This article has been accepted for publication in a future issue of this journal, but has not been

fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIA.2018.2878143, IEEE
Transactions on Industry Applications

Renewable Energy Sources Integration and

Control in Railway Microgrid
Soukaina Boudoudouh and Mohammed Maaroufi

 the regenerated power is far from the consumed power. The

Abstract— The traffic rail increase implies an increase in the techniques of the energy recovery are listed in [1]. Authors in
electric energy consumption. Hybridizing the railway sub- [2] have proposed an intelligent control to plan the
stations with hybrid energy sources based on renewable energy regenerative energy recovery. In fact, the line voltage increase,
sources and storage units connected to a DC bus may be a occurring due to the regenerative braking energy causes the
solution to contribute to the partial independence of energy traction materials damage, also the regenerative energy
producers in the sector of traffic rail. A smart control is highly generates a bidirectional energy flow on the line. If it is not
recommended in order to avoid disturbing the traffic or the consumed, it can be exported to the main network.
energy quality of railway lines. This paper proposes a reversible, A configuration of the substation that respects the
self-adaptive, autonomous and intelligent distributed generator reversibility without modifying the actual infrastructure, the
connected to the catenary thanks to the DC bus distributed
proximity to the trains' consumption and facilitating the
control by the multi-agent system. The results analysis has shown
regenerative energy consumption that it is not recovered by
that the proposed control architecture can be a solution to face
the issues related to the traffic railway issues. the circulating trains is proposed in this paper. It consists of
hybridizing the substation by Hybrid Power Generation
Index Terms— Railway microgrid, braking and tracking System (HPGS) tied to a DC bus that it is directly connected
energy, renewable energy sources, penalty costs, multi agent to the catenary as depicted in Fig.1. It offers all services
system, Jade, MacsimJX, Matlab Simulink brought by the hybrid systems [3] and does not require any
changes in the existing architecture of the substation. Linking
renewable energy sources (RES) and storage units to the DC
bus, connected to the catenary is the Hybrid Substation (HSS)
I. INTRODUCTION studied in this paper. The proposed HSS characteristics are
he subscribed power exceeding cost weighing the presented in Table I. It consists of a PV and wind turbine
T energy bill of the railway networks actors, its reduction
will undoubtedly alleviate this increase. The placement
generators (PVG and WTG) and battery storage units.
Two RES from different nature are combined, to minimize
of a new DC substation and/or the existing DC their discontinuity and also to avoid their oversizing that can
substation upgrades are the solutions actually implemented. occur if only one source is available. Following their main
From an economic point of view, these solutions do not solve characteristic the RES supply the railway line in
the problems related to the losses generated by the power complementarity. A storage unit such as the battery is tied to
routing from the substation to the trains running on the line. It the RES. It provides the energy necessary to meet the peak
also introduces other investment costs related to the power demand and consumes the excess of regenerative
transmission and distribution network. Moreover, in the case energy in the line thanks to the associated converter. Indeed
of the renewal or extension, if the substation is powered by the interconnection of the RES and the battery to a DC bus in
fossil energy sources, the energy produced will have a direct the catenary can be viewed as distributed generator DG that
impact on the environment and its cost will be high following reduces the subscribed power thanks to the RES generation,
the refraction of this type of energy source. provides the energy necessary during acceleration phases to
Reducing the subscribed power, eliminating the voltage suppress the subscribed power exceeding and eliminates the
drop in the line due to the acceleration and leading to the drop voltage in the line thanks to the battery discharge and the
subscribed power exceeding and avoiding the voltage rise due RES generation simultaneously, and avoids the voltage
to the deceleration by consuming the total of the regenerative increase due to the deceleration process by consuming the
energy not recovered by the other trains in the line, are the energy excess, thanks to the battery charge. The overall power
main issues related to the traffic railway. Indeed, it can occur of the proposed HSS is given in (1). According to (1) the
that the regenerative power exceeds the consumed power, or trains will no more consume the energy needed for circulation
only from the substation but from a judicious use of RES
This paragraph of the first footnote will contain the date on which you generated power PRES, battery power PBattery, and substation
submitted your paper for review. It will also contain support information, power PSubstation thanks to the Distributed Energy Management
including sponsor and financial support acknowledgment. For example, “This (DEM).
work was supported in part by the U.S. Department of Commerce under Grant
BS123456”.
Soukaina Boudoudouh, Mohammadia School of Engineers (EMI), Ptrains  PSubstation  PRES  PBattery (1)
Mohammed V University, Rabat, Morocco ([email protected])
Mohammed Maâroufi, Mohammadia School of Engineers (EMI), Mohammed
V University, Rabat, Morocco ([email protected]).

0093-9994 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIA.2018.2878143, IEEE
Transactions on Industry Applications

Substation
Wind power
controller
WTG

PV power
controller
PVG

Battery

DC BUS

Tracking/Braking

Fig. 1. HSS architecture

TABLE I control is highly recommended. This paper proposes DEM by

Railway microgrid components’ specifications Multi Agent System (MAS) to enable the implementation of
Subscribed power PS=2.5MW railway microgrid with HSS based on HPGS.
MAS is extensively suggested as a suitable approach to
RES maximum power PRES=0.5MW
manage complex distributed system [9] [10] such as the
Battery PBat=3MW railway microgrid and it can give solution to a distributed
Line resistance RL=0.025Ω/Km control scenario. A large number of studies regarding the
Line voltage VL=1500 Volts (DC)
distributed control applications have been developed by MAS
technology [11] [13].
The subscribed power exceeding cost weighing the energy
The integration of the distributed generation such as RES bill of the railway networks actors, its reduction will
in the railway network should meet several technical undoubtedly alleviate this increase. The penalty costs related
constraints [4] [6]. The potential interest and the possibility of to the railway reflects the system's adaptation to energy
integrating RES in the railway substation have been studied in consumption. If the energy production covers consumption,
[7] [8], where it was pointed out that in order to improve the there is no penalty, and then the penalty cost is zero. On the
efficiency and reliability of the system, hybridizing a other hand, if there is a certain lack or excess of energy, the
stationary system based on RES in the traffic railway can only penalty cost Cpenality depends on the square of the missing or
be achieved by thinking of an intelligent approach, because surplus energy EExcess/lack according to (2).
the maximum railway traffic consumption is rarely correlated
with maximum RES production. C penalty  Cu  E 2 Excess / Lack
The HPGS consists of a multi-source system with (2)
decentralized energy sources with different capacities and Cu : Unit Cost of elements
different generation. To derive all the benefits of the braking
energy and to reduce the tracking energy cost and also in order In the proposed railway microgrid, penalty costs are
to avoid disturbing traffic or the energy quality of the railway related to the integration of intermittent energy sources as RES
lines. The energy flow and information coming from each and to the balance between supply and demand and the
source should be controlled simultaneously while integrating subscribed power exceeding.
the HPGS to the substation. That allows the judicious and The penalty costs associated to the RES correspond to the
effective use of each source. Therefore, a need for a DEM is risk of unavailability during peak periods, this cost is covered
raised. by the battery in association with the available RES.
The objective of the proposed DEM is to design a Concerning the penalty costs related to the balance between
reversible, active, intelligent, self-adaptive, and autonomous supply and demand, it correspond to the acceleration and
DG connected to the catenary. Thereby, a DC bus distributed deceleration phases and to the surplus of the unused consumed
control where the HPGD is tied is required. Furthermore, the energy as in the case where the RES are available while the
railway electrical network can be observed as a microgrid, trains running on the line are running with a constant traction
because it is depicted as a chain of distributed loads when the speed or at the trains start or stop, where the trains speed
trains are in acceleration mode and also as distributed running on the line is zero. The penalty cost associated with
generators when the trains are in deceleration mode. Having this excess energy is totally covered by the battery for reuse in
regard to this, a very sophisticated distributed DC bus voltage

0093-9994 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIA.2018.2878143, IEEE
Transactions on Industry Applications

the event of peak or acceleration phases. Finally, the penalty by assigning the current reference Irefdc calculated during the
cost related to the subscribed power exceeding is covered voltage control.
simultaneously by the RES and the battery. The 'battery agent' may also refuse the 'line manager agent'
The proposed railway microgrid control by MAS is quoted proposal if its SOC is not correlated. The DC bus current
in section II. Section III shows the reliability of the proposed control is then supported by other agents. The case where the
microgrid, by detailing the simulation results. Section IV SOC is not correlated with the proposal of the 'line manager
concludes the paper. agent' is not taken into account.
The battery is current-controlled in order to regulate the
II. RAILWAY CONTROL BY MAS bus voltage and balance the energy flow in the DC bus. The
The control of the railway microgrid follows the same current reference Irefdc is the current to be assigned to the
principle of a power flow control in a DC bus [14]. The converters and becomes Ibattdc in order to regulate the bus
current flowing in the line reveals different possible scenarios; voltage as it was previously pointed out. The calculation of the
a train or several trains brake, or a train brakes while others DC bus voltage is given by (3) which Cdc is the DC bus
accelerate. Therefore, the average current represents the capacitor. Irefdc is calculated via the bus voltage vdc during
outcome of the various possible scenarios that may occur. It is voltage control. The difference between the line current
viewed by the DC bus as reversible load. The positive (substation ISubstation and trains consumption Itrains) and the
resultant current expresses the acceleration phase and the RES current IRES is seen by the DC bus as a bidirectional load;
negative resultant current represents the deceleration phase. consumes energy in the event of acceleration and provides
The intelligence of the proposed control by MAS allows energy in case of deceleration. The difference between these
the battery to be charged from the RES generated power while currents is the current to be controlled by the battery via Irefdc.
trains are stopped and during deceleration periods, to reuse the
stored energy during acceleration periods. It also allows the vdc 
1
Cdc 

I Battery dc  I RES  I substation  Itrains dt
(3)

RES to reduce the subscribed power while the trains speed is
constant.
Using the MacsimJX tool [15] [16], there are three agents
that receive three Simulink signals. All the necessary Moreover, when the trains speed on the line is equal to
calculations are carried out by Simulink while the decision zero (at the start or the stop), the current to be controlled by
making is performed by the agents in the Jade environment the battery is the current generated by the RES, for this reason,
[17]. the battery consumes the RES current when the trains are
The resulting current signal which informs about the stopped on the line or during the start.
various scenarios that may occur on the line is the input signal When the trains’ speed running on the line is constant, the
of the 'line manager agent'. The processing of this data by the objective will be to reduce the subscribed power thanks to the
agent indicates whether a request for energy supply is RES generation. Consequently, the RES will supply the line
required, due to an acceleration in order to suppress the directly in order to reduce the subscribed power. On the other
voltage drop on the line, or a consumption demand in order to hand, if the difference between the resulting current on the line
remove the voltage increase due to deceleration is necessary. (from the different circulating, accelerating and decelerating
The 'line manager agent' then sends this information to trains) and the RES generated current is always considered as
agents interested by this information and who can provide the the current to be controlled by the battery, there will be an
supply or consumption of energy service to meet the excess of energy on the line when trains’ speed is constant
acceleration and deceleration phases requirements. Actually, because the RES production will exceeds the trains
these agents are 'battery agent' and 'RES manager agent'. consumption moving at fixed speed. Then the battery will
On the other side, the 'battery agent' receives from consume this excess of energy of renewable origin instead of
Simulink the signal indicating its state of charge SOC. If the integrating it in the line in order to reduce the subscribed
proposal of the 'line manager agent' is correlated with the power as it is desired.
SOC, the 'battery agent' accepts the proposal and begins to In order to achieve this objective, the proposed smart
supply the energy needed during the acceleration which is not control by MAS assumed in this case that on the DC bus the
recovered by the RES, that also participate in the supply of power flows are balanced, that means, the RES generated
energy during acceleration. The battery consumes the total power will not be integrated into the DC bus. ie there will be a
excess of energy on the line due to deceleration and also due zero reference current, Irefdc = 0 and the converter associated to
to the RES generation which is not consumed at deceleration the battery receives a set-point equal to zero IBref = 0. This is
time. the role of the 'RES manager agent'.
The 'battery agent' sends its decision of acceptance to the
converter associated to the battery in Simulink environment,

0093-9994 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIA.2018.2878143, IEEE
Transactions on Industry Applications

iTrains
iTS
iT  I RES Vdc
S

SpeedTrains
Line manager Agent
RES manager Agent
Ye No
s IR  0
Processing Data
Send Send
controller proposal to proposal to
Decision Making agents able agents able
to provide to provide
Wind power supply consume
controller service service
IRES
Wind Turbines Agent Send request

PV power
controller

PVs
Battery Agent
Voltage
Controller Processing Data
Vref
PI
controller
Decision Making
Irefdc

Battery converter

Irefdc Agent
IB
IBref
0 PI dc

ILB

Fig. 2. Railway microgrid control principle

The 'RES manager agent' receives from Simulink the line power exceeding. On the other hand if a deceleration occurs
circulation cycle signal, which informs about the trains speed the battery consumes the total of the braking energy that is not
on the line, and receives in parallel from Jade the 'line recovered by other trains and the energy produced by the RES
manager agent' proposal. If there is a demand to supply energy considered in excess in the event of deceleration.
supply, it means that the resulting current in the line is If the traffic speed is constant, the 'RES manager agent'
positive. On the other hand, the processing of the information will integrate a zero current on the DC bus, which means that
coming from Simulink which provides information about the the bidirectional load current in the DC bus is considered to be
trains speed through the line will allow the 'RES manager equal to zero, in order to have IrefDC = IBref = 0. This lets to
agent' to take the decision to accept or to reject the 'RES consider a DC bus where the power flows are balanced. On
manager agent' proposal. the other hand, the generation of the RES in the case where the
If the circulation speed is non-constant and a supply circulation speed is constant supplies the line directly in order
request is suggested, the 'RES manager agent' will integrate to reduce the subscribed power. The demonstration is given
into the DC bus the resulting current which is the resulting later by analyzing the simulation results. The proposed railway
traction and braking current on the line and the RES current as microgrid control principle is illustrated in Fig.2. It is shown
a bi-directional load as explained above. Therefore, the current that each agent depending on received signals and messages; it
to be monitored by the battery is the difference current, i.e., if controls autonomously the corresponding elements. It gathers
acceleration is produced the battery supplies the energy that is the needed data to determine the set-points following the
not covered by the RES in order to eliminate the subscribed

0093-9994 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIA.2018.2878143, IEEE
Transactions on Industry Applications

PS=2,5 MW

2,5MW 2,15MW 1.9MW 1,7MW

A B C D E
Fig. 3. The available power in the railway line (P S=2.5 MW)

PS=2.5 MW

2MW 1.274MW 1.427MW 1.24MW

A 0.5MW B 0.487MW C 0.473MW D 0.46MW E

DG
Fig. 4. The available power in the railway line while integrating a DG
(PRES=0.5 MW)
PS=2.5 MW 0.51MW PS=2,5 MW 3.51MW

3,51MW 3,51MW

3MW
A B A B
3MW 3MW
DG

Fig. 5.a . The peak load demand near to Fig. 5.b . The peak load demand near to
substation consumption with a DG substation consumption without a DG

PS=2.5 MW 0.38MW PS=2.5 MW 3.38MW

3,38MW 3,38MW

3MW D
A DG 2.4MW A 2.4MW
D

Fig. 6.a . The crossing point peak load Fig. 6.b . The crossing point peak load
demand consumption with a DG demand consumption without a DG

decision-making. Therefore, the agents are autonomous and DG here depicts the RES which their maximum power is
self-adaptive. about 0.5 MW.
The RES do not always generate the maximum power.
III. RESULTS AND DISCUSSION However the objective here is not to activate the battery to let
the RES reach their maximum power. The battery is turned on
Two configurations are studied. The first one illustrates the to remove the rise voltage and aiming solving the peak load
railway line when a DG is integrated, and aims to show the demand causing the drop voltage occurring for example near
reaction of the hybrid system tied to the DC bus which is to substation as shown in Fig.5.a and Fig.5.b.
viewed as a DG, in the different cases. The second The maximum power that can be delivered by the hybrid
configuration validates the strategy of energy management by system tied to the DC bus is about 3 MW. It is shown in
MAS proposed in the previous section, by taking full Fig.5.a that the peak load demand is supplied by the hybrid
advantages of the capacity, characteristics and limitations of system and therefore the subscribed power is not exceeded.
each element in the HS following a driving cycle scenario. Contrarily to Fig.5.b the peak load demand is provided by the
The objective is to improve the energy quality of the whole substation and exceeds the subscribed power by 1.512 MW in
system by the energy flow distributed and intelligent control. order to meet the peak load demand occurring near to
Fig.3 shows the available power in the line. The subscribed substations.
power is equal to 2.5 MW. The distance between each point in The crossing point between two trains is considered as the
the line is about 5 Km. Due to the losses in the line, the second peak load. Fig.6.a and Fig.6.b illustrate a crossing
available power decreases while going away from the point occurring at point D that consumes 2.4 MW. It is shown
substation. When a DG is integrated as illustrated in Fig.4, the in Fig.6.a that the hybrid system which is viewed as a DG
subscribed power is reduced, also the power coming from the provides its maximum power to meet the peak load demand
substation. The subscribed power can be reduced by 0.5MW. without exceeding the subscribed power. It is shown in Fig.6.b
Therefore, voltage drop across the catenary system due to the that the peak load demand is covered by the substation letting
long distance between two substations is also reduced. The

0093-9994 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIA.2018.2878143, IEEE
Transactions on Industry Applications

the subscribed power exceeds the conventional value by 0.88

MW.
To cut the reversibility with the substation, to take full
advantages of the energy recovery and to control the energy
flow coming from each element following their own capacity,
an intelligent strategy of energy management is introduced in
the previous section that aims to derive all the benefits of the
braking energy and to reduce the tracking energy cost thanks
to the proposed HS architecture. By the following the
proposed control strategy by MAS is validated following a
driven cycle simulated between A and B.
For the simulation, the circulation cycle shown in Fig.7
was considered. The circulation circuit considered lasts 120 s
and simulates three consecutive accelerations, and cruise
phases at different speed values. The braking phase starts at 80
s from 80 km / h, which represents the maximum speed.
Fig.8 shows the resulting current of the different tracking Fig. 7. The considered driving cycle
and braking trains on the line following the circulation cycle
shown in Fig.7. Fig.9 shows the current generated by the RES
during the simulation time. Fig.10 and Fig.11 show the battery
reaction and the different currents on the line during the
simulation respectively.
At the start and until 12 seconds, the speed on the line was
equal to 0. The aim is to take full advantages of the RES
generation during the trains stop. In Fig.10 and Fig.11 it is
shown that the battery takes control of the DC bus current and
it is charged with the total current generated by the RES
during this period. Fig.12 shows the SOC evolution during the
simulation. It is indicated that it is equal to 70%, from which
the 'battery agent' agrees to consume the energy surplus.
Then three consecutive accelerations took place, the RES
and the battery simultaneously supply and share the requested
current during the three accelerations such that the battery
covers the difference between the current resulting from
various traction and braking operations on the line and the
current generated by the RES. During the first acceleration the Fig. 8. Resultant current from tracking and braking process
battery covered almost 90% of the current required by the
acceleration while the RES covered 10%. Always in Fig.11, it
is indicated that during the second acceleration the battery
covered 88% of the acceleration energy and the RES provided
the 12%. Finally the third acceleration was covered at 91% by
the battery and 9% by the RES. During the three consecutive
accelerations the battery accepts to supply energy depending
on its SOC as shown in Fig.12.
Between the three accelerations, the speed on the line is
constant. Depending on the objective of the proposed control
by MAS, it can be seen that the battery does not consume the
excess of the RES and the battery current is zero during the
periods when the speed is constant, because the energy
management strategy by MAS desires that this excess of
energy coming from the RES participates in the reduction of
the subscribed power. So the 'RES manager agent' does not
supply the DC bus by the RES but directly feeds the line.
Fig.13 shows the subscribed power gain. It is reduced to Fig. 9. Current generated by RES
maximum to 2 MW and at a minimum to 2.24 MW.
During the deceleration phase, the 'RES manager agent'
decides to supply the DC bus by the RES in order to take

0093-9994 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIA.2018.2878143, IEEE
Transactions on Industry Applications

Fig. 10. Battery DC-DC converter output current Fig. 13. Subscribed power gain

Fig. 11. The line current Fig. 14. Subscribed power exceeding removal

Fig. 15. Penalty cost removal

Fig. 12. Battery SOC evolution

0093-9994 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIA.2018.2878143, IEEE
Transactions on Industry Applications

penalty costs related to the subscribed power exceeding and

the RES intermittence and also to the acceleration and
deceleration were suppressed, thanks to the simultaneous
control of the battery with the generation of the RES. The
results also showed the stability and continuity of the system
thanks to the effectiveness of the proposed control.

References
[1] R.R.Pecharroman and al, “Riding the Rails to DC Power Efficiency:
Energy efficiency in dc-electrified metropolitan railways”, IEEE
Electrification Magazine, vol. 2, No.3, pp. 32 – 38 , 2014.
[2] Boudoudouh Soukaina, and Mohammed Maâroufi. "Smart control in a
DC railway by Multi Agent System (MAS)." In Electrical Systems for
Aircraft, Railway, Ship Propulsion and Road Vehicles & International
Transportation Electrification Conference (ESARS-ITEC), International
Conference on, pp. 1-6. IEEE, 2016.
[3] Hajizadeh, A., Golkar, M. A., "Intelligent power management strategy
Fig. 16. DC bus voltage evolution of hybrid distributed generation system", Elsevier, electrical Power and
Energy systems, pp. 783-795, 2007
[4] H.Ibrahim and al “Integration of Wind Energy into Electricity Systems:
advantage of both, the regenerated energy due to the braking Technical Challenges and Actual Solutions”, Energy proceedia, vol. 6,
process and also consume the total energy from the RES, pp. 815_824, 2011.
which enable the battery to charge, in order to be discharged [5] B. Robyns and al, “Electricity production from renewables energies”,
during critical periods. ISTE Wiley 2012.
During the last phase, the trains are stopped. Consequently, [6] M. Alvarez-Herault, “Architectures of the distribution futir networks in
their speed is equal to zero. The battery consumes the current the presence of decentralized production”, Phd thesis, University of
coming from the RES which supply during this phase the DC Grenoble, 2009.
bus following the 'RES manager agent' decision in order to [7] R.Faranda.;S. Leva “ Energetic Sustainable.developement of railway
stations”;. IEEE Power engeneering society general meeting pp 1-6,
achieve the DEM objectives in the proposed railway Tampa, FL, 2007.
microgrid. [8] H. Hayashiya and al, “Necessity and possibility of smart grid technology
Fig.14 shows the remove of the subscribed power application on railway power supply system”, proceeding of 14th
exceeding, by the battery and the RES during the acceleration European conference onpower electronics and applications (EPE 2011),
phases. Fig.15 shows the suppression of the penalty cost pp 1-10, Bermingham, 2011.
related to the acceleration and the subscribed power overrun [9] Khamphanchai, W.;Pipattanasomporn, M., and Rahman, S., “A Multi-
by the battery and the RES. Also the removal of penalty cost Agent System for Restoration of an Electric Power Distribution Network
with Local Generation,” IEEE Power and Energy Society General
related to the deceleration and the energy excess coming from Meeting, 2012.
the RES exhaustively via the battery. Fig.16 illustrates the
[10] W. Khamphanchai, M. Kuzlu and M. Pipattanasomporn, “A Smart
evolution of the DC bus voltage during the simulation. It is Distribution Transformer Management with Multi Agent Technologies,”
constant equal to the reference voltage throughout the IEEE PES Innovative Smart Grid Technologies, 2013
simulation time, which proves the success of the control by [11] D. E. Olivares, C. A. Canizares, and M. A. Kazerani, “Acentralized
MAS of the proposed railway microgrid. optimal energy management system for microgrids,” in Proc. IEEE
Power Energy Soc. Gen. Meet., Jul. 2011, pp. 1–6.
IV. CONCLUSION [12] A. H. Etemadi, E. J. Davison, and R. Iravani, “Adecentralized robust
control strategy for multi-DER microgrids—Part I: Fundamental
This paper deals the DEM by MAS in the railway concepts,” IEEE Trans. Power Del., vol. 27, no. 4, pp. 1843–1853, Oct.
microgrid with HSS based on HPGS to meet the limitations of 2012.
rail transportation systems in terms of energy saving. The [13] AL. Kulasekera, RARC. Gopura, KTMU. Hemapala, et al (2011), “A
HPGS consists of a multi-source system with decentralized Review on Multi-Agent Systems in Microgrid Applications,” In:
Proceedingsof IEEE PES Conference on Innovative Smart Grid
energy sources with different capacities and a different Technologies India (ISGT 2011), Kollam, Kerala, India, December 1-3,
generation, therefore, judicious use and integration of each 2011.
element were respected. Reducing the subscribed power, [14] S.Boudoudouh; M.Ouassaid; M. Maâroufi, ‘Multi Agent system in a
eliminating the voltage drop in the line due to the acceleration distributed energy management of a multi sources system with a hybrid
and leading to the subscribed power exceeding and avoiding storage’ 3th renewable and sustainable energy conference (IRSEC,
2015), IEEE conference
the voltage rise due to the deceleration by consuming the total
of the regenerative energy not recovered by the other trains in [15] S. Boudoudouh et M. Maâroufi, ‘Real Time Battery State Of Charge
Estimation In Smart Grid Application by Multi Agent System’ Journal
the line, remain the main issues that should be taking into of Hydrogen Energy, 2017.
account while hybridizing the substation without modifying [16] S.Boudoudouh, M. Maâroufi (2017), Real Time Distributed Systems
the existing architecture. Thereby, this paper meets the Modeling and Control: Application to Photovoltaic Fuel Cell
mentioned limitations and constraints by designing reversible, Electrolyser System, journal of Engineering Science and. Technology
active, intelligent, self-adaptive, and autonomous DG Review (jestr), Vol 10, issue 1
connected to the catenary thanks to the distributed DC bus [17] Fabio Bell feminine, Telecom Italy, “Giovanni Caire, telecom
voltage control by MAS. It was shown the ability of the Italy”Developing Multi Agent Systems with JADE “John Wiley& sons
2007”.
proposed control to reduce the subscribed power and to omit
the subscribed power overrun by the RES generation and the
storage system which is represented by the battery. The

0093-9994 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.