This article has been accepted for inclusion in a future issue of this journal.

Content is final as presented, with the exception of pagination.

IEEE TRANSACTIONS ON SMART GRID 1
Stability of a dc Distribution System for Power
System Integration of Plug-In Hybrid Electric
Vehicles
Mansour Tabari, Student Member, IEEE, and Amirnaser Yazdani, Senior Member, IEEE
AbstractThis paper proposes a method for enhancing the
stability of a dc distribution system that integrates plug-in
hybrid electric vehicles (PHEVs) with an ac power grid. The
dc distribution system is interfaced with the host ac grid via
a voltage-sourced converter and can also embed photovoltaic
(PV) modules. Thus, bidirectional dcdc electronic power con-
verters act as battery chargers and interface the PHEVs with
the dc distribution system, while PV modules are interfaced with
the dc distribution system via unidirectional dcdc converters.
The dc distribution system is expected to be more efcient and
economical than a system of acdc battery chargers directly
interfaced with an ac grid, but it is prone to instabilities due
to the constant-power property of the dcdc converters. Using
a nonlinear control strategy, the proposed stability enhancement
method mitigates the issue of instability by altering the power
setpoints of the battery chargers, bidirectional dcdc converters,
without a need for changing system parameters or hardware.
The paper presents mathematical models for the original and
modied systems and demonstrates that the proposed technique
expands the stable operating region of the dc distribution system.
Simulation results demonstrate the effectiveness of the proposed
method for a study system in the PSCAD/EMTDC software
environment.
Index TermsConstant-power property, dc distribution sys-
tem, dc grid, dc-voltage control, energy storage, plug-in hybrid
electric vehicle (PHEV), smart grid, stability enhancement.
I. INTRODUCTION
T
HE OUTLOOK of a large number of electric vehicles,
including plug-in hybrid electric vehicles (PHEVs), and
their potential impacts on the power system have motivated
much research recently [1], [2]. The U.S. Department of
Energy projects that more than about one million PHEVs will
have been sold in the U.S. by 2015 and the current incentives
by the U.S. government will promote and increase the sales
of PHEVs [3].
In public parking areas where a sizable number of PHEVs
are interfaced with the power grid [4], such ancillary services
as the provision of back-up power for commercial facilities,
Manuscript received November 8, 2013; revised April 9, 2014 and June 10,
2014; accepted June 10, 2014. Paper no. TSG-00839-2013.
M. Tabari is with the Department of Electrical and Computer Engineering,
University of Western Ontario, London, ON N6A 5B9, Canada (e-mail:
mtabarik@uwo.ca).
A. Yazdani is with the Department of Electrical and Computer
Engineering, Ryerson University, Toronto, ON M5B 2K3, Canada (e-mail:
yazdani@ryerson.ca).
Color versions of one or more of the gures in this paper are available
online at http://ieeexplore.ieee.org.
Digital Object Identier 10.1109/TSG.2014.2331558
voltage support [3], frequency regulation [5], peak shaving,
reactive-power support [6], and integration of photovoltaic
(PV) panels [7] can be offered by PHEV batteries. To provide
such services, bidirectional battery chargers must be employed
to enable energy exchange between PHEVs and the host grid.
In most proposed integration strategies, acdc power-
electronic converters act as the battery chargers and are
directly interfaced with the power grid [8]-[9]. As a perceived
technically and economically superior alternative, dc distri-
bution systems have recently been proposed in which dcdc
converters act as the battery charges [10][14], especially for
public areas. A dc distribution system can more efciently
host PV modules and interfaced with the power grid via
a central acdc power-electronic converter. However, due to
the constant-power property of dcdc converters [15], [16], it
becomes unstable if the powers absorbed by the battery charg-
ers exceed certain values [13]. This phenomenon inicts a
limit on the maximum power that can be imported to charge
the batteries and, consequently, precludes full utilization of the
installed capacities and prolongs the charging times. Therefore,
it is imperative to:
1) systematically characterize the phenomenon and identify
the prevailing constraints;
2) devise a stability enhancement technique, in order to
push the limits and expand the stable operating region
of the dc system.
To mitigate the aforementioned issue of instabilities caused
by constant-power elements in a dc distribution system, various
methods have been proposed in the literature, [17][19]. The
method proposed in [17] stabilizes a dc-link electric propulsion
system where a dcac converter drives an induction motor,
by altering the torque setpoint of the motor. The proposed
technique, therefore, is applied to a dc systemwith one constant-
power element; there is no analysis for multiple constant-power
elements. The techniques proposed in [18] and [19] deal with a
system in which a dcdc converter is assumed to be supplying
another constant-power element. However, both techniques
require information about the internal state variables and access
to the pulse width modulation (PWM) signal of the dcdc
converter. Moreover, the studied systems include only one
dcdc converter and one constant-load element.
Expanding upon the idea proposed in [17], this paper pro-
poses a control technique for expanding the stable operating
region of a dc distribution system integrating PHEVs via bidi-
rectional dcdc converters (battery chargers), such that the dc
1949-3053 c 2014 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 inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
2 IEEE TRANSACTIONS ON SMART GRID
Fig. 1. dc bus for distribution of power in a parking lot for PHEVs.
system and its PHEVs can import larger powers from the host
ac grid. The proposed technique is simple, does not require
information internal to the system or its embedded convert-
ers, and does not need hardware modications. Rather, it only
employs local measurements and individual power setpoints
and, therefore, can be exercised in a decentralized fashion.
These, in turn, permit the use of commercially available dcdc
converters (battery chargers), expected to further reduce the
overall cost of the system. The proposed technique is also
applicable to other dc distribution systems, e.g., shipboard
power systems, that have multiple power-electronic converters.
II. SYSTEM CONFIGURATION
Fig. 1 illustrates a dc distribution system, for example, in a
parking lot, that hosts PHEVs and PV modules. In this system,
dcdc converters are utilized as battery chargers for the PHEVs
and also for interfacing the PV modules. Moreover, a central
voltage-sourced converter (VSC) interfaces the dc distribution
system to the host ac grid. A communication network [20] is
used to enable the exchange of metering and control informa-
tion for a management unit, to and from the dcdc converters
and the central VSC. The management unit calculates the lim-
its of the power exchange setpoints and sends them to the
dcdc converters, to ensure that the dc system operates in its
stable operation region. The PHEV owners, on the other hand,
can set state-of-charge (SoC) limits for their vehicles, to per-
mit power exchanges only if the SoC resides within a certain
range. For example, if the SoC is above 70%, then energy
can be sold to the rest of the system, whereas if the SoC is
below 40%, then the vehicle should buy energy from the rest
of the system. The aforementioned limits (determined based
on the trip plans, electricity price, and other factors) translate
into power setpoints for the corresponding dcdc converters.
For example, for a PHEV with 20 kWh of battery capacity, if
the SoC limit for energy export is 70% and the present SoC is
85%, then 3 kWh (that is, 15% of the battery capacity) can be
sold by the PHEV to the rest of the system, meaning that the
power setpoint of this particular PHEV can be set to export
9 kW of power in 20 min, or 3 kW of power in 1 h, and so
on [13].
Fig. 2. Schematic diagram of the dc system.
Fig. 2 shows a simplied schematic diagram of the dc sys-
tem. As the diagram indicates, the dc system consists of a
central VSC, a network of RL branches representing the dc
distribution network, and a multitude (3, here) of dcdc power
electronic converters that represent either the battery chargers
of the PHEVs or the power-electronic interface of the PV sys-
tem. The central VSC is interfaced with the host ac power
grid via a three-phase tie reactor with a per-phase inductance
of L
s
; the resistance R
s
represents the aggregate effect of the
ohmic power losses of the tie reactor and conduction power
losses of the VSC. The central VSC regulates the dc volt-
age of the network. The PHEV battery chargers can draw or
deliver power, the PV system delivers power, and the host
ac grid compensates for any generation-consumption power
mismatch, via the central VSC. A study system, whose param-
eters are reported in Table I, is used for stability analysis and
performance evaluation in this paper.
III. MATHEMATICAL MODEL
A. Central VSC
The central VSC and its control scheme act as a controlled
dc-voltage power port [21] and regulate the dc voltage of the
network (Fig. 3 illustrates the concept). The VSC is current-
controlled, such that its output real power, P
s
, rapidly tracks
the real-power setpoint, P

s
, issued by a dc-voltage regulation
loop (in this paper, variables with superscript * signies the
setpoints). P
dc
denotes the power that the rest of the dc sys-
tem delivers to the VSC. As Fig. 3 indicates, a measure of
P
dc
is incorporated in the control loop, as a feed-forward sig-
nal, to mitigate the dynamic coupling between the dc-voltage
regulation loop and the rest of the dc system.
As discussed in [21], the transfer function from

P
s
to v
dc
is
G
v
(s) =
v
2
dc
(s)

P
s
(s)
=

2
C

s +1
s
(1)
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
TABARI AND YAZDANI: STABILITY OF A DC DISTRIBUTION SYSTEM FOR POWER SYSTEM INTEGRATION OF PHEVs 3
Fig. 3. Simplied block diagram of the controlled dc-voltage power port.
where the time constant is a function of the steady-state
value of P
dc
, i.e., P
dc0
=
2L
s
P
dc 0
3V
2
s
(2)
where V
s
is the peak value of v
s
, the line-to-neutral voltage of
ac grid at the point of common coupling (PCC). As discussed
in [21], is negative (corresponding to a nonminimum phase
plant) if P
dc0
is negative. Therefore, the dc-voltage regulation
loop should be tuned for a case where the maximum amount
of power ows from the ac grid toward the dc system to ensure
stability also in other operating conditions.
In the remainder of this paper, it is assumed that the con-
trolled dc-voltage power port of Fig. 3 rapidly and tightly
regulates v
dc
at its setpoint v

dc
, irrespective of the steady-state
value or transient excursions of P
dc
; it is assumed that the
dynamics of v
dc
decay to zero rapidly.
B. dcdc Converters
The PHEV batteries exchange energy with the dc system
through corresponding bidirectional dcdc power-electronic
converters which are referred, hereafter, to as the battery
chargers. Fig. 4 shows a simplied schematic diagram of a
full-bridge dcdc converter which, with no loss of general-
ity, is assumed to represent a battery charger (e.g., for the
ith PHEV of the dc system). The battery current, i
Bi
, is regu-
lated at its setpoint, i

Bi
, by a feedback control loop in which a
compensator, K
i
(s), processes the error (i

Bi
i
Bi
) and generates
the control signal u
i
. A measure of the battery voltage, V
bati
,
is then added to u
i
, and the resulting signal is divided by a
measure of the network-side terminal voltage of the converter,
v
i
, to generate the pulse-width modulating (PWM) signal of
the converter. In turn, i

Bi
is calculated by dividing the battery
power setpoint, P

Bi
, by V
bati
.
Assuming a fast and accurate current-control loop, the bat-
tery power P
Bi
equals P

Bi
. On the other hand, P
Bi
is almost
equal to the power that leaves the network-side port of the
converter, P
ti
; the approximation is plausible in view of the
typically small battery-side lter resistance R
Bi
and inductance
L
Bi
(due to the typically large switching frequency of the con-
verter), as well as negligible power losses of the converter.
Fig. 4. Schematic diagram of the bidirectional dcdc converter as a battery
charger for i th PHEV.
Therefore, P
ti
P

Bi
. Then, in a steady state, the terminal
capacitance C
i
is effectively open and the power that the bat-
tery charger delivers to the rest of the dc system, P
PHEVi
, is
equals P
ti
and, therefore, P

Bi
.
Hence, P

Bi
should be determined based on value of power
one desires the battery charger to deliver to the rest of the
dc system, i.e., the setpoint P
PHEVi
(Fig. 4). The desirable
fast current control can be ensured by proper design of K
i
(s).
For example, if K
i
(s) is of the proportional-integral (PI) type,
then its gains can be determined for a rst-order closed-loop
transfer function of the form
I
Bi
(s)
I

Bi
(s)
=
1

Bi
s +1
(3)
for which the time constant
Bi
is a design choice.
The current that leaves the positive network-side terminal
of the converter is i
ti
= P
ti
/v
i
, which based on the conclusion
that P
ti
P

Bi
= P

PHEVi
, can be written as
i
ti
=
1
v
i
P

PHEVi
. (4)
Hence, from its network-side port, the converter can be
represented by a dependent current source whose value is
i
ti
= P

PHEVi
/v
i
, as shown in Fig. 5. If P

PHEVi
is an exoge-
nous signal, i.e., if it does not depend on any other variable
of the dc system, then P
ti
is constant (from the viewpoint
of dynamics) and, therefore, the converter presents itself as a
constant-power element to the external world; any change in v
i
is counteracted by a corresponding change in i
ti
, such that P
ti
remains constant at P

PHEVi
. This constant-power property of
regulated power-electronic converters is known for giving rise
to system instabilities [22], as demonstrated in Section IV of
this paper.
In a PV system, the power-electronic interface is internally
regulated so as to maintain the dc voltage of its host mod-
ules at a value that corresponds to the maximum-powerpoint
(MPP) under the prevailing sunlight and temperature condi-
tions. Therefore, the model of Fig. 5 can also be used for
a PV system, except that P

PHEVi
is replaced with P
PV,mpp
.
Based on aforementioned discussions, Fig. 6 represents the
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
4 IEEE TRANSACTIONS ON SMART GRID
Fig. 5. Simplied model of the dcdc converter.
Fig. 6. Equivalent circuit for the analysis of the dc system of Fig. 2.
equivalent circuit of the dc system of Fig. 2; P

PHEVi
repre-
sents the power exchange setpoint of the ith battery charger,
and P
PV,mpp
signies the maximum power of a PV array as
imposed by the corresponding maximum powerpoint tracking
(MPPT) scheme.
Based on Fig. 6, the following nonlinear differential equa-
tions describe the dynamic behavior of the dc system of
Fig. 2:
L
d(i
1
+i
2
+i
3
)
dt
+L
1
di
1
dt
= v
1
v
dc
R(i
1
+i
2
+i
3
) R
1
i
1
(5)
L
1
di
1
dt
L
2
di
2
dt
= v
1
v
2
R
1
i
1
+R
2
i
2
(6)
L
2
di
2
dt
L
3
di
3
dt
= v
2
v
3
R
2
i
2
+R
3
i
3
(7)
C
1
dv
1
dt
=
P

PHEV1
v
1
i
1
(8)
C
2
dv
2
dt
=
P

PHEV2
v
2
i
2
(9)
C
3
dv
3
dt
=
P
PV,mpp
v
3
i
3
. (10)
The steady-state values of the variables of the system can
be calculated by setting the derivatives to zero in (5)(10), and
solving the nonlinear equations for every given set of power
exchange setpoints P

PHEV1
, P

PHEV2
, and P
PV,mpp
.
The linearized state-space representation of (5)(10) can be
expressed by
x = Ax (11)
where x = [

i
1

i
2

i
3
v
1
v
2
v
3
]
T
is the vector of state variables,
and denotes small-signal perturbation around steady-state
operating points, and components of A can be determined
Fig. 7. Stable and unstable regions in the charing mode.
Fig. 8. Modication of the power exchange setpoints of the dcdc converters.
based on resistance and inductance of the cables, the capac-
itance of the converters, and steady-state operating point of
the converters. The state-space model of (11) can be used
for stability analysis of the dc system and in Section IV, it
will be shown that the dc system is prone to instability when
the battery chargers are drawing energy from the ac power
grid.
IV. STABILITY ENHANCEMENT
Constant-power elements have been shown to have detri-
mental effects on the stability of their host dc distribution
systems, if the direction of ow from the distribution sys-
tem toward the constant-power elements [13], [15], [22].
Thus, to identify the boundary between the stable and unsta-
ble regions associated with the study system, an algorithm
has been executed in MATLAB software environment, such
that for any given set of power exchange setpoints P
PHEV1
and P
PHEV2
, the linearized state-space model has been exam-
ined in terms of stability, and the boundary between the
stable and unstable operating regions has been depicted in
Fig. 7, when the PV system generates an aggregate power
of 10 kW.
It is observed that the system becomes unstable if the power
absorbed by the battery chargers exceed the levels marked
by the boundary curve. The instability, in turn, prevents the
battery chargers from operating at their full capacities and,
consequently, results in under-utilization of the investment and
prolongs the charging times.
To mitigate the aforementioned instability issue, a ltered
version of the network-side voltage of each dcdc converter
is used to alter the power exchange setpoint of the converter,
as illustrated in Fig. 8. Thus, the power exchange setpoint of
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
TABARI AND YAZDANI: STABILITY OF A DC DISTRIBUTION SYSTEM FOR POWER SYSTEM INTEGRATION OF PHEVs 5
Fig. 9. Simplied model of the modied dcdc converter.
the ith dcdc converter, P

Bi
, is determined by the following
nonlinear equations:
P

Bi
=

v
i
w
i

PHEVi
(12)
dw
i
dt
=
v
i
w
i

i
(13)
where,
i
and w
i
are time constant, and output of the lter on
the network-side dc voltage, v
i
, respectively.
Fig. 9 shows a simplied model of the modied dcdc con-
verter. The current that ows to the dc system, i
ti
, can be
found as
i
ti
=
1
w
i
P

PHEVi
. (14)
As illustrated in Appendix B, if
i
is large compared to time
scale of dynamics of interconnection cables in the dc network
then the instability of the dc system can be avoided in all oper-
ating points of the system. The steady-state gain of the lter
that converts v
i
to w
i
is unit, therefore, the modication of (12)
will not change the steady-state operating point of the dcdc
converter. It should be pointed out that the proposed technique
modies the power exchange setpoints of the converters, using
local measurements. Thus, it does not require access to any
variables internal to the converters, does not need hardware or
software modications, and can be exercised in a decentralized
fashion. Therefore, it can be implemented using commercially
available battery chargers.
By applying the proposed modication in Fig. 8 to the dcdc
converters of the system of Fig. 2, (8)(10) are replaced by
the following differential equations:
C
1
dv
1
dt
=
P

PHEV1
w
1
i
1
(15)
dw
1
dt
=
v
1
w
1

1
(16)
C
2
dv
2
dt
=
P

PHEV2
w
2
i
2
(17)
dw
2
dt
=
v
2
w
2

2
(18)
C
2
dv
3
dt
=
P
PV,mpp
w
3
i
3
(19)
dw
3
dt
=
v
3
w
3

3
. (20)
Fig. 10. Boundaries between stable and unstable regions for unmodied and
modied systems with different values of
i
.
Fig. 11. Boundaries between stable and unstable regions for different values
of power generation of PV modules.
The linearized versions of (5)(7), and (15)(19) can be
expressed in the state-space form as
x
mod
= A
mod
x
mod
(21)
where x
mod
= [

i
1

i
2

i
3
v
1
w
1
v
2
w
2
v
3
w
3
]
T
is the new vector
of state variables, and the matrix A
mod
is introduced by (22)
and (23); the uppercase letters denote the steady-state values
of the state variables, which can be calculated by setting the
derivatives to zero in (5)(7), and (15)(19).
Stability of the modied system has been checked for the
study system in MATLAB software environment. Fig. 10
shows the boundaries between the stable and unstable oper-
ating regions for different values of
i
. In this case, the PV
system is generating a power of 10 kW and power consump-
tion of PHEVs are variable. Fig. 10 depicts that if the values
of
i
increase then the stable operating region of the DC
system expands. This agrees with the analysis presented in
Appendix B.
It should be noted that an increase in the PV power gener-
ation slightly expands the stable region, as Fig. 11 indicates.
However, the improving effect of an increased PV power gen-
eration is insignicant compared to the effect of the time
constant in the proposed modication. Moreover, PV power
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
6 IEEE TRANSACTIONS ON SMART GRID
A
mod
=

RL
2
L
3
+R
1
(LL
2
+LL
3
+L
2
L
3
)
den
RL
2
L
3
R
2
LL
3
den
RL
2
L
3
R
3
LL
2
den
RL
1
L
3
R
1
LL
3
den
RL
1
L
3
+R
2
(LL
1
+LL
3
+L
1
L
3
)
den
RL
1
L
3
R
3
LL
1
den
RL
1
L
2
R
1
LL
2
den
RL
1
L
2
R
2
LL
1
den
RL
1
L
2
+R
3
(LL
1
+LL
2
+L
1
L
2
)
den

1
C
1
0 0
0 0 0
0
1
C
2
0
0 0 0
0 0
1
C
3
0 0 0
L(L
2
+L
3
)+L
2
L
3
den
0
LL
3
den
0
LL
2
den
0
LL
3
den
0
L(L
1
+L
3
)+L
1
L
3
den
0
LL
1
den
0
LL
2
den
0
LL
1
den
0
L(L
1
+L
2
)+L
1
L
2
den
0
0
P

PHEV1
V
2
1
C
1
0 0 0 0
1

1

1

1
0 0 0 0
0 0 0
P

PHEV2
V
2
2
C
2
0 0
0 0
1

2

1

2
0 0
0 0 0 0 0
P
PV,mpp
V
2
3
C
3
0 0 0 0
1

3

1

(22)
den = LL
1
L
2
+LL
1
L
3
+LL
2
L
3
+L
1
L
2
L
3
(23)
Fig. 12. DC system response when the power ows from the ac power grid
to the DC system.
generation is not deterministic and, therefore, one cannot rely
on it to enhance the stability.
V. SIMULATION RESULTS
The study system was simulated in the PSCAD/EMTDC
software environment. Figs. 12 and 13 demonstrate the per-
formance of the dc system in the normal conditions when the
steady-state operating point lies inside the stable region.
Initially, the system is in a steady state and, as Fig. 12
shows, the controlled DC-voltage power port regulates v
dc
at
480 V. At t = 0.5 s, the PV system starts to generate a power
Fig. 13. DC system response when the power ows from the dc system to
the ac power grid.
of 20 kW. At t = 0.6 s, the batteries in PHEV2 start to absorb
a power of 20 kW, as indicated by P

PHEV2
, and at t = 0.7 s,
the batteries in PHEV1 start to charge with a power of 30
kW, as indicated by P

PHEV1
. Therefore, in a steady state, a
power of about 30 kW ows from the ac power grid to the
dc system. As Fig. 12 illustrates, the dc system remains stable
under the aforementioned sequence of events and the central
VSC regulates the v
dc
at 480 V.
Fig. 13 shows the system performance for a case where the
power ow is from the dc system toward the ac power grid. The
PV system starts to generate a power of 20 kW at t = 0.5 s,
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
TABARI AND YAZDANI: STABILITY OF A DC DISTRIBUTION SYSTEM FOR POWER SYSTEM INTEGRATION OF PHEVs 7
Fig. 14. Responses of the dc system in the rst scenario. (a) Unmodied dc system. (b) Modied dc system with
i
= 0.1 ms. (c) Modied dc system with

i
= 0.2 ms. (d) Modied dc system with
i
= 2 ms.
Fig. 15. Responses of the dc system in the second scenario. (a) Unmodied dc system. (b) Modied dc system with
i
= 0.1 ms. (c) Modied dc system
with
i
= 0.2 ms. (d) Modied dc system with
i
= 2 ms.
as Fig. 13 shows. At t = 0.6 s, the batteries in PHEV2 start to
consume a power of 20 kW that compensates the PV system
generation so no power ows to the ac power grid. At t = 0.7 s,
the batteries in PHEV1 start to generate 30 kW, as indicated
by P

PHEV1
in Fig. 13. Consequently, in a steady state, a power
of around 30 kW ows from the dc system to the ac power grid.
The dc system remains stable and v
dc
is regulated at 480 V for
all of the aforementioned disturbances.
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
8 IEEE TRANSACTIONS ON SMART GRID
Figs. 14 and 15 demonstrate the effectiveness of the proposed
modicationtoenhancethestabilityof thedcdistributionsystem.
Performance of the system has been has been analyzed in
two working scenarios. In the rst scenario, as Fig. 14 shows,
the system is working at an operating point where the power
consumptions of PHEV1 and PHEV2 are 60 kW and 50 kW,
respectively, and the PV system generates a power of 10 kW.
Based on Fig. 10, this operating point situates at the verge of
instability for the unmodied dc system. Thus, a change in the
power consumption of PHEV1, at t = 1.3 s, from 60 kW to 77
kW, leads to instability for the unmodied dc system while for
the modied dc systems with both
i
= 0.1 ms,
i
= 0.2 ms, and

i
= 2 ms, the systems remain stable and converge to the new
setpoints. The instability rst occurs to P
1
, P
2
, and P
3
, while
the dcdc converters try to keep the power levels of PHEV1
and PHEV2 constant at their setpinots. The central VSC keeps
the dc voltage, v
dc
, constant.
In the second scenario, as Fig. 15 shows, the system works
at another operating point where the power consumptions of
PHEV1 and PHEV2 are 60 kW and 50 kW, respectively, and
the PV system generates 10 kW. A change in the power con-
sumption of PHEV1, at t = 1.3 s, from 60 kW to 85 kw,
moves the operating point outside of the stable region for
the unmodied dc system and the modied dc system with

i
= 0.3 ms, as predicted in Fig. 10. Fig. 15 shows that this
transition destabilizes both the unmodied dc system and the
modied dc system with
i
= 0.1 ms, whereas the modied
dcdc converters with
i
= 0.2 ms and
i
= 2 ms remains sta-
ble and converge to the new setpoints. The instabilities in this
case also rst occur to P
1
, P
2
, and P
3
. Figs. 14 and 15 show
that, the response settles more rapidly at its post-disturbance
value as the time constant
i
is increased.
VI. CONCLUSION
A method was proposed for enhancing the stability of a dc
distribution system intended to integrate PHEVs with an ac
power grid. The dc distribution system is interfaced with the
host ac grid via a VSC and can also embed PV modules. Thus,
bidirectional dcdc power-electronic converters act as battery
chargers and interface the PHEVs with the dc distribution sys-
tem, while PV modules are interfaced with the dc distribution
system via unidirectional dcdc converters. It was demon-
strated the proposed stability enhancement method mitigates
the issue of instability by altering the power setpoints of the
battery chargers, without a need for changing system param-
eters or hardware. The paper presented mathematical models
for the original and modied systems and demonstrated that
the proposed technique expands the stable operating region of
the dc distribution system. Simulation studies were conducted
to demonstrate the effectiveness of the proposed method for a
study system in the PSCAD/EMTDC software environment.
APPENDIX A
STUDY SYSTEM PARAMETERS
Table I reports the parameters of the study dc system. The
parameters are used for stability analysis and simulations.
TABLE I
STUDY SYSTEM PARAMETERS
Fig. 16. Proposed modication on a single constant-power element.
APPENDIX B
STABILITY ENHANCEMENT FOR SINGLE
CONSTANT-POWER ELEMENT ON DC LINE
Fig. 16 illustrates the application of the proposed modica-
tion to a circuit consisting of a dc voltage source, an intercon-
nection line, and a constant-power element. In the circuit of
Fig. 16, E represents the source voltage, L and R, respectively
signify the line inductance and resistance, and C denotes the
line-side terminal capacitance of the constant-power element.
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
TABARI AND YAZDANI: STABILITY OF A DC DISTRIBUTION SYSTEM FOR POWER SYSTEM INTEGRATION OF PHEVs 9
The dynamics of the circuit of Fig. 16 are governed by the
following differential equations:
L
di
dt
= Ri +v E (24)

dw
dt
= v w (25)
C
dv
dt
= i
P

w
. (26)
It can be shown that the existence condition for steady-state
operating point of the system is
P

E
2
4R
. (27)
It also can be shown that the steady-state voltage of the
constant-power element, V, is always greater than half of dc
voltage source, E, for all operating points
V
E
2
. (28)
The linearized form of (25) is
C
d v
dt
=

i +
P

V
2
w. (29)
The eigenvalues of the linearized system can be found by
solving the following characteristic equation:
LC
3
+RC
2
+

V
2
L +RC +

+
P

V
2
R +1 = 0.
(30)
Based on RouthHorwiz stability test, the system is stable
if and only if
P

>
V
2
R
(31)
and
P

>
RCV
2
L

V
2
L
(32)
and
R

V
2
L +RC +

> L

1 +
P

V
2
R

. (33)
Considering (27) and (28), inequality (31) is always true
for all operating points of the system. It can be shown that by
choosing as
>
L
R
RC (34)
the inequalities (32) and (33) always hold, and therefore, the
system is always stable in its operating point range. This means
that by choosing a proper value for , based on parameters of
the dc line, the proposed modication stabilizes the dc system
with the constant-power element.
As inequality (34) indicates, a dc system with larger
network-side dcdc converter capacitances and/or smaller
interconnection cable L/R ratios requires smaller values for
the time constant
i
to ensure stability. In the other words,
larger network-side capacitances and smaller cable L/R ratios
have stabilizing effects on the dc system.
REFERENCES
[1] D. P. Tuttle and R. Baldick, The evolution of plug-in electric vehicle-
grid interactions, IEEE Trans. Smart Grid, vol. 3, no. 1, pp. 500505,
Mar. 2012.
[2] K. Clement-Nyns, E. Haesen, and J. Driesen, The impact of charging
plug-in hybrid electric vehicles on a residential distribution grid, IEEE
Trans. Power Syst., vol. 25, no. 1, pp. 371380, Feb. 2010.
[3] K. Sikes et al., Plug-in hybrid electric vehicle market intro-
duction study, U.S. Dept. Energy, Washington, DC, USA,
Final Rep. ORNL/TM-2010/46, Jul. 2010.
[4] W. Su and M. Chow, Performance evaluation of an EDA-based large-
scale plug-in hybrid electric vehicle charging algorithm, IEEE Trans.
Smart Grid, vol. 3, no. 1, pp. 308315, Mar. 2012.
[5] S. Han, S. Han, and K. Sezaki, Development of an optimal vehicle-
to-grid aggregator for frequency regulation, IEEE Trans. Smart Grid,
vol. 1, no. 1, pp. 6572, Jun. 2010.
[6] M. Kisacikoglu, B. Ozpineci, and L. M. Tolbert, Examination of a
PHEV bidirectional charger system for V2G reactive power compen-
sation, in Proc. 25th Annu. IEEE Appl. Power Electron. Conf. Expo.
(APEC), Palm Springs, CA, USA, Feb. 2010, pp. 458465.
[7] Y. Gurkaynak and A. Khaligh, Control and power management of a grid
connected residential photovoltaic system with plug-in hybrid electric
vehicle (PHEV) load, in Proc. 24th Annu. IEEE Appl. Power Electron.
Conf. Expo. (APEC), Washington, DC, USA, Feb. 2009, pp. 20862091.
[8] M. Yilmaz and P. T. Krein, Review of battery charger topologies,
charging power levels, and infrastructure for plug-in electric and hybrid
vehicles, IEEE Trans. Power Electron., vol. 28, no. 5, pp. 21512169,
May 2013.
[9] A. Khaligh and S. Dusmez, Comprehensive topological analysis of con-
ductive and inductive charging solutions for plug-in electric vehicles,
IEEE Trans. Veh. Technol., vol. 61, no. 8, pp. 34753489, Oct. 2012.
[10] Y. Du, X. Zhou, S. Bai, S. Lukic, and A. Huang, Review of non-isolated
bi-directional dcdc converters for plug-in hybrid electric vehicle charge
station application at municipal parking decks, in Proc. 25th Annu.
IEEE Appl. Power Electron. Conf. Expo. (APEC), Palm Springs, CA,
USA, Feb. 2010, pp. 11451151.
[11] S. Bai, Y. Du, and S. Lukic, Optimum design of an EV/PHEV charging
station with dc bus and storage system, in Proc. Energy Convers. Congr.
Expo. (ECCE), Atlanta, GA, USA, Sep. 2010, pp. 11781184.
[12] S. Bai and S. Lukic, Unied active lter and energy storage system for
an MW electric vehicle charging station, IEEE Trans. Power Electron.,
vol. 28, no. 12, pp. 57935803, Dec. 2013.
[13] M. Tabari and A. Yazdani, A dc distribution system for power system
integration of plug-in hybrid electric vehicles, in Proc. IEEE Power
Energy Soc. Gen. Meet. (PES), Vancouver, BC, Canada, Jul. 2013,
pp. 15.
[14] Y. Du, S. Lukic, B. Jacobson, and A. Huang, Review of high
power isolated bi-directional dcdc converters for PHEV/EV dc charg-
ing infrastructure, in Proc. Energy Convers. Congr. Expo. (ECCE),
Phoenix, AZ, USA, Sep. 2011, pp. 553560.
[15] A. Emadi, A. Khaligh, C. H. Rivetta, and G. A. Williamson, Constant
power loads and negative impedance instability in automotive systems:
Denition, modeling, stability, and control of power electronic con-
verters and motor drives, IEEE Trans. Veh. Technol., vol. 55, no. 4,
pp. 11121125, Jul. 2006.
[16] D. Marx, P. Magne, B. Nahid-Mobarakeh, S. Pierfederici, and B. Davat,
Large signal stability analysis tools in dc power systems with constant
power loads and variable power loadsA review, IEEE Trans. Power
Electron., vol. 27, no. 4, pp. 17731787, Apr. 2012.
[17] S. D. Sudhoff, K. A. Corzine, S. F. Glover, H. J. Hegner, and
H. N. Robey, DC link stabilized eld oriented control of electric propul-
sion systems, IEEE Trans. Energy Convers., vol. 13, no. 1, pp. 2733,
Mar. 1998.
[18] A. M. Rahimi and A. Emadi, Active damping in DC/DC power elec-
tronic converters: A novel method to overcome the problems of constant
power loads, IEEE Trans. Ind. Electron., vol. 56, no. 5, pp. 14281439,
May 2009.
[19] A. Emadi and M. Ehsani, Negative impedance stabilizing controls
for PWM dcdc converters using feedback linearization techniques,
in Proc. 35th Intersoc. Energy Convers. Eng. Conf. Exhibit (IECEC),
vol. 1. Las Vegas, NV, USA, Jul. 2000, pp. 613620.
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
10 IEEE TRANSACTIONS ON SMART GRID
[20] P. Kulshrestha, L. Wang, M. Y. Chow, and S. Lukic, Intelligent energy
management system simulator for PHEVs at municipal parking deck in
a smart grid environment, in Proc. IEEE Power Energy Soc. Gen. Meet.
(PES), Calgary, AB, USA, Jul. 2009, pp. 16.
[21] A. Yazdani and R. Iravani, Voltage-Sourced Converters in Power
Systems: Modeling, Control, and Applications. Hoboken, NJ, USA:
Wiley, 2010.
[22] A. Kwasinski and C. N. Onwuchekwa, Dynamic behavior and stabiliza-
tion of dc microgrids with instantaneous constant-power loads, IEEE
Trans. Power Electron., vol. 26, no. 3, pp. 822834, Mar. 2011.
Mansour Tabari (S11) is currently pursuing the
Ph.D. degree in electrical engineering from the
University of Western Ontario (UWO), London, ON,
Canada.
He is a Research Assistant with the Department
of Electrical and Computer Engineering, UWO. His
research interests include design, dynamic modeling,
and control of dc distribution systems and integration
of electric vehicles with the power system.
Amirnaser Yazdani (M05SM09) received the
Ph.D. degree in electrical engineering from the
University of Toronto, Toronto, ON, Canada, in
2005.
He was an Assistant Professor with the University
of Western Ontario (UWO), London, ON. He is cur-
rently an Associate Professor at Ryerson University,
Toronto. His research interests include modeling and
control of electronic power converters, renewable
electric power systems, distributed generation and
storage, and microgrids. He is a co-author of the
book Voltage-Sourced Converters in Power Systems (IEEE-Wiley Press, 2010),
and an Associate Editor of the IEEE TRANSACTIONS ON SUSTAINABLE
ENERGY.