified by testing.
Supercapacitor SESS
(Maxwell) Estimated
Power (MW)
Energy (MJ)
2,200
75
80
75%
90%
83%
Size (m # m # m)
7 # 1.3 # 1.8
15 # 1.6 # 2
14 # 1.5 # 2
Footprint (m )
10
24
21
Weight (t)
30
15
22
10
10
20
1.5
2.6
(1) The cost of a BESS is much higher than those of FESS and SESS, but if we use a hybrid electric battery with a cost of US$1,500/kWh instead of the GigaCell and
allow an additional 50% for a cooling system, the BESS cost drops to US$2.7 million and falls in the range of the other technologies.
TCI is actually a thyristor controlled rectifier (TCR) connected backwards to transfer the energy from the dc bus to
the ac line. The current ratings are about half of the forward rectifier, and this reduces the cost by about 2025%.
||| 77
RTCR Estimated
1.5
12
Efficiency (%)
96
9294, estimated
97, estimated
Size (m # m # m)
2.4 # 1 # 2.3
NA
Footprint (m )
2.4
7.5
5.25
Weight (t)
2.65
NA
NA
20+
20+
20+
0.22
0.32
0.6
0.850.9
0.850.9
No
No
Yes
||| 79
Conclusion
The ESS cost is about US$1 million per MW for all technologies, four to ten times higher than the cost of the reversible
converter. The energy savings achieved by adding ESSs to
diode rectifier substations provide a payback time of well
over 30 years, thus making it economically unfeasible. If
improving train performance in a particular local spot is
required, the ESS application can be justified.
Using recuperating traction power stations provides
substantial cost benefits in comparison with wayside
ESS and can be justified just on energy savings.
Using diode rectifiers in combination with dc/ac converters (either PWM or line commutated) achieves energy
savings but does not improve train voltage performance.
Upgrading diode rectifiers to the RTCR allows us to
both improve the performance in the forward operation
by increasing the voltage and reducing the voltage regulation, as well as save money by recuperating braking energy. For new systems using RTCR, we can also achieve
additional savings by reducing the number of substations through increased spacing.
Acknowledgments
The author would like to thank the following individuals
for the helpful discussions and information (in alphabetical order): John Calvello (Kawasaki), Paul Forquer
(Powell Electric), Salwa Foulda (Bombardier), Vitaly
Lusherovich (BART), Dick Newark (KTSi), Chuck Ross
(PGH Wong Engineering), Bob Schmitt (NYCTA), Rick
Tetrault (VYCON), Nathan Waissman (Maxwell), and
Rick Wolf (Myers Controlled Power).
Author Information
Vitaly Gelman received his M.S.E.E. degree from Moscow
Power Engineering University in 1976. He designed variable
frequency drives from 1976 to 1982, first in Moscow and
from 1979 at Ramsey Controls in New Jersey. He worked
for ABB in New Jersey from 1982 to 1983 designing and
troubleshooting high power rectifiers. Since 1984, he has
worked for VG Controls, developing various controllers
and protection relays for traction and other heavy
References
[1] V. Gelman, Energy savings with reversible thyristor controlled rectifier, in Proc. 2009 Joint Rail Conf., Pueblo, Colorado, USA, Mar. 45,
2009, pp. 4146.
[2] J. G. Yu, M. P. Schroeder, and D. Teumim, Utilizing wayside energy
storage substations in rail transit systemsSome modeling and
simulation results, in Proc. APTA Rail Conf., 2010.
[3] J. M. Ortega, H. Ibaiondo, and A. Romo, Kinetic energy recovery on
railway systems with feedback to the grid, in Proc. 9th World Congress on Railway Research, May 2226, 2011.
[4] K. Ogura, K. Nishimura, T. Matsumura, C. Tonda, E. Yoshiyama, M.
Andriani, W. Francis, R. Schmitt, A. Visgotis, and N. Gianfrancesco,
Test results of a high capacity wayside energy storage system using
Ni-MH batteries for DC electric railway at New York City Transit, in
Proc. Green Technologies Conf. (IEEE-Green), IEEE, 2011.
[5] L. Romo, et al., Cutting traction power costs with wayside energy
storage systems in rail transit systems, in Proc. 2005 Joint Rail Conf.,
JRC2005, Pueblo, Colorado, Mar. 1618, 2005, pp. 187192.
[6] A. Cooper and P. Moseley, Advanced lead-acid batteriesThe way
forward for low-cost micro and mild hybrid vehicles, World Electric
Veh. J., vol. 3, 2009.
[7] (2013). Kawasaki GIGACELL batteries characteristics. [Online].
Available: http://www.khi.co.jp/english/gigacell/elec/elec3.html;
http://www.khi.co.jp/english/gigacell/elec/elec4.html; http://www.khi.
co.jp/english/gigacell/elec/elec.html
[8] K. Nakano, S. Takeshima, and J. Furakawa, Technological trends
in lead-acid batteries for automotive applications, Furukawa Rev.,
no. 32, pp. 4955, 2007.
[9] WP2B ENERGY RECOVERY, Overview of braking energy recovery
technologies in the public transport field, Deliverable 1Mar. 2011,
Ticket to Kyoto.
[10] (2013). ABB thyristor inverters. [Online]. Available: http://www.abb.com/
industries/ap/db0003db002812/b3f5beb3855ffd82c125779d0052bf8d.
aspx
[11] M. Meinert, K. Rechenberg, G. Hein, and A. Schmieder. (2008). Energy efficient solutions for the complete railway system, Siemens.
[Online]. Available: http://www.uic.org/cdrom/2008/11_wcrr2008/
pdf/PS.2.32.pdf
[12] M. Schroeder, J. G. Yu, and D. Teumim. (2010). Guiding the selection and application of wayside energy storage technologies for rail
transit and electric utilities. [Online]. Available: http://onlinepubs.
trb.org/onlinepubs/tcrp/tcrp_webdoc_51.pdf
[13] (2011). Wayside Energy Storage ProjectProgress Update. [Online].
Available: http://www.septa.org/sustain/blog/2011/07-15.html
[14] A. Okui, S. Hase, H. Shigeeda, T. Konishi, and T. Yoshi, Application
of energy storage system for railway transportation in Japan, in
Proc. 2010 Int. Power Electronics Conf, pp. 31173123.
[15] Lincoln Electrics technical brief. (1999). Evaluating inverter-duty
motor insulation systems using corona inception voltage. [Online].
Available:
http://shared.a2zinventory.com/mfg/LincolnMotors/
Library/tb-100.pdf
[16] (2013). Panasonic PEVE Ni-MH battery. [Online]. Available: http://
www.peve.jp/en/product/np2/index.html
[17] T. Markel. (2006). Plug-in hybrid-electric vehicles: Current status,
long-term prospects and key challenges. [Online]. Available: http://
www.nrel.gov/docs/fy06osti/40239.pdf
[18] S. Venkatesan. (2011). Battery systems for EV/HEV/PHEV applications. [Online]. Available: http://www.et.eng.wayne.edu/images/
PDF/2011_Summer_Workshop/doe%20slides_venkatesan_2011.pdf
[19] Maxwell Technologies. (2012). Maxwell 125V heavy duty transportation module. [Online]. Available: http://www.maxwell.com/products/
ultracapacitors/docs/datasheet_bmod0063_1014696.pdf
[20] A. Schomburg. (2011). Ticket to Kyoto Siemens TCI. [Online]. Available: http://www.tickettokyoto.eu/sites/default/files/download/file/
Siemens_TCI_T2K_Annuel_Event_May2011.pdf
[21] J. Ortega. (2011). Ingeber system for kinetic energy recovery & Metro
Bilbao experience. [Online]. Available: http://www.vialibre-ffe.com/
pdf/Estaciones_reversibles.pdf
[22] Bombardier Transportation. (20112013). EnerGstor wayside energy storage. [Online]. Available: http://www.bombardier.com/files/
en/supporting_docs/BT-ECO4-EnerGstor.pdf