Content uploaded by Mohsen Ahmadi
Author content
All content in this area was uploaded by Mohsen Ahmadi on Jun 04, 2019
Content may be subject to copyright.
Dynamic Load Control at a Bidirectional DC Fast
Charging Station for PEVs in weak AC grids
Mohsen Ahmadi, N. Mithulananthan and Rahul Sharma
School of Information Technology & Electrical Engineering
University of Queensland
Brisbane, Australia
mohsen.ahmadi@uq.net.au; mithulan@itee.uq.edu.au; rahul.sharma@uq.edu.au
Abstract— Widespread acceptance of Plug-in Electric Vehicles
depends on the presence of charging infrastructures with filling
times close to those of petrol stations. On the other hand, the roll
out of fast charging stations which have significant power demand
would cause several issues in the power grid such as overload,
voltage drop and power loss. Therefore, to avoid costly upgrades
in the existing power grids, it is necessary to have fast charging
stations able to provide fast PEV charging while keeping the
power grid intact. This paper proposes a bidirectional DC fast
charging station with dynamic load control to address voltage
drop and overload issue in weak distribution grids. By dynamic
control of charging process, the proposed charging station
maintains the voltage at the point of common coupling and
prevents overloading in the incoming lines even when the
charging power demand is more than the charging station
capacity.
Index Terms— charging stations, electric vehicles, power
conversion, power distribution, voltage control.
I. I
NTRODUCTION
Plug-in Electric Vehicles (PEVs) are progressively gaining
more interest due to their potential to conserve energy, protect
the environment, and prevent the dependence on fossil fuels.
Hence, replacement of internal combustion engine vehicles
with electric vehicles is going to increase in the coming years
[1], [2]. Many countries have set several targets for the
widespread use of PEVs [2].
Widespread acceptance of PEVs depends on the presence
of well-established infrastructure for fast charging with filling
times close to those of petrol stations.
Currently, most PEV charging is expected to happen
overnight at home which may take 1-17 hours depending on the
PEV battery and the charger [3], [4]. PEVs are a new type of
load with high level of uncertainty in terms of power demand,
location in the power grid and load profile. The uncoordinated
charging of a large number of individual PEVs during peak
demand periods can overload the distribution grid, cause grid
instabilities, power losses and voltage drops [5], [6]. Therefore,
roll-out of fast charging stations is unlikely in the absence of
appropriate load management strategy. This is because fast
charging stations with several charging slots and large power
demands, make the aforementioned problems more severe. This
is of more importance particularly in populated locations with
high power demands or locations with weak AC grids.
Reference [7] investigates the effect of high power demand of
fast charging of PEVs on an existing distribution grid. A
charging station with 8 fast-charging PEVs with a total load of
2 MW is considered in four locations in the grid. The results
showed the number of PEVs which can be charged at the same
time in locations with weak AC grids is lower than 8, even with
a dedicated distribution transformer. Reference [8] investigates
the effect of the simultaneous fast charging of 50 fully depleted
PEV allocated in 5 groups, on the voltage profile of a test
distribution grid. The result shows that for both 400V and 11kV
voltage levels, the system voltages goes below the safe
operation levels. The results provided in [9] based on a
simulation scenario with 8 fast chargers in a rural distribution
grid, also show a 3%-8% voltage variation in adjacent buses.
Therefore, it can be concluded from the results of [7]-[9] that a
charging station must be able to provide voltage and/or power
compensation in the weaker zones of the distribution grid or in
the populated locations with high power demands.
On the other hand, fast charging stations can overcome the
inherent problems associated with PEV charging by controlling
the charging process considering the grid condition and PEVs
power demands. Moreover, they can benefit the grid by
providing ancillary services such as node voltage regulation.
Reference [10] addresses the need for node voltage regulation
by proposing a fast charging station capable of bidirectional
power transfer. The results show that reactive power
compensation allows the charging station to charge 6 PEVs at
40 kW without driving the distribution system beyond safe
voltage limits. Although charging facilities with higher power
demands are introduced in the standards and expected in future,
the effect of them on the operation of the charging station is not
studied. Based on charging levels introduced by the Association
of Automotive Engineers (SAE) [4], Level 3 chargers are called
fast DC chargers and can provide 90kW-240kW of charging
power.
Although the concept and the necessity of fast charging is
well recognised, most of the existing technological
developments and the academic literature focus on slow
charging of individual PEVs. By more fast-charge compatible
PEVs expected to be manufactured in near future, it is important
to provide solutions to prevent adverse effects of the large
demand of fast charging. Hence, this paper proposes a Level 3
DC fast charging station with dynamic load control and
evaluates its performance in a weak AC grid. The proposed
control method dynamically adjusts the charging power of the
PEV chargers based on the total charging power demand, the
level of the Point of Common Coupling (PCC) voltage and the
capacity of the charging station and distribution transformer.
Moreover, it adjusts the reference PCC voltage within the
allowable operational range accordingly, to achieve the highest
possible charging rate considering the grid condition and
charging power demand. Consequently, it can regulate the PCC
voltage and prevents overloading of the charging station and the
distribution transformer, even when the charging power
demand is more than the charging station capacity.
II. DC
CHARGING STATION
Most commercially available DC fast chargers don’t
provide bidirectional power flow with the grid. Whereas the
possible benefits of PEVs can be only realised by a bidirectional
interface. This paper uses the DC fast charging station topology
shown in Fig. 1 that is able to provide bidirectional power
exchange with the grid. The DC fast charging station consists
of a bidirectional AC/DC converter, a DC bus and DC/DC
chargers as illustrated in Fig. 1.
Figure 1. Topology of DC fast Charging Station [9], [10].
There are several topologies proposed for PEV charging
suitable for unidirectional and bidirectional charging at
different power levels [3]. In this paper, a universal two-level
voltage-source converter (VSC) as shown in Fig. 2 is selected
due to its capability in providing bidirectional power exchange
and low distortion operation. By employing proper control
strategy, exchange of the instantaneous real (P), and reactive
power (Q) of the VSC with the AC grid can be controlled
independently. Therefore, not only a unity power factor can be
achieved, but P and Q can be injected to the grid for loss
reduction and voltage support. This makes it a proper interface
for power exchange between the charging station and the grid.
Figure 2. Topology and control of the front-end converter.
Fig. 3 shows an equivalent circuit for connection of the
charging station to the grid. The power flow between the
charging station and the grid can be controlled in two ways [11]
as depicted in Table I. This results in two main approaches for
controlling the exchanged P and Q in the system of Fig. 2,
namely, voltage-mode and current-mode control. The latter
method features an inherent overcurrent and fault protection,
robustness against parameter variation in the AC system and the
VSC system, and better dynamic performance [12]. Therefore,
a current-mode control scheme is used in this paper as shown
in Fig. 4(a) and (b). The real and reactive power, P(t) and Q(t),
at the PCC are:
()=
[()()+()()] (1)
()=
[−()()+()()]. (2)
In (1) and (2), Vpccd, Icd, Vpccq and Icq are converter voltage and
current components in direct axis and quadrature axis,
respectively. In steady state, Vpccq= 0, therefore, P and Q can be
controlled by icd and icq, if the control system fast tracks the
reference currents idref and iqref:
()=
(3)
()=
. (4)
Figure 3. Equivalent circuit of the grid-connected charging station.
TABLE I. P
OWER
F
LOW
C
ONTROL
M
ETHODS FOR THE
C
HARGING
S
TATION
Control
Method
Control
Variables P Q
Voltage-mode VC(t) and δ
×
()
1−
()
Current-mode Ic(t) and θ
×
×()
×
×()
As it can be seen in Fig. 4(b) and (c), the reference real and
reactive powers, Pref and Qref, are derived by regulating the
output DC voltage and the input AC voltage, respectively.
Then, the reference currents are derived based on (3) and (4). In
the next stage, the converter currents, icd and icq , are controlled
using the derived reference currents and then appropriate
modulation signals, md and mq, are produced as shown in Fig.
4(a). Using this control method, the front-end AC/DC converter
is capable of exchanging P and Q bidirectionally and
independent from each other [12]. As it is shown in Fig. 4(b)
and (c), the reference real and reactive powers are limited to
±Pmax and ±Qmax, respectively. The limits for P and Q keeps the
instantaneous power exchange under the maximum capacity of
converter, and also prevents overloading and high current
excursions during transients.
Figure 4. (a) Current-control of VSC (b) Vdc Control (c Vpcc Control.
III. P
ROPOSED
C
ONTROL
A
PPROACH
According to [4], Level 3 DC chargers should be able to
provide 90kW-240kW for battery charging. Therefore, for a 10-
slot Level 3 fast charging station, 900kW to 2400kW may be
needed, in the case of simultaneous charging in all 10 slots. This
may exceed the real power rating of the charging station.
Moreover, particularly when the charging station is connected
to a weak AC grid, excessive current of PEV charging may take
the PCC voltage below the acceptable limit and also make the
charging station unstable. As it can be seen from Fig. 4(c), in
order to maintain the PCC voltage at the desired value, a certain
amount of reactive power, Qref, is required to be injected into
the grid. When there are PEVs being charged at the station,
some current is drawn from the grid, causing more voltage drop
at the PCC. Therefore, more reactive power is needed to be
injected into the grid. When Qref reaches the limits of reactive
power exchange i.e. ±Qmax, the controller cannot further
compensate the PCC voltage and additional increase in
charging demand leads to more voltage drops and eventually
makes the front-end converter unstable.
To avoid these problems, this paper proposes a dynamic
load control method to protect the grid from voltage drop and
the charging station and distribution transformer from
overloading while providing fast charging for the connected
PEVs.
A. Dynamic Load Control
Fig. 5 shows the proposed control loop which is to be used
instead of the original controller of Fig. 4(c). The idea behind
the proposed controller is to adapt the amount of charging
power so that the required reactive power for voltage
compensation, Qref, is kept under the maximum reactive power
capacity of the charging station, Qmax.
As it can be seen in Fig. 5, Qref is compared with Qmax and
the error is fed to a PI compensator to produce the Charging
Factor (CF). The CF represents the percentage of load that
should be reduced in order to maintain the PCC voltage. In this
way, not only the PCC voltage is maintained, but the DC-link
voltage is also maintained. This is because the proposed
controller keeps DC voltage regulator of Fig. 4(b) stable by
preventing excessive voltage drop at the PCC.
Further, the proposed controller adjusts the reference PCC
voltage based on the value of the CF and according to a
heuristically developed look-up table. The output of the lookup
table is set between 0.95 and 1 and this consequently change
the reference voltage in the range of 0.95 p.u.-1 p.u, based on
the grid and the loading conditions. When CF is close to zero
due to light charging power, the output of the look-up table is
close to 1. The output of the look-up table drops quickly
whenever the CF tends to values more than zero. This allows
the charging station to maintain the PCC voltage at the lower
level in the permissible range, when the loading and/or grid
conditions constitute.
B. DC/DC Chargers
The commonly used method to fast charge PEV lithium-ion
battery is constant-current/constant-voltage (CC/CV) method.
This paper uses a charging method based on the constant-
current/reduced-constant-current (CC/RCC) method proposed
in [10] which can safely charge the battery by preventing the
battery voltage from exceeding the maximum battery voltage
and is faster than conventional CC/CV method. Furthermore,
the method in [10] uses one PI controller for both operating
modes instead of two.
The controller applies a constant current to the battery and
monitors the battery voltage. Whenever the battery voltage, Vbat
reaches a predefined value which is lower than the maximum
allowable voltage, Vmax, the reference current, Iref is reduced to
a lower level. This causes a slight drop in the battery voltage
and the battery is charged with a reduced constant current. This
process repeats to top-up the State of Charge (SoC) of the
battery until the reference current reaches zero and the charging
process ends.
This paper uses a modified charging control method which
complements the proposed control method introduced in the
previous section. It uses the CF obtained from the dynamic load
controller to adjust the reference current of the DC/DC charger
in order to regulate the total charging power demand based on
the grid condition. It uses a modified reference current, Ibat_ref
which is the original reference current reduced by CF in
percent. Besides, it uses hysteresis current control instead of PI
current control. Since the current controller as the inner loop
must be faster than the outer dynamic load control loop, the
Figure 5. Proposed dynamic load control loop.
hysteresis control method is chosen. This is due to the proven
superior dynamic performance of hysteresis controller in
comparison to PI method. Since the reference current and the
DC-link voltage are nearly constant, hysteresis current control
gives a fixed switching frequency which is favourable for filter
design. Fig. 6(a) and (b) show the buck DC/DC charger and its
control block diagram. Note that in Fig. 6(b) the saturation
block is used to keep the modified reference current, Ibat_ref,
between 0%-100% of the original reference current, Iref.
IV. E
VALUATION OF
F
AST
DC
C
HARGING
S
TATION
A 1.6 MVA charging station with 10 slots of Level 3 DC
chargers is modelled, as illustrated in Fig. 7. The limits for real
and reactive power exchanges (Pmax and Qmax) is set to 1131 in
the control loops depicted in Fig. 4 and Fig. 5. The switching
frequency and DC-link voltage of the front-end converter are
set to 3.2 kHz and 1600 V, respectively. The charging station is
connected via a dedicated distribution transformer into an 11
kV grid. Another load is also connected to the grid at the PCC.
The PCC is connected to the subtransmission substation
through a 10 km, 50 mm2 overhead line. The allowable voltage
margin at PCC is assumed ±5% according to [13].
Table II shows two types of PEV with different battery
capacities and charging powers. These two PEV types are used
to evaluate the performance of the charging station in PEV
charging and PCC voltage regulation in different loading
situations. The state of charge of the batteries are considered to
be 0% to demonstrate a worst case scenario.
The common battery modules used in PEVs are lithium-ion
batteries. In this paper the model proposed in [14] is used in
simulations due to its adequate accuracy.
Figure 7. The fast charging station connected to weak AC grid.
TABLE II. PEV
T
YPES USED IN
T
HE
S
IMULATIONS
PEV Type Battery
Type
Battery
Capacity
Charging
Power
Initial
SoC
Type I Li-ion 65 Ah 100 kW 0%
Type II Li-ion 130 Ah 200 kW 0%
In the following parts, the charging station is evaluated
under the same load condition with and without the proposed
dynamic load control. The performance of the charging station
is evaluated by using Simulink/SimPowerSytems software.
Because of resource limitations, Simulink cannot simulate the
whole charging process that may take up to 60 minutes.
Therefore, every 10 minutes is represented by 1 second in
charging process.
A. Fast Charging Without Dynamic Load Control
The simulation results of the charging station without
dynamic load control is shown in Fig. 8. As it can be seen in
Fig. 8 (d), prior to t = 0.1 s, the charging station is not connected
to grid and the PCC voltage is 0.96 p.u. At t = 0.1 s, the charging
station is connected to the grid and injects reactive power into
the grid to maintain the PCC voltage at 1 p.u. which is equal to
398.4 V. However, the required reactive power to maintain the
PCC voltage is more than the reactive power limit of the
controller and the PCC is maintained at 0.99 p.u., as shown in
Fig. 8 (b) and (d). As shown in Fig. 8(a), until t = 0.4 s, there is
no PEV connected to the charging station and the absorbed real
power is due to the front-end converter losses. At t = 0.4 s, 7
Type I PEVs are connected to the charging station and draw
750 kW from the grid. There is no change in the reactive power
injection into the grid, since the reactive power controller is
already injecting at its maximum capacity. Therefore, the PCC
voltage drops to 0.92 p.u. which violates the minimum voltage
level. However, the charging station itself is stable and
operating normally. At t = 0.9 s, 4 Type II PEVs are connected
to the charging station which make the charging demand more
than the maximum real power allowed by the controller.
Therefore, the drawn power from the DC-link becomes more
than the injected power into it and thus the controller loses the
control over the DC-link voltage. Consequently, at t = 0.95 s
when the DC-link voltage reaches the minimum voltage
required for normal operation of the converter, the controller
loses its control over real and reactive power as well. This leads
to large fluctuations in power, current and DC-link voltage, as
well as drawing of highly distorted currents from the grid.
As the PEV batteries need less power in RCC mode, the
power demand of PEVs gradually decreases and at t = 3.1 s, the
DC-link voltage starts to recover and the controller becomes
stable again at t = 3.15 s. The charging process finishes at t =
3.45 s. The charging times of the first and the second groups are
2.42 s and 2.55 s respectively (equivalent to 24 and 25 minutes).
Although the charging station gains its stability again, there
are unacceptable undervoltages at the PCC and very large
overcurrents in the power electronic switches, the capacitor and
the incoming line inductors and conductors.
B. Fast Charging With Dynamic Load Control
The simulation results of the charging station with dynamic
load control is shown in Fig. 9. At t = 0.4 s, the total demand of
the connected PEVs is 700 kW, therefore the main controller
tries to provide this amount of real power to the connected
PEVs. Consequently, this causes the PCC voltage to go below
the acceptable level and the main controller tries to inject
reactive power into the grid to compensate this voltage drop.
Since the required reactive power for voltage drop
compensation is more than the capacity of the charging station,
the proposed dynamic load controller, reduces the charging rate
of the PEVs by adjusting the charging factor, CF, as shown in
Fig.10(c). At the same time, the dynamic load controller
reduces the PCC voltage reference, accordingly, within the
Figure 6. (a) Buck converter (b) Hysteresis current-controller.
acceptable margin, as is shown in Fig. 9(c). This provides more
real power available, when it is needed. Finally, the charging
station absorbs approximately 450kW from the grid, as it can
be seen in Fig. 9(a) and 10(a). This is the maximum real power
which can be absorbed while maintaining the PCC voltage
within the permissible operational margin. Therefore, at t = 1.4
s, charging load of the second group of PEVs, doesn’t change
the steady state absorbed real power. However, at the load side,
the charging power available for PEVs are reduced again by
CF, as shown in Fig. 10 (b) and (c).
Figure 8. Simulation results without dynamic load control.
Figure 9. Simulation results with dynamic load control.
Figure 10. (a) Total charging power (b) EV groups charging power
(c) Charging factor.
At t = 4.64 s the charging of the first group enters to RCC
mode and consequently, its charging power gradually
decreases. However, there is no reduction in the power
absorbed from the grid. This is because the released capacity is
now dedicated to the remaining PEVs. The charging of the first
group finishes at t = 5.56 s.
At t = 5.76 s, the PEVs in the second group enter to RCC
mode of charging and the power demand starts gradually
decreasing until t =7 s which is the end of the charging process.
With the proposed dynamic load controller, the SoCs of the
batteries are topped up from 0% to about 95% in both groups in
approximately 52 and 56 minutes.
V. C
ONCLUSION
A fast DC charging station with dynamic load control is
presented. The charging station is able to control the rate of
charging of the PEVs while maintaining the PCC voltage. This
feature is highly desirable when the charging station has high
number of fast charging slots and is connected to a highly
loaded or to a weak AC grid. When the total power demand of
connected EVs are lower than available power allowed by the
dynamic load controller, EVs get charged at highest charging
rate. On the contrary, if the charging power demand is higher
than available power, the dynamic load controller reduces the
rate of charging, accordingly. A 1.6 MVA rated charging
station with ten PEV charging slots connected to a weak AC
grid is simulated with realistic parameters in a worst case
scenario. The results show that by employing the proposed load
control method, fast charging of multiple PEVs with total
power demand higher than the rating of the charging station is
possible, without adverse impact on the PCC voltage.
References
[1] A. G. Boulanger, A. C. Chu, S. Maxx, and D. L. Waltz, "Vehicle
Electrification: Status and Issues," Proceedings of the IEEE, vol. 99, pp.
1116-1138, 2011.
[2] N. Tanaka, "Technology roadmap: Electric and plug-in hybrid electric
vehicles," International Energy Agency, Tech. Rep, 2011.
[3] M. Yilmaz and P. T. Krein, "Review of Battery Charger Topologies,
Charging Power Levels, and Infrastructure for Plug-In Electric and
Hybrid Vehicles," Power Electronics, IEEE Transactions on, vol. 28, pp.
2151-2169, 2013.
[4] SAE International. EV Charging Speeds and Configurations. Available:
http://www.sae.org/smartgrid/chargingspeeds.pdf
[5] C. H. Dharmakeerthi, N. Mithulananthan, and T. K. Saha, "Overview of
the impacts of plug-in electric vehicles on the power grid," in Innovative
Smart Grid Technologies Asia (ISGT), 2011 IEEE PES, 2011, pp. 1-8.
[6] R. C. Green, W. Lingfeng, and M. Alam, "The impact of plug-in hybrid
electric vehicles on distribution networks: a review and outlook," in
Power and Energy Society General Meeting, 2010 IEEE, 2010, pp. 1-8.
[7] M. Etezadi-Amoli, K. Choma, and J. Stefani, "Rapid-Charge Electric-
Vehicle Stations," Power Delivery, IEEE Transactions on, vol. 25, pp.
1883-1887, 2010.
[8] Y. Jia Ying, V. K. Ramachandaramurthy, T. Kang Miao, A.
Arulampalam, and J. Selvaraj, "Modeling of electric vehicle fast
charging station and impact on network voltage," in Clean Energy and
Technology (CEAT), 2013 IEEE Conference on, 2013, pp. 399-404.
[9] D. Aggeler, F. Canales, H. Zelaya-De La Parra, A. Coccia, N. Butcher,
and O. Apeldoorn, "Ultra-fast DC-charge infrastructures for EV-
mobility and future smart grids," in Innovative Smart Grid Technologies
Conference Europe (ISGT Europe), 2010 IEEE PES, 2010, pp. 1-8.
[10] J. Y. Yong, V. K. Ramachandaramurthy, K. M. Tan, and N.
Mithulananthan, "Bi-directional electric vehicle fast charging station
with novel reactive power compensation for voltage regulation,"
International Journal of Electrical Power & Energy Systems, vol. 64, pp.
300-310, 2015.
[11] H. Saadat, "Power System Analysis,(2nd)," ed: McGraw-Hill Higher
Education, 2009.
[12] A. Yazdani and R. Iravani, Voltage-sourced converters in power
systems: modeling, control, and applications: John Wiley & Sons, 2010.
[13] QLD electricity legislation. Available: http://www.legislation.qld.gov.au
[14] O. Tremblay and L.-A. Dessaint, "Experimental validation of a battery
dynamic model for EV applications," World Electric Vehicle Journal,
vol. 3, pp. 1-10, 2009.
