1544
4, APRIL 2013
Energy Management Strategy of the DC Distribution
System in Buildings Using the EV Service Model
Gilsung Byeon, Member, IEEE, Taeyoung Yoon, Student Member, IEEE, Seaseung Oh, Member, IEEE,
and Gilsoo Jang, Senior Member, IEEE
Abstract—In this paper, a new energy management strategy
(EMS) for a dc distribution system in buildings is being proposed.
The dc distribution system is considered as a prospective system
according to the increase of dc loads and dc output type distribu-
tion energy resources (DERs) such as photovoltaic (PV) systems
and fuel cells. Since the dc distribution system has many advan-
tages such as feasible connection of DERs and electric vehicles
(EVs), reduction of conversion losses between dc output sources
and loads, no reactive power issues, it is very suitable for industrial
and commercial buildings interfaced with DERs and EVs. The es-
tablishment of an appropriate EMS based on the economic point
of view can reduce energy costs of buildings and provide benefits
to participants in energy management. Applicable elements for the
dc distribution system are identified and the real-time decision-
making-based algorithm for minimizing operating costs is pro-
posed in this paper. The EV service model for the EMS to offer
incentive to EV owners who participate in battery discharging is
described. To verify the performances of the proposed algorithm,
computer simulation and economic analysis are being performed
where the results show that the proposed EMS reduces energy
costs, motivates EV owners, and can be applied to the dc distribu-
tion buildings.
Index Terms—Dc distribution, dc system, electrical vehicle (EV),
energy management, EV charging/discharging.
I. INTRODUCTION
ECENTLY, environmental concerns have become a hot
R issue in the field of power systems. The countries ratified
to the Kyoto protocol have agreed to reduce the emission of
greenhouse gases. It is achievable through the reduction of var-
ious losses associated with generation, transmission, distribu-
tion, and consumption of electrical power, since this reduces the
amount of energy that is needed to be produced thereby reduc-
ing pollution and greenhouse gas emissions. In power systems,
renewable resources and electrical vehicles (EVs) are consid-
ered as good alternatives. Distributed energy resources (DERs)
and EVs produce and utilize dc voltage. Therefore, these units
have to be interfaced with dc systems which are more efficient
than ac systems. Furthermore, dc distribution systems are being
Manuscript received March 30, 2012; revised July 3, 2012; accepted July 21,
2012. Date of current version October 26, 2012. This work was supported by
the National Research Foundation of Korea (NRF) Grant funded by the Korea
Government (MEST) (No. 2012-0006275). Recommended for publication by
Associate Editor J. Choi.
G. Byeon, T. Yoon, and G. Jang are with the School of Electrical Engi-
neering, Korea University, Seoul 136-701, Korea (e-mail: bgsean@korea.ac.kr;
xodudgenius@korea.ac.kr; gjang@korea.ac.kr).
S. Oh is with the Department of Energy Efficiency, Korea Institute of Energy
Research, Daejeon 305-343, Korea (e-mail: shung@kier.re.kr).
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TPEL.2012.2210911
0885-8993/$31.00 © 2012 IEEE
IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 28, NO.
spotlighted as a good energy solution as the number of buildings
that have high percentage of dc loads continues to increase.
The dc distribution system is not a new concept and presently
several studies, some of the main topics being enlisted below,
have been performed for setting up the detailed aspects of the dc
distribution system. In [1], the authors performed several pieces
of research related to dc power system analysis and modeling. A
dc distribution system for remote customers by using a fly-back
converter was proposed in [2]. The VSI control scheme that
enhances the dc distribution system reliability and extendibility
and the dc bus voltage control method was proposed in [3].
In [4], a grounding scheme for improving neutral voltage and
fault current issues was proposed.
From the viewpoint of the energy storage device applications
in dc distribution systems and microgrids, the energy storage
control and strategies are described to meet transient load de-
mands of “more electric aircraft” (MEA) in [5]. In the dc-bus
system, energy storage devices are implemented to improve the
reliability and energy efficiency of distributed generation sys-
tems such as photovoltaic (PV) system [6]. In [7], a composite
energy storage system that contains both ultracapacitor and bat-
tery and control strategies in a microgrid is proposed. DC voltage
control and power balancing schemes using a grid-connected
converter and an energy storage device for a dc microgrid are
described in [8].
From the viewpoint of energy management, energy manage-
ment strategy (EMS) schemes for a droop or mode-controlled
microgrids have been proposed in [9]–[14]. Most of the afore-
mentioned control schemes are developed mainly to stabilize the
dynamic behavior of the system where voltage and frequency-
droop characteristic methods are the representative algorithms.
In order to plan the day-ahead energy management schedul-
ing for conventional generators and PV systems, an operational
power planning based on power prediction and load forecast-
ing has been proposed in [15]. The comparison of the source-
following and grid-following power control strategies for DERs
and microgrid is introduced in [16]. To calculate the optimal
size and allocation of ESS, “net present value” (NPV)-based
economic analysis has been performed in [17]. In [18], using an
improved SMES controller, the stabilization and control scheme
of the power flow of microgrids incorporating wind generation
is proposed. The EMS system that contains power forecasting,
ESS management, and optimization module has been proposed
in [19]. In [20], a polynomial control method to control dc bus
voltage and manage embedded energy using a battery and a su-
percapacitor is described. As can be seen in the aforementioned
papers, various aspects have been studied based on EMS in dc
distribution systems and microgrids. In many papers, in-depth
BYEON et al.: ENERGY MANAGEMENT STRATEGY OF THE DC DISTRIBUTION SYSTEM IN BUILDINGS USING THE EV SERVICE MODEL
Fig. 1. Concept of a bipolar-type dc distribution system with the EV station.
studies on the transient control and stabilization system with
energy storage devices have been conducted and various EMS
schemes have been developed.
In this paper, a new EMS to minimize operating costs for a
dc distribution system in buildings is proposed where the pro-
cess that determines the optimal output of the nonutility source
and incoming power from the external grid are being described.
The proposed EMS method in this paper is not based on op-
timization but is based on the EMS method utilizing real-time
decision making. The method does not involve optimization
based on the long-term data and has the advantage of deriving
the best judgment for the corresponding moment using the var-
ious data achieved in real time. The technique when compared
to the general optimization techniques can operate in real time
owing to lower computation complexity and can be directly ap-
plied to a new system. The service model for EV owners who
participate in the battery discharging is also proposed. To verify
the proposed algorithm, simulation tests and economic analysis
studies are performed. Also, other parameters such as output
power, load capacity, state of charge (SOC) of EVs, and energy
cost index are shown and discussed. This paper is organized
as follows. Section II describes the dc distribution system in
buildings and its components in detail. Section III represents
the EMS to minimize energy costs and provide benefits to EV
owners. The results of the simulation tests and economic anal-
ysis studies are described and discussed in Section IV. Some
comments about future works and conclusions are presented in
Section V.
II. DC DISTRIBUTION SYSTEM IN BUILDINGS
A. Bipolar-Type DC Distribution System
The basic circuit of the bipolar-type dc distribution sys-
tem in building is represented in Fig. 1. Generally, the dc
1545
distribution system can be configured using the unipolar type or
the bipolar type [21], [22]. In the unipolar type, all components
are connected directly between ± conductors. This case is easy
to implement and decreases the number of power electronics
components. However, in the unipolar system, the transmission
capacity of the system is smaller compared to the bipolar type
with the same voltage level (i.e., 200 and ±200 V) and does not
have a wide range of choices in dc voltage level. In the bipo-
lar type, the system requires more components and the system
results in an unbalanced situation when the loads are not identi-
cally connected. However, the bipolar type is regarded more suit-
able for the dc distribution system in buildings when the advan-
tages such as multiple selective voltage level and enough trans-
mission capacity for nonutility generators, DERs, and EVs are
being considered [9]. Therefore, the bipolar-type dc distribution
system is selected as a basis model of the future-oriented build-
ing’s distribution system in this paper. The three-level bridge
block implements a three-level power converter that consists of
three arms of power switching devices. Each arm consists of
four switching devices along with their antiparallel diodes and
two neutral clamping diodes as shown in Fig. 2. This converter
is controlled using the dc voltage-regulated PWM as shown in
Fig. 3. It is assumed that this dc building has more dc loads
than ac loads to fit the purpose of the dc distribution system
being dealt in this paper. The unidirectional dc/dc converters
are used as the dc load-side converters which can be supply 200
or 400 V dc voltage using two switching devices as shown in
Fig. 2. Following the global trend related to the introduction of
DERs and cogeneration systems, uncontrollable sources such as
PV systems and controllable sources such as gas, diesel turbine
generator, or fuel cell can be connected to the dc distribution sys-
tem using power electronics devices. And the EV that received
attention in a power system recently is suitable to connect to the
dc distribution system too. A controlled dc/dc boost converter
1546
Fig. 2. Simulation circuit of a bipolar-type dc distribution system.
Fig. 3. DC regulator control block.
is implemented to adjust the varying PV system voltage to a
400 V dc voltage. A three-phase PWM current-controlled ac/dc
converter is used to interconnect the diesel ac generator with the
dc distribution system. Several bidirectional dc/dc converters
are implemented to model the EV charging/discharging station.
By changing the operating mode, the batteries of EVs can be
charged or discharged. DC circuit breakers and relays are set up
essentially in interconnection points to protect the system from
disturbances.
B. EV Battery Model
In this paper, the EV battery model is implemented based
on a modified Shepherd model [23]. This model assumes that
the charging and discharging characteristics are identical. The
no-load voltage is calculated with a nonlinear equation and the
actual SOC is derived from the actual battery current. These
IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 28, NO. 4, APRIL 2013
parameters are being calculated by (1). This equation represents
the discharging characteristics of the battery model available in
the MATLAB/Simulink library [24]

Q
E = E0 − K ·  + A · exp(−B idt) (1)
Q − idt
where E is the no-load voltage (V), E0 is the battery nominal
voltage (V), K is the polarization voltage (V), Q is the battery
capacity (Ah), A is the exponential voltage (V), and B is the
exponential capacity (Ah−1 ).
This model is widely used in many EV dynamic studies be-
cause of its simple implementation. In this paper, EV batteries
are implemented based on modified parameters obtained from
commercial EVs.
C. Controllable and Uncontrollable Sources
In recent years, the number of buildings that have nonutility
cogeneration system to produce electricity and heat has been in-
creasing. These buildings can benefit economically because they
simultaneously generate their own consumption power and use
the waste heat for heating and cooling purposes. Controllable
sources that can be used as cogeneration systems in buildings
are diesel, gas turbine, fuel cells, and so on. When a control-
lable source which generates the ac voltage is introduced, it can
be connected to the dc distribution system using an ac/dc con-
verter. A comparison with the generation cost of the nonutility
source and the price of public electricity is an important fac-
tor in the cost-based EMS. Uncontrollable sources such as the
PV systems and wind generators support power management
in buildings by peak shaving. If these systems are installed in
buildings, generation costs are not incurred except for the capi-
tal and maintenance cost. Depending on the system conditions,
an energy storage device might be required to mitigate their
BYEON et al.: ENERGY MANAGEMENT STRATEGY OF THE DC DISTRIBUTION SYSTEM IN BUILDINGS USING THE EV SERVICE MODEL 1547

unpredictable output and improve energy efficiency. In this pa-

per, a small-scale diesel generator and a PV system is used
as a controllable and uncontrollable source, respectively. These
components are modeled using the MATLAB–Simulink toolbox
SimPowerSystem [24]. The output power of a diesel generator
is changed by the reference signal from the energy operator of
the building. The PV system used in this paper is composed of
cascaded PV dc-building modules (PV-DCBM) [25], [26].

III. EMS
It is reasonable to install the dc distribution system in a build-
ing that has many dc loads in the viewpoint of energy system
operator. How to configure the distribution system is closely
related to energy efficiency in a building. If a well-designed
distribution system in the building is constructed, it helps cut
down the overall energy consumption expenses. In addition to
the design of the distribution system, the establishment of the
EMS is also important. To improve profits and reduce costs, the
EMS should check all energy consumption indices, calculate
costs, and perform the optimal behavior.
If the data related to the building load, the EV charging,
and climatological statistics are accumulated in a long time, a
long-term EMS can be developed. However, since a long-term
strategy is mainly based on the prediction algorithm, it could be
sometimes inaccurate and increases the computational burden.
The major drawbacks of a long-term strategy are as follows:
1) its application is almost impossible in buildings that do
not have accumulated data;
2) most energy management strategies based on prediction
or other complicated techniques are difficult to implement
and are hard to understand and modify for general energy
operators;
3) It could have a slow speed due to computational burden.
Therefore, the EMS based on the real-time cost data can be
more suitable for new buildings since it is easier to understand
and operate. This paper proposes the EMS that utilizes real-time
power and cost data, and helps operator behave rationally and
automatically for dc distribution buildings.
Fig. 4. Flowchart of the proposed EMS.
In Fig. 4, the flowchart of the proposed EMS is represented.
First, the cost and power data related to nonutility generators
(controllable sources), DERs (uncontrollable sources), EVs, the are inserted in this algorithm. Therefore, the proposed EMS
utility grid, and the price of electricity from the utility grid are changes the mode of EVs to minimize the incoming power from
collected in real time. Then, the optimal generation output is the utility grid. To implement this EMS, some conditions should
determined using the aforementioned data. If the amount of be satisfied. A detail description about the optimal generation
loads and EV charging power is larger than the sum of output output calculation and the EV service model is represented in
power of nonutility generators and DERs, the system should next subsections.
use the incoming power from the utility grid. In this system,
the dc distribution system is always connected to the utility A. Optimal Nonutility Generation Output Calculation
grid. However, if the price of electricity from the utility grid is
high, the EMS should reduce the incoming power which is the The total energy cost required for loads and charging EV in
moment when the EV service model is being introduced. This the building is shown as follows:
model is composed of selective standby and discharging of EVs. ECtotal = ECtotal − ECheat
gen +ECUG +ECtotal EV (2)
The goal of selective standby is to decrease the total charging
power for EVs (PEV C ). If the system needs the incoming power where ECtotal is the total energy cost of electrical components.
despite of the selective standby, the EMS starts the discharging ECgen , ECUG , ECEV total , and ECheat are energy costs of
of EVs. To avoid frequent changes of mode, dead band blocks the nonutility generator, incoming power from the utility grid,
1548 IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 28, NO. 4, APRIL 2013

TABLE I
EV OPERATIONS ACCORDING TO THE ELECTRICITY PRICE

 
Fig. 5. Flowchart of the optimal generation output calculation.
ECi,gen = Ci,0 + Ci,1 · Pi,gen + Ci,2 · Pi,gen
2
(3)
Pi,gen heat m in ≤ Pi,gen ≤ Pi,gen m ax (heat load following)
0 ≤ Pi,gen ≤ Pi,gen m ax (electrical load following)

m
ECtotal gen = ECi,gen (4)
i=1

where Pi,gen is the ith nonutility generator output power (kWh),

Pi,gen m ax is the maximum output power of ith nonutility gen-
erator output power (kWh), and Pi,gen heat m in is the power
produced to meet the minimum heating requirements of the
building (kWh)
The energy cost of incoming power from the utility grid is
obtained by
 
 m n

Fig. 6. Graphical description of several modes for the EV service model. ECUG = CUG · Ptotal load − Pi,gen −PPV − Pi,EV D .
i=1 i=1
(5)
charging/discharging power of EVs, and cost savings by uti- The amount of total load is
lizing heat produced by the cogeneration system, respectively.  
n

In (2), the operator does not need to get the real-time value
of ECheat . This cost is obtained when the total energy cost of
the building is calculated on a daily basis. This value is equal
to the amount of energy needed to be purchased externally in
the absence of cogeneration in the corresponding system. If the
cogeneration system generation cost is lower than the price of
electricity from the utility grid and the electrical load is larger
than the heat load, it can be assumed that the minimum thermal
energy required in this building should be produced. However,
it is better not to following the heat load pattern in all other
cases. The cost function of the ith nonutility generator is shown
in [27]
Ptotal load = PDC Load + Pi,EV C (6)
i=1
where CUG is the price of electricity from the utility grid
($/kWh), PPV is the BIPV system output power (kWh),
PDC load is the dc load demand (kWh), Pi,EV C is the charging
power for the ith EV (kWh), and Pi,EV D is the discharging
power for the ith EV (kWh).
Generally, the total energy cost is calculated by the sum of
nonutility generators’ generation cost and the price of incoming
power from the utility grid. Since cogeneration systems and
the EV service model are included in the proposed EMS, the
energy cost for EVs’ charging (ECtotal EV ) is added and the
BYEON et al.: ENERGY MANAGEMENT STRATEGY OF THE DC DISTRIBUTION SYSTEM IN BUILDINGS USING THE EV SERVICE MODEL 1549

TABLE II
PARAMETERS OF THE TEST SYSTEM

cost savings by utilizing the thermal energy (ECheat ) can reduce a contract anytime and set the required SOC parameters
the total cost. If the nonutility generator is not a cogeneration (SOCm in , SOCdischarging ).
system but a typical generator for only electricity production, 2) If there is no disconnection by the EV owner, the build-
terms (ECheat , Pi,gen heat m in ) related to the thermal energy are ing energy operator must ensure charging more than the
removed. If it is assumed that all EV under the certain SOC must minimum SOC by contract (SOCm in in Fig. 6).
be charged, the energy cost for EVs’ charging (ECtotal EV ) is 3) If the EV’s SOC reaches the maximum status by contract
an independent variable from the output power of any nonutility (SOCm ax in Fig. 6), the building energy operator can stop
generator (Pi,gen ). To obtain the minimum total energy cost, (2), the EV charging to avoid dummy charging.
which is a function of Pi,gen , is differentiated thereby resulting 4) Through energy management actions such as peak load
in (7), which is a simplified equation derived from (2)–(5), that shaving, the charging of EVs can be temporarily stopped
provides the nonutility generators’ optimal generation value for EVs that exceed the SOCm in value (selective standby
Fig. 6)
Popt gen = (CUG − C1 )/2C2 . (7) 5) For cases when the load level is higher, the discharg-
ing of EVs can be performed for EVs that exceed the
The optimal generation output is determined by the coeffi- SOCdischarging value (discharging of EVs in Fig. 6).
cients of the generation cost function and the price of electricity According to the proposed EV service model, the charging
from the utility grid. If the amount of load that is given by the cost of each EV and the total energy cost of all EVs are shown
difference of PTotal load and PPV is larger than the optimal in
generation output, the output power of a nonutility generator
is constant. However, if the amount of load is smaller than
the calculated optimal generation output, the output power of 
k
a nonutility generator is the same as the amount of load. The ECi,EV = CEV C · Pi,j,EV C − CEV D
flowchart of the optimal generation output calculation algorithm j =1
is represented in Fig. 5.

l
· Pi,j,EV D (8)
B. EV Service Model j =1

In order to utilize EVs as energy storage devices, the energy 

n 
k

operator must provide benefits to the EV owners. The proposed ECtotal EV = CEV C · Pi,j,EV C
service model for EV charging and discharging is as follows: i=1 j =1
1) This contract is not compulsory. If the EV owner thinks 
n 
l
that this contract is not necessary for him, the EV owner − CEV D · Pi,j,EV D (9)
can cancel the contract. Conversely, the EV owner can sign i=1 j =1
1550 IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 28, NO. 4, APRIL 2013

Fig. 7. Dynamic simulation results (∗N.G: nonutility generator). (a) Load and EV charging profile (Case I). (b) Load and EV charging profile (Case II). (c) Load
and EV charging profile (Case III). (d) Generation and EV discharging profile (Case I). (e) Generation and EV discharging profile (Case II). (f) Generation and
EV discharging profile (Case III). (g) Power output of nonutility generator. (h) Incoming power from utility grid. (i) Incoming and EV discharging power (Case
I). (j) EV charging power.
BYEON et al.: ENERGY MANAGEMENT STRATEGY OF THE DC DISTRIBUTION SYSTEM IN BUILDINGS USING THE EV SERVICE MODEL 1551

where ECi,EV is the energy cost (charging cost) of the ith

EV, ECtotal EV is the sum of the energy cost of all EVs,
CEV C , CEV D are charging/discharging costs per 1 kW, and
Pi,j,EV C , Pi,j,EV D are the jth charging/discharging power of
the ith EV. If the EV is connected to the system, this EV re-
peats the charging and discharging process by the proposed
EMS. The EV owner should pay the cost for the total accumu-
lated charging power (CEV C · kj=1 Pi,j,EV C ), but will re-
ceive an incentive
for the total accumulated discharging power
(CEV D · lj =1 Pi,j,EV D ). The energy operator should deter-
mine the variable CEV D by considering many factors such as
cost savings by peak shaving, battery life issues, loss of charging
opportunities, and so on. If the variable CEV D is determined
properly, the EV owners can support the energy management
for the buildings and receive the benefits for discharging. Also,
EVs operation is affected by the electricity price operation to
prevent the reckless discharging process. EVs operation condi-
tions under the contract are described in Table I. Under these
conditions, EVs can avoid contradictory charging/discharging
process caused by the changing electricity price.

C. Interconnection With the Utility Grid

From the economic point of view, reducing the amount of
total load during peak load periods is the easiest and most effi-
cient way to reduce energy costs because the price of electricity
from the utility grid is more higher than the generation cost of
the nonutility generator. By using an appropriate EV service
model, limiting the charging of EVs and shaving peak load by
EV batteries are suitable solutions. However, factors that need
to be considered such as stabilizing the system and preventing
power outages in buildings are more important than financial
issues. Therefore, the interconnection with the utility grid is
always on standby mode. Despite efforts to reduce peak load,
if the amount of load is larger than the internal power supply,
the power system of building must allow the injection of power
from the utility grid. Therefore, the interconnection with the
utility grid is considered in preference to the EV service model. Fig. 8. Comparison of the SOC changes in three cases. (a) SOC of EVs (Case
I). (b) SOC of EVs (Case II). (c) SOC of EVs (Case III).
However, the amount of this incoming power should be mini-
mized. Besides, if the price of electricity from the utility grid
is lower than the nonutility generation price, it is desirable to
in Table II. It is assumed that this building is a small-scale load
minimize the amount of the nonutility generation and maximize
(maximum load is about 100 kW) and has the EV station that ac-
the incoming power from the utility grid.
commodates eight EVs. Out of the eight EV stations, only seven
EVs enter into an agreement with the building energy operator.
IV. CASE STUDY Coefficients of the NG are selected based on the power mar-
Focused on dynamic behaviors, three simulation cases were ket simulation software, M-core which is made by MASTER
performed and compared to validate the utilization of the pro- SPACE [28]. In this software, we chose “Pankyo co-generator’s
posed EMS in dc distribution buildings through the EV charg- generation and cost curve coefficients” and fuel cost is based on
ing/discharging model. LNG costs on February 2010. The heat demand curve used in
Case I: Peak shaving simulation using the proposed EMS. the case studies is based on the heat demand curve in weekdays
Case II: Simulation using only the charging model. of the real commercial building. To reflect the commute time,
Case III: Simulation using only the charging model but the the PV pattern is implemented based on the real daily traffic
final SOCs of EVs are the same as those of case I. data.
These simulation cases are simulated using MATLAB– The total load profiles for both cases are represented in
Simulink software and each economic analysis is conducted Fig. 7(a)–(c). The basic load profiles for these cases [pink area
simultaneously. The test system to be analyzed is shown in in Fig. 7(a)–(c)] are the same. However, the EV charging power
Figs. 1 and 2. The parameters of its components are represented is different from each other owing to the difference of their SOC
1552 IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 28, NO. 4, APRIL 2013

TABLE III
EV CHARGING/DISCHARGING PROFILE (CASE I)

TABLE IV
EV CHARGING PROFILE (CASE II)

TABLE V
EV CHARGING PROFILE (CASE III)

TABLE VI
ENERGY COST COMPARISON (BUILDING, $)

TABLE VII
ENERGY COST COMPARISON (EV OWNERS, $)
BYEON et al.: ENERGY MANAGEMENT STRATEGY OF THE DC DISTRIBUTION SYSTEM IN BUILDINGS USING THE EV SERVICE MODEL
profile (Fig. 8). The power outputs of the nonutility generator
for three cases are represented in Fig. 7(g). According to the
price of electricity from the utility grid, the optimal genera-
tion output is varied. The output is increased when the price
is high in medium and heavy load periods (from 9 o’clock
onward). During the medium load period, the optimal gener-
ation output is determined as 70 kW. During the heavy load
period, the nonutility generator produces the maximum power
output. If the price becomes lower, the output is decreased (from
23 o’clock onward). The reason of difference between three
cases in Fig. 7(g) (18–21 o’clock) is the additional EV charg-
ing behavior of Case I [Fig. 7(j)]. Fig. 7(i) shows the power of
EV discharging and the utility grid in Case I. During the heavy
load period, the proposed EMS scheme makes EVs discharge
power. The SOCs of these EVs are higher than the defined SOC
value that permits discharging (SOCdischarging ). The discharg-
ing power decreases the amount of incoming power from the
utility grid. Fig. 7(d)–(f) shows the total output power for the
total load of this building in three cases. BIPV and discharging
output power contribute to peak shaving in Case I. However,
only BIPV output power reduces the power from the utility grid
in Cases II and III. In Fig. 7(h), the incoming power from the
utility grid in three cases is represented. This result shows that
the proposed EMS can reduce the incoming power effectively.
The EV charging power graphs of three cases are shown in
Fig. 7(j). Since the selective standby algorithm is not applied in
Case II, the EV charging power of Case II is more than that of
Case I from 10 to 12 o’clock. However, the EV charging power
of Case I is larger than that of Case II from 18 to 21 o’clock ow-
ing to the additional charging. But in Case III, the EV charging
power is smaller than other two cases due to low final SOCs. The
SOC of each EV in three cases is represented in Fig. 8. When
the EV owners connect the EVs to the EMS in the morning, all
the EVs under SOCm in value start charging. Then the EVs that
exceed the SOCm in value continue charging or are standing by
according to the amount of total load and the price of electricity
from the utility grid. During heavy load periods, EVs exceeding
SOCdischarging value start discharging and reduce the incoming
power from the utility grid. Although the EV is charging or dis-
charging, its owner can stop the operation and leave the building.
In Cases II and III [Fig. 8(b) and (c)], all EVs are charging or
are under standby mode without any discharging process. The
connection and disconnection time and the initial and final SOC
of EVs in three cases are represented in Tables III–V. In Cases
I and II, the final SOC of some EVs is different in both cases
because the proposed EMS that initiates EV discharging is only
applied to Case I. In Case I, some EVs experience the charg-
ing/discharging process, so the final SOC is different from that
of Case II. In Case III, all final SOCs of EVs are the same as
those of Case I in order to compare the economic analysis of
Case I. In Case I, seven EVs are contracted to discharge dur-
ing heavy load periods (contracts: H) and one EV (EV 8) is
not contracted for the discharging process. The detailed energy
cost indices of the building are described in Table VI. The cost
savings from producing heat and maintenance fee of the BIPV
system are the same in all cases. In Case I, the generation cost
of the nonutility generator is more higher than that of Cases II
1553
and III owing to the additional charge of EVs. However, the cost
of incoming power from the utility grid is smaller than that of
Cases II and III owing to the EV discharging. In Case II, since
EVs are charging in heavy load time, the EV charging fee is the
highest. By the agreement of the EV service model, the cost of
discharging has been paid to the EV owner in Case I. The charge
to be paid for the charging power and incentives for the discharg-
ing power are represented in Table VII. It is noteworthy that the
EV5 owner paid a low charge for charging the same power by
participating in the discharging process. Other EV owners who
allowed the discharge process also receive the same benefit in
proportion to the discharging power. Especially, the EV which
is charged for many hours in light load time made a lot of profit.
In the case when the building has the EV charging stations, the
proposed EMS can obtain additional benefits. If the number of
EVs is increased or the long-term operation is done, it can bring
forth more benefits to the building and EV owners. In addition
to financial benefits, the power being supplied from utility grid
is being reduced thereby securing more generation reserves and
the discharging process can encourage active participation of
the EV owners in the energy management.
V. CONCLUSION
In this paper, a new EMS of the dc distribution system in build-
ings has been proposed. This strategy is based on the real-time
acquisition data and determines the optimal behavior of com-
ponents to reduce total energy costs. In addition to the EMS,
the service model for EV owners who participate in the energy
management has been also proposed. A comprehensive result
through simulation tests and economic analysis has shown that
the proposed strategy and EV service model can be utilized for
management solutions in the dc distribution system. Since the
proposed EMS does not need any accumulated data and predic-
tion algorithms, the operator and engineer can implement the
algorithm more efficiently thereby resulting in a more effec-
tive operation of the building energy management system. In
future works, a study related to the combination of stability en-
hancement and research on the estimation of discharging incen-
tive coefficients considering the additional charging/discharging
process are required and are under progress.
REFERENCES
[1] K. Fleischer and R. S. Munnings, “Power systems analysis for direct
current (DC) distribution systems,” IEEE Trans. Ind. Appl., vol. 32, no. 5,
pp. 982–989, Sep./Oct. 1996.
[2] B. Han, G. Ledwich, and G. Karady, “Study on resonant fly-back converter
for DC distribution system,” IEEE Trans. Power Del., vol. 14, no. 3,
pp. 1069–1074, Jul. 1999.
[3] P. Karlsson and J. Svensson, “DC bus voltage control for a distributed
power system,” IEEE Trans. Power Electron., vol. 18, no. 6, pp. 1405–
1412, Nov. 2003.
[4] M. Baran and N. R. Mahajan, “DC distribution for industrial systems:
Opportunities and challenges,” IEEE Trans. Ind. Appl., vol. 39, no. 6,
pp. 1596–1601, Nov./Dec. 2003.
[5] H. Zhang, F. Mollet, C. Saudemont, and B. Robyns, “Experimental val-
idation of energy storage system management strategies for a local dc
distribution system of more electric aircraft,” IEEE Trans. Ind. Electron.,
vol. 57, no. 12, pp. 3905–3916, Dec. 2010.
[6] K. Sun, L. Zhang, Y. Xing, and J. Guerrero, “A distributed control strategy
based on DC bus signaling for modular photovoltaic generation systems
1554
with battery energy storage,” IEEE Trans. Power Electron., vol. 26, no. 10,
pp. 3032–3045, Oct. 2011.
[7] H. Zhou, T. Bhattacharya, D. Tran, T. S. T. Siew, and A. M. Khambad-
kone, “Composite energy storage system involving battery and ultracapac-
itor with dynamic energy management in microgrid applications,” IEEE
Trans. Power Electron., vol. 26, no. 3, pp. 923–930, Mar. 2011.
[8] L. Xu and D. Chen, “Control and operation of a DC microgrid with variable
generation and energy storage,” IEEE Trans. Power Electron., vol. 26,
no. 4, pp. 2513–2522, Oct. 2011.
[9] H. Kakigano, Y. Miura, and T. Ise, “Low-voltage bipolar-type DC micro-
grid for super high quality distribution,” IEEE Trans. Power Electron.,
vol. 25, no. 12, pp. 3066–3075, Dec. 2010.
[10] F. Katiraei and M. R. Iravani, “Power management strategies for a micro-
grid with multiple distributed generation units,” IEEE Trans. Power Syst.,
vol. 21, no. 4, pp. 1821–1831, Nov. 2006.
[11] S. Chakraborty, M. D. Weiss, and M. G. Simões, “Distributed intelligent
energy management system for a single-phase high-frequency ac micro-
grid,” IEEE Trans. Ind. Electron., vol. 54, no. 1, pp. 97–109, Feb. 2007.
[12] E. Barklund, N. Pogaku, M. Prodanovic, C. Hernandez-Aramburo, and
T. C. Green, “Energy management in autonomous microgrid using
stability-constrained droop control of inverters,” IEEE Trans. Power Elec-
tron., vol. 23, no. 5, pp. 2346–2352, Sep. 2008.
[13] F. Katiraei, R. Iravani, N. Hatziargyriou, and A. Dimeas, “Microgrids
management-controls and operation aspects of microgrids,” IEEE Power
Energy, vol. 6, no. 3, pp. 54–65, May/Jun. 2008.
[14] J.-Y. Kim, J.-H. Jeon, S.-K. Kim, C. Cho, J. H. Park, H.-M. Kim, and
K.-Y. Nam, “Cooperative control strategy of energy storage system and
microsources for stabilizing the microgrid during islanded operation,”
IEEE Trans. Power Electron., vol. 25, no. 12, pp. 3037–3048, Dec. 2010.
[15] D. Lu, H. Kanchev, F. Colas, V. Lazarov, and B. Francois, “Energy man-
agement and operational planning of a microgrid with a PV-based active
generator for smart grid applications,” IEEE Trans. Ind. Electron., vol. 58,
no. 10, pp. 4583–4592, Oct. 2011.
[16] T. Zhou and B. Francois, “Energy management and power control of a
hybrid active wind generator for distributed power generation and grid
integration,” IEEE Trans. Ind. Electron., vol. 58, no. 1, pp. 95–104, Jan.
2011.
[17] C. Chen, S. Duan, T. Cai, and B. Liu, “Optimal allocation and economic
analysis of energy storage system in microgrids,” IEEE Trans. Power
Electron., vol. 26, no. 10, pp. 2762–2773, Oct. 2011.
[18] M. G. Molina and P. E. Mercado, “Power flow stabilization and control
of microgrid with wind generation by superconducting magnetic energy
storage,” IEEE Trans. Power Electron., vol. 26, no. 3, pp. 910–922, Mar.
2011.
[19] C. Chen, S. Duan, T. Cai, B. Liu, and G. Hu, “Smart energy management
system for optimal microgrid economic operation,” IET Renew. Power
Gener., vol. 5, no. 3, pp. 258–267, May 2011.
[20] M. Camara, B. Dakyo, and H. Gualous, “Polynomial control method of
DC/DC converters for DC-bus voltage and currents management bat-
tery and supercapacitors,” IEEE Trans. Power Electron., vol. 27, no. 3,
pp. 1455–1467, Mar. 2012.
[21] A. Sannino, G. Postiglione, and M. H. J. Bollen, “Feasibility of a DC
network for commercial facilities,” IEEE Trans. Ind. Appl., vol. 39, no. 5,
pp. 1499–1507, Sep./Oct. 2003.
[22] T. Kaipia, P. Salonen, J. Lassila, and J. Partanen, “Possibilities of the low
voltage DC distribution systems,” presented at the NORDAC 2006 Conf.,
Stockholm, Sweden, Aug. 2006.
[23] O. Tremblay, L. A. Dessaint, and A. I. Dekkiche, “A generic battery model
for the dynamic simulation of hybrid electric vehicles,” in Proc. IEEE Veh.
Power Propulsion Conf., Sep. 2007, pp. 284–289.
[24] B. Singh and J. Solanki, “Load compensation for diesel generator based
isolated generation system employing DSTATCOM,” IEEE Trans. Ind.
Appl., vol. 47, no. 1, pp. 238–244, Jan./Feb. 2011.
[25] G. R. Walker and P. C. Sernia, “Cascaded DC–DC converter connection
of photovoltaic modules,” IEEE Trans. Power Electron., vol. 19, no. 4,
pp. 1130–1139, Jul. 2004.
IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 28, NO. 4, APRIL 2013
[26] S. Duan, B. Liu, and T. Cai, “Photovoltaic DC-building-module-based
BIPV system—Concept and design considerations,” IEEE Trans. Power
Electron., vol. 26, no. 5, pp. 1418–1429, May 2011.
[27] A. J. Wood and B. F. Wollenbergy, Power Generation, Operation, and
Control. New York: Wiley, 1984.
[28] Web page for M-Core software of MATERSPACE. (2011). [Online].
Available http://www.masterspace.co.kr/renewal/product/product01.asp
Gilsung Byeon (S’06–M’11) received the B.S. de-
gree from the Department of Electrical Engineering,
Korea University, Seoul, Korea, in 2006, where he is
currently working toward the Ph.D. degree.
His research interests include power system mod-
eling, control, and simulation algorithm analysis.
Taeyoung Yoon (S’2) received the B.S. degree from
the Department of Electrical Engineering, Korea Uni-
versity, Seoul, Korea, in 2010, where he is currently
working toward the M.S. degree.
His research interests include dc power system
modeling, control, and power quality.
Seaseung Oh (S’02–M’08) received the B.E. and
M.E. degrees from Korea University, Seoul, Korea.
He also received the Ph.D. degree from Korea Uni-
versity in 2007.
He was a Postdoctoral Researcher at Seoul Na-
tional University, Seoul, Korea and a Research Pro-
fessor in BK21 Division, Department of Electrical
Engineering , Myongji University, Seoul. He is cur-
rently a Senior Researcher with Korea Institute of
Energy Research, Daejeon, Korea. His research inter-
ests include computer applications in power system
analysis, operation, and control.
Gilsoo Jang (S’95–M’97–SM’06) received the B.S.
and M.S. degrees from Korea University, Seoul, Ko-
rea, and the Ph.D. degree from Iowa State University,
Ames, in 1997.
He was a Visiting Scientist in the Department
of Electrical and Computer Engineering, Iowa State
University, for one year and a Researcher with Ko-
rea Electric Power Research Institute, Daejeon, Ko-
rea, for two years. He is currently a Professor with
the School of Electrical Engineering, Korea Univer-
sity. His research interests include power quality and
power system control.