A Grid Interface Current Control Strategy For DC Microgrids
A Grid Interface Current Control Strategy For DC Microgrids
A Grid Interface Current Control Strategy For DC Microgrids
TRANSACTIONS ON ELECTRICAL MACHINES AND SYSTEMS, VOL. 5, NO. 3, SEPTEMBER 2021 249
A Grid Interface Current Control Strategy
for DC Microgrids
Muhannad Alshareef, Zhengyu Lin, Senior Member, IEEE, Fulong Li, Member, IEEE,
and Fei Wang, Senior Member, IEEE
Abstract—In this paper, a grid interface current control [2]. Microgrid components can be connected through direct
strategy is presented for a DC microgrid, which aims to reduce the current (DC) links or alternating current (AC) links. With the
disturbance from PV generation and the load variation to the development of power electronics technologies, DC energy
main grid without a grid interface converter. The grid interface
systems have gained an increasing research interest in the last
current is directly controlled by a battery DC-DC converter
within the DC microgrid. Based on a comprehensive analysis of few years. DC based energy systems have a natural interface to
the battery DC-DC converter and interface current control, the DC renewable generation, energy storage and DC electronics
control system has been mathematically modelled. This enabled loads [3]. Compared with conventional AC systems, DC
two transfer functions to be derived that reflect the dynamic systems have advantages of simpler control, higher reliability
response of the inductor current to the duty cycle variation (inner and efficiency [4]-[6].
loop), and the dynamic response of the grid interface current to
In [7], the effects of a renewable source within the microgrid
the inductor current variation (outer loop). Experimental study
has been done to assess the effectiveness of the proposed control on LV distribution network operation were investigated.
strategy. The experimental results indicate that the proposed Simulations of daily PV and load profiles were made to analyze
control strategy has a good performance to control the grid fluctuations in utility voltage and line losses. This research
interface current without an interface converter, regardless the given valuable insight into the operation of PV microgrids in
variations of both PV and the load conditions. weak LV networks. The findings indicated that the utility
voltage was impacted by PV and load fluctuations, while line
Index Terms— Battery, DC microgrid, DC-DC converter,
losses in the LV network were also observed. The authors
photovoltaic (PV) system.
determined that it was essential to utilize an advanced
controller for eliminating these fluctuations [7].
I. INTRODUCTION In the study conducted in [8], an interleaved interface
converter was designed for the purposes of controlling the
D ISTRIBUTED generation (DG) plays an important role in power flow between a DC microgrid and the DC bus of a main
modern power systems, which includes renewable energy
resources, energy storage systems and innovative types of
DC grid. The controller was designed to be more effective to
prevent PV and load disturbances instead of just tracking the
load, such as electric vehicles. However, without a proper
reference provided. In order to ensure that the DC bus voltage
coordinated operation, it will cause a variety of issues to power
systems, such as degraded voltage profiles, congestion in the remains at a constant level, it is important to ensure that the PV
transmission line and a decrease in frequency reserves [1]. and load variations that are generated within a DC microgrid
To accommodate more DGs in power grids and reduce are managed efficiently. Such variations can lead to a series of
carbon emission, structured microgrids have been recognized voltage variations, which could be transferred to sensitive loads.
as the key energy infrastructure for future smart energy systems However, this study requires a DC/DC interleaved converter to
act as the interface connecting between the DC microgrid and
the main DC grid, which could increase the cost of the system.
Manuscript received August 18, 2020; revised January 22, 2021; accepted The researchers in [9] presented a novel power balancing
March 09, 2021. date of publication September 25, 2021; date of current
version September 18, 2021.
approach that could be used for low voltage DC microgrids.
This work has received funding from the U.K. EPSRC UKRI Innovation The study was aimed at extracting the MPP from the PV, while
Fellowship scheme (EP/S001662/2), and the European Union’s Horizon 2020 simultaneously mitigating the effects of fluctuations of PV
research and innovation programme under grant agreement No.734796. output power on the main AC grid. Additionally, a supervision
M. Alshareef is with Department of Electronics and Communication
Engineering in Al-Qunfudhah, Umm al-Qura University, Makkah, Saudi layer was developed for the purpose of optimizing the usage of
Arabia (email: mjshareef@uqu.edu.sa) the battery energy storage system (BESS), thus compensating
Z. Lin and F. Li are with the School of Mechanical, Electrical and for any increase/decrease in voltage caused by the droop
Manufacturing Engineering, Loughborough University, Loughborough, LE11
3TU, United Kingdom (e-mail: Z.Lin@ieee.org and F.Li@lboro.ac.uk) controller.
F. Wang is with the School of Mechatronic Engineering and Automation, Also, microgrids can significantly benefit via the
Shanghai University, Baoshan District, Shanghai 200444, China (email: participation in pricing between supply and demand. As
f.wang@shu.edu.cn)
(Corresponding Author: Zhengyu Lin) suggested in [10], an aggregator is an organization that collects
Digital Object Identifier 10.30941/CESTEMS.2021.00028 the power generated by multiple microgrids and then sells that
2096-3564 © 2021 CES
250 CES TRANSACTIONS ON ELECTRICAL MACHINES AND SYSTEMS, VOL. 5, NO. 3, SEPTEMBER 2021
power to the utility with the aim of achieving better prices. it is necessary to investigate the dynamic behavior of the
Thus, it is desirable for microgrids to deliver power to the bi-directional DC-DC converter used for batteries. This
utility based on the aggregator’s request, regardless of the requires the analysis and derivation of a small signal model of a
intermittent nature of the renewable generation and load bi-directional converter, as well as the grid interface current
demand variations. control between the DC microgrid and the main DC grid. This
In this study, a DC microgrid interface current control will allow two transfer functions to be derived, which are
strategy is proposed for the DC distribution network to ensure representative of the dynamic response of the inductor current
that PV and load disturbances are not transferred to the main of the battery convertor to the duty cycle variation, as well as
DC grid. The proposed interface current control strategy is the dynamic response of the grid interface current to the
based on the power converter control for the battery energy inductor current variation. Subsequently, two transfer functions
storage within the DC microgrid, and no additional interface will be employed to design the proposed control scheme’s
converter is required between the DC microgrid and the main double loop PI compensators. As suggested in [11], the
DC grid. Consequently, the costs of the system can be reduced bi-directional convertor’s charging and discharging modes
while still maintaining a similar control performance to that share the same transfer function provided by the converter.
demonstrated in [8]. Hence, a unified controller can be tolerated in situations where
The rest of this paper is organized as follows: Section II the two switches are supplemented by a single controller. The
presents the DC microgrid system configuration used in this Boost operation mode is selected in this study for the controller
study. The small signal modelling of the system is described in design. The bi-directional DC-DC converter is shown in Fig. 2.
Section III. Section IV explains the proposed control strategy
along with the compensators design. Section V reports the
experimental findings and the paper ends with conclusions in
Section VI.
II. CONFIGURATION OF THE DC MICROGRID SYSTEM
The block diagram of the DC microgrid system employed in
this study is shown in Fig. 1. The DC microgrid is connected to
the DC main grid without grid interface converter. All the
converters within the DC microgrid work at current mode, so
the DC microgrid as a whole is connected to the DC main grid
at current mode. By control the interface current Im, the power
flow can be controlled between the DC main grid and the DC
microgrid. Fig. 1. The DC microgrid configuration for this study.
Within the DC microgrid, the PV array is connected to the
DC bus via a Boost converter. MPPT algorithm is used for
tracking the maximum power for the PV array, therefore, the
output power from the PV varies with the weather conditions.
The batteries are linked to the DC bus using a bi-directional
DC-DC converter. Load is connected to the DC bus directly,
and the power consumption is changed based on user
requirement.
With the development of DC microgrid technologies, it is
expected that multiple DC microgrids will be connected to form
Fig. 2. The bi-directional DC-DC converter circuit diagram.
a DC main grid. Individual DC microgrid can be connected to
the DC main grids via different ways, for example, using a The general average equation for the inductor voltage of the
bi-directional DC/DC converter to control the power flow power converter can be derived as:
between the DC microgrid and main DC grid as proposed in [8].
In this study, as shown in Fig. 1, a power inductor is used as the di° v° Vdc d% 1 D v±
L
b dc
(1)
interface between DC microgrid and DC main grid. A grid dt L1
interface current control strategy is proposed so that the grid
interface current Im can be controlled without an interface where Vdc , L1 and D denote the steady state variables of DC bus
converter. The variations of PV current IPV and load current voltage, power converter inductor and duty cycle, respectively,
ILoad are considered as the disturbances for the grid interface while, v± % °
dc , d and vb represent the small signal perturbation
current control. variables of DC bus voltage, duty cycle and battery voltage,
respectively.
III. SMALL SIGNAL MODELLING AND ANALYSIS The general averaging equation of the capacitor current of
In order to develop the grid interface current control scheme, the converter can be obtained as follows:
ALSHAREEF et al. : A GRID INTERFACE CURRENT CONTROL STRATEGY FOR DC MICROGRIDS 251
d v± 1 D i°L ? I L d% v±dc V s Vm s
im s dc
dc
(2) (7)
dt C RC sL2 r
where IL, C, R, and D represent the steady state variables of The assumption is made that Vm s remains constant,
inductor current, capacitance and the load resistance of power
meaning that its dynamic response can be ignored when the
converter, and duty cycle whereas, v± ° %
dc , iL and d denote the Equations are modelled.
small signal perturbation variables of DC bus voltage, inductor By taking the derivative of equation (7):
current and duty cycle, respectively.
v±dc s
To simplify the analysis, the battery voltage Vb can be i°
m s (8)
assumed as a constant, so that v° sL2 r
b is zero and can be ignored in
the following analysis. Rearrange equation (8)
The Laplace transfer function of Equations (1) and (2) can be v± °
dc s im s sL2 r (9)
represented as follows: By substituting Equation (9) into Equation (6):
1 D ± Vdc %
°
siL s vdc s d s (3) i°m s sL2 r L1 I L s 1 D Vdc
L1 L1
%
d s L 2
L1Cs 2 1 s 1 D
1 D ° 1 ± I R
sv±
dc s iL s vdc s L d% s (4)
C RC C So, the transfer function of the grid interface current to duty
By rearranging equations (3) and (4), the transfer functions cycle variation ° im s / d% s can be obtained:
of the inductor current to duty cycle variation i° %
L s / d s and °
im s v±dc s L1 I L s 1 D Vdc
±
the DC bus voltage to duty cycle variation vdc s / d% s can be % %
d s d s sL2 r 2 L1 2
IV. THE PROPOSED CONTROL STRATEGY
In the proposed interface current control strategy, the
charge/discharge current of the battery is controlled to prevent
PV and load variations from being transferred to the main grid,
Fig.3. Simplified diagram of the battery converter with the grid interface
current. and control the grid interface current between the DC microgrid
Based on Fig.3 the following Equation can be obtained: and the main DC grid.
252 CES TRANSACTIONS ON ELECTRICAL MACHINES AND SYSTEMS, VOL. 5, NO. 3, SEPTEMBER 2021
To control the grid interface current Im a double loop Fig.5 illustrates the block diagram of the inner current
interface current controller is developed, as shown in Fig.4. controller. The inductor reference current is acquired
I mref is the reference current, which could be based on the from the outer control loop and subsequently moved to the
request from an aggregator or a central controller. The inner current control loop.
reference current varies based on the request, but for a short
period of time, it can be assumed as a constant.
The proposed controller structure is similar to the double
loop voltage control for a single Boost converter. The inner
control loop is responsible for regulating the battery converter
inductor current. Instead of controlling the output voltage, the
outer control loop in this controller is responsible for regulating
the grid interface current.
Fig.5. Inner current control block diagram.
The PV continually operates at maximum power point, and is
therefore considered as a variable current source, whereas the The transfer function of Gid is given as follow:
DC load is regarded as a variable current sink. The variables Gid, i° s V Cs 2 I L 1 D
Gid L dc (13)
Gii, G pi1 , G pi2 , H1, and H2 represent the transfer function of the %
d s L Cs 2 L1 s 1 D 2
1
inner control loop, outer control loop, inner PI compensator, R
outer PI compensator, inner sensor and outer sensor, The inner current control loop compensator G pi1 transfer
respectively. function is derived as:
A comparison is made between the grid interface current Im Ki
and the reference current value I mref ; subsequently, the current G pi1 K p1 1 (14)
s
error is delivered to the PI compensator G pi2 , which generates The open loop transfer function of the inner current control
the required current I Lref for the inner control loop. A loop Gol1 is formulated as:
comparison is then made between the reference value I Lref and Gol1 Gid G pi1 H1 (15)
the inductor current IL and the current error is delivered to the The transfer function of the closed loop gain Gcl1 is:
PI compensator G pi1 . After this, the PI compensator generates
Gid G pi1
the required duty cycle D, which is transferred to the PWM Gcl1 (16)
1 Gid G pi1 H1
generator in order to create switching pulses that correspond to
the battery converter. The bandwidth of the inner current Fig.6 shows the control block diagram, where the battery
controller is fixed at approximately fsw/10, where fsw represents converter parameters used in the control design and
the switching device switching frequency, which is 25 kHz for experimental research are listed in Table I.
this study.
The inner compensated control loop crossover frequency
must be significantly lower than the main switching frequency.
This is because the switching frequency can be rejected along
with its related harmonics within the system control loop.
Additionally, the outer control loop should have a slower
response time in comparison to the inner control loop, as the
former generates the reference for the latter and it must have
greater speed to allow the inner control loop to track the
generated reference by the outer control loop. Therefore, the
bandwidth of the outer current control loop is maintained a
level below that of the inner current control loop.
Fig.6. Battery converter control block diagram.
TABLE I
PARAMETERS AND OPERATING POINTS OF BATTERY CONVERTER
Vb IL Vdc D P R L1 L2 Cdc
(V) (A) (V) (W) (Ω) µH µH µF
the regulation of the output at a frequency lower than the loop
crossover frequency. Hence, the objective was to boost the low
frequency loop gain in order to reduce the steady state error in
the output while maintaining a sufficient phase margin. The
proportional integral (PI) compensator has the capability to
manage this based on the parameter specifications, as the
compensated loop gain of the inner control loop has a crossover
frequency of 2.33 kHz and a phase margin of 63.2℃ . The
calculated parameters of PI are formulated as: K p1 = 0.25 and
K i1 = 1800.
Fig.9. Experimental setup of the DC microgrid system.
Fig.11.Experimental results for step change in load current.
Fig.10.Schematic diagram of the complete DC microgrid system used in the
experimental work.
TABLE II Fig.12. Experimental results for step change in load current under Zoom 1.
SPECIFICATION OF THE DC MICROGRID SYSTEM
Component Type Specification
DC power supply GW Instek GPS-2303 2 A
Battery pack YPC33-12×12 24 V
QPX 1200S 1200-watt
Main DC grid 48 V
DC power supply
DC electronic load EA-EL 2400-25 100 Ω
Branch resistors
- 10 Ω
converter
Bi-directional converter - 200W
experienced a step change and then was regulated to 1 A after
approximately 75 ms.
Fig.13. Experimental results for step change in load current under Zoom 2.
The zoom in experimental waveforms at t = t2 is shown in Fig.
13. It shows that the grid interface current has a step change and current consumes power from the DC bus.
then was regulated to 1 A after approximately 75 ms. As shown in Fig.14, at the beginning, the PV current was 0 A,
The performance of the proposed control strategy was and the battery pack was discharged to supply the grid interface
compared to that demonstrated in [8] which has an interface current of 1 A. At t = t1, the PV current was raised to 2 A.
converter. It shows that the control in [8] has a different Consequently, the Im (grid interface current) is increased from 1
response time vary from 20 ms to 150 ms, which means that the A to approximately 3 A. To maintain the grid interface current
proposed control strategy in this paper has similar performance Im at 1 A, the battery was changed to operate in charging mode
as [8], and exhibits adequate performance in terms of the using a charge current of approximately 1 A for the purpose of
rejection of the load current disturbances, while effective absorbing the excess PV supply and regulating the Im according
regulation of the grid interface current can be achieved without to its reference value. At t = t2, the PV current was returned to 0
an interface converter. A, and the battery worked as discharging mode with a discharge
current of approximately 1 A.
B. Case2: Step Change in PV Generation Fig.15 shows the zoom in experimental waveforms at
The aim of this case study is to assess the effects of PV approximately t = t1. It indicates that the grid interface current
disturbances on the grid interface current controller. experiences a step increase change and is then regulated to 1 A
In the experiment, a reversed channel scope was used to read after approximately 50 ms.
the PV current as the assumption was made that the direction of Fig.16 illustrates the zoom in experimental waveforms at
the PV current was opposite to that of the load current. As the approximately t = t2. It indicates that the grid interface current
PV current injects power to the DC bus, whereas, the load experiences a step decrease change and is regulated after
ALSHAREEF et al. : A GRID INTERFACE CURRENT CONTROL STRATEGY FOR DC MICROGRIDS 255
University. His research interests include photovoltaic Fei Wang (S’07-M’11-SM’16) received
modeling and control, intelligent control, nonlinear systems the B.Sc. degree in electrical engineering
control, and optimization techniques such as genetic algorithm, and the M.Sc. degree in power electronics
particle swarm optimization, and control and protection of DC from Zhejiang University, Hangzhou,
microgrid. China, in 2002 and 2005, respectively, and
the Ph.D. degree in power electronics from
the Eindhoven University of Technology,
Eindhoven, The Netherlands, in 2010. He
Zhengyu Lin (S’03-M’05-SM’10) was with the Philips Lighting Electronics
Global Development Center, Shanghai, China, from 2005 to
received the B.Sc. and M.Sc. degrees
2006.
from the College of Electrical
He has been a Faculty Member with the School of
Engineering, Zhejiang University,
Mechatronic Engineering and Automation, Shanghai
Hangzhou, China, in 1998 and 2001,
respectively, and the Ph.D. degree from University, Shanghai, since 2010, and became a Professor in
Herio-Watt University, Edinburgh, U.K., March, 2018. He has authored/co-authored more than 100
technical papers, one academic book, and also 13 authorized
in 2005. He is currently a Reader in the
invention patents. His current research interests include
School of Mechanical, Electrical and
distributed generation, electrical drives, power quality, LED
Manufacturing Engineering, Loughborough University,
drivers, and smart grid.
Loughborough, UK.
His research interests include power electronics and its
applications in renewable energy, energy storage, motor drives,
microgrids, and multi-energy systems. He is currently holding
an EPSRC UKRI Innovation Fellowship on DC microgrids.
Fulong Li (S'16-M'20) received the B.S.
degree in electrical engineering from
Yangzhou University, Yangzhou, China, in
2015, and Ph.D. from Aston University,
Birmingham, UK, in 2019. He is currently a
Research Associate in Loughborough
University, U.K.
His current research interests include control
and stability analysis of power electronics
converters applied in DC microgrids and distributed power
systems, and energy management system design of microgrids.