Received: May 20, 2017 228

Droop Control of Bi-Directional DC-DC Converter for Improved Voltage

Regulation and Load Sharing in DC Microgrid

Narasimha Prasad Tulasi1* Lakshmi Devi Aithepalli2

1
Department of Electrical & Electronics Engineering,
PVP Siddhartha Institute of Technology, Vijayawada, Andhra Pradesh, India
2
Department of Electrical & Electronics Engineering, S.V. University,
Tirupati, Andhra Pradesh, India
* Corresponding author’s Email: narasimhaiete@gmail.com

Abstract: DC microgrids are gaining popularity as most of the residential loads like Televisions, BLDC fans, LED
bulbs, Mobile phones and other electronic gadgets are DC in nature, but to connect Distributed DC generator (DG) to
DC bus, DC-link voltage is to be maintained at the reference value. In the proposed work, a transformer coupled dual
input converter control strategy is modified by replacing existing Perturb & Observe Maximum power point tracking
controller with 𝑉 − 𝑑𝑝/𝑑𝑣 based droop controller, which amalgamates DC bus voltage regulation and Maximum
power point tracking (MPPT) of PV source and battery state of charge (SOC) in single controller, and eluding
controller switching between voltage regulation and MPPT modes. The proposed system will also work as
Decentralized droop control for load sharing, and applicable for both islanded and grid-connected operations. Load
Power sharing between DG sources is achieved according to its droop coefficient values in DC side of the inverter,
thus suitable for DC Microgrid, and results are verified by simulating the above system in MATLAB/Simulink.
Keywords: Isolated type transformer coupled boost converter, Single phase VSI based PV system, Solar and battery
based PV system, DC microgrid, Droop control, −𝑑𝑝/𝑑𝑣 , Load sharing.

meet the load requirements under low light

1. Introduction conditions. Thus, it needs another dc-dc converter for
battery charge/discharge control, and for conversion
Photovoltaic (PV) source is one of the significant
of DC to AC, one more converter is required [1]. But
players in the world’s energy portfolio, which will be
this increases switching losses and overall cost of the
one of the biggest contributions to electricity
system and reduces reliability. When isolation is non-
generation among all the renewable sources. Hence,
uniform, ordinary MPPT algorithms doesn’t hold
it is a challenge to supply stable and continuous
good, complex MPPT algorithm are required [2]. In
power using these sources. This can be addressed by
[3], an isolated multiport dc–dc converter is discussed
efficiently integrating energy storage elements.
for simultaneous power management of multiple and
Further, such systems must require power electronic
different types of renewable energy sources and
converters to form an interface between the PV array,
converter, which are useful for MPPT control of a
the battery and the load/grid.
wind/solar hybrid generation systems. A method in
The main focus of the work is on DC bus voltage
which the use of separate dc-dc converter for MPPT
regulation and load sharing between distributed
action is avoided and is projected in [4], Similar
generators (DG) for DC microgrid. Solar irradiation
approach has also been reported in [5], for application
is time varying in nature and there is a necessity to
in grid-connected scheme. Nevertheless, the
track its maximum power point (MPP) all the time for
aforesaid schemes have to bypass the PV array by a
effective utilization. It requires a DC-DC converter,
diode and an inductor when PV power goes to zero.
backup storage such as a battery is also necessary to
International Journal of Intelligent Engineering and Systems, Vol.12, No.3, 2019 DOI: 10.22266/ijies2019.0630.23
Received: May 20, 2017 229

This results in overall gain reduction as PV and frequency domain according to their dynamic time
battery are connected in series. In [6, 7], a transformer scales, but this is also mode based adaptive droop
coupled dual-input converter is proposed in order to control scheme. Intermittency in solar or wind
tackle these issues of MPPT and battery charging requires battery storage system. If system rating is
/discharging control and it is followed by a high, supported storage system cost also increases to
conventional full bridge inverter, meant for meet the connected loads in the microgrid. A new
residential applications. But in this control strategy, configuration is proposed in [16] involving a series
control switching is done between PV and battery to DC electric spring (series ES), photovoltaic (PV)
maintain DC bus voltage regulation. In [8], to feed panels, and a load to lessen the battery storage
the dc loads, a low capacity multi-port converter for capacity of DC microgrids. But it includes tedious
a hybrid system is presented. This has PV array and mathematical analysis and results are based on
battery linked in series and is intended for the optimizing the parameters.
photovoltaic powered lighting system. However, the PV Storage independent microgrid is proposed in
existence of resonant elements make it sensitive to [17] where MPPT based droop control is explained.
parameter variation, acceptable variation in the duty Active and reactive power reference based load
ratio of the switches is constrained within a specific sharing, conventional droop control and MPPT based
range, and voltage gain is rather limited. Typically, droop control are proposed to extract power from
load sharing amongst different DGs is based on its DGs and send to load/grid, but here PV and storage
individual power ratings. In microgrids, control is are considered independently [18-20]. A multi-input
separated as primary, secondary and tertiary controls transformer coupled bidirectional DC-DC converter
[9]. Tertiary control is the system level control which is proposed in [21-24], here DC voltage regulation is
does the energy management within the microgrid taken care by Inverter in grid-connected mode. But in
and provides essential power references for different Islanded mode of operation, battery and DC-DC
DGs. Secondary control takes power references from converter have to look out and PV is forced to operate
the above i.e., tertiary level and controls power by in non-MPPT mode to maintain voltage regulation.
regulating the bus voltages and shares the current Here, there is variation in dc link voltage for load
among DG sources according to their rating, and changes and also a mode switching is used for
Primary control is controlling voltage or current of a reference input to MPPT voltage. In [23], a new
particular DG, i.e., it offers device level control for control strategy for MPPT tracking is proposed for
power converters. The Power developed by one DG maintaining voltage regulation in addition to the
reduces due to its inherent weather dependent nature, maximum power point tracking (MPPT) in DC
causes unequal voltages, and reduces system load microgrids, which addresses the problem of mode
sharing accuracy. In [10] the slope of the droop curve switching. This control configuration avoids
is adjusted to circumvent this problem. In [11, 12] reconfiguration of control during PV mode switching,
Master-slave control technique is projected for droop but here a separate DC-DC converter is used to aid
control in Islanded microgrids, but if master fails, the battery to supply loads when PV output is reduced
complete microgrid will be shutdown. A noise- than load requirement. In [18], MPPT based droop
resilient voltage observer is presented in [13] to take control strategy is proposed to maintain load sharing
the mean voltage across the microgrid, this value is between DGs, but here also two separate DC-DC
set to voltage regulator to produce a correction factor converters are used.
and adjust the present voltage set point which makes In this paper, a transformer coupled bidirectional
the voltage level preserved in microgrid as desired by DC-DC converter topology [16] is considered with
tertiary control, here difficulty in cyber layer makes solar and battery as inputs and a modified 𝑣 −
the system to lose its control. A multi-time scale 𝑑𝑝/𝑑𝑣 droop control strategy is implemented to
droop based power management scheme is proposed avoid mode switching between MPPT and voltage
in [14], but mode selection switch is there in it and its regulation modes, and simultaneously does MPPT,
time-based control has some limitations in randomly battery SOC control, improve DC bus voltage
changing irradiation patterns in solar PV power regulation and under variations in load compared to
extraction. Above works are based on centralized earlier works. The control system is also worked as
control of microgrid, a decentralized method is decentralized droop control to maintain load sharing
proposed in [15], here various energy sources with between Distributed Generators in DC side of the
fast and slow response are combined into a dc inverter in islanded and grid-connected operating
microgrid in a coordinated and flexible manner by mode, thus this proposed control strategy is suitable
shaping the converter output impedances in for DC Microgrid.

International Journal of Intelligent Engineering and Systems, Vol.12, No.3, 2019 DOI: 10.22266/ijies2019.0630.23
Received: May 20, 2017 230

PV Array

Bi- Synchro
LCL
Battery Diarectional nizing
DC/AC Filter
DC-DC Switch
Converter Grid

L
o

DC Bus
a
V-dp/dv Droop d
DG-1 Controller

DG -2

DG -n

Figure. 1 Basic structure of Distributed Generator with PV system with Battery Backup for Islanded/Grid-connected
operation

𝑑𝑐 𝑣
2. System operation 𝐼𝑝𝑣 = {𝑁𝑝 𝐼𝑝ℎ − 𝑁𝑝 𝐼𝑟𝑠 {𝑒𝑥𝑝 (𝑞 𝐴𝑘𝑇 (𝑣𝑑𝑐 +
The proposed system consists of a PV array, a
𝑁𝑠 𝑅𝑠𝑒 𝐼𝑝𝑣
DC-DC converter to accommodate PV source, a
𝑁𝑠
))}} − 𝐼𝑟𝑠ℎ (1)
battery and its control structure. Section 2.1 to 2.3
discusses PV array, converter operation, existing and
modified MPPT control strategies, and section 3 Where 𝑣𝑟 is the cell reference temperature, 𝐼𝑠𝑐𝑟 is the
illustrates the existing and proposed strategies. short-circuit current of one PV cell at the reference
temperature and irradiation level, and 𝐾𝑣 is a
2.1 PV array temperature coefficient. 𝑣𝑑𝑐 is PV arrays voltage, 𝑁𝑝
and 𝑁𝑠 are the number of PV cells connected in
The commonly accepted solar cell model is a one
parallel and in series respectively, 𝐼𝑝ℎ is the photo-
diode model [25]. When cell is exposed to light, the
DC current is generated. The typical equivalent current of a single solar module which is proportional
circuit of PV cell is shown in Fig. 2. PV arrays consist to both illumination and surface area, 𝐼𝑟𝑠 is the
of ‘Ns’ series connected solar cells per string and ‘Np’ saturation current of diode which is related to
such strings are connected in parallel which form a temperature, k is Boltzmann constant (1.38 × 10−23
PV array. Thus, the equivalent circuit is represented J/K), q is electric charge (1.6 × 10−19 C), T is
by a current source parallel to an ideal diode. The operating temperature (in kelvins), A is P-N junction
mathematical model of PV arrays is given as ideal factor. 𝑅𝑠 is series resistance of the single solar
module, and 𝐼𝑟𝑠ℎ is the current flowing through the
𝑆 shunt resistance. Ignoring the power absorbed by the
𝐼𝑝ℎ = [𝐼𝑠𝑐𝑟 + 𝑘𝑣 (𝑣 − 𝑣𝑟 )] inductance and the resistance of the Voltage source
1000 converter (VSC) interface, the generally accepted

International Journal of Intelligent Engineering and Systems, Vol.12, No.3, 2019 DOI: 10.22266/ijies2019.0630.23
Received: May 20, 2017 231

solar cell model is a single diode model of the solar Ig ID

cell to model the Kyocera KC200GT solar array. 

2.2 Principle of operation of converter PV Cpv Vpv
s1 Ipri Ip Isec D1 C2
Array
  Im
A dual input DC-DC converter is used for IL L VDC
simultaneous maximum power point tracking control VLm Vpri Vsec
  VL  
(MPPT) of a solar PV generation system consisting Ibatt Vbatt 
Cb
s2 Ic1 1: n C3
of photovoltaic (PV) panels and battery, operation  C1 VC D2

and control of this topology is given in [21, 22], In 


this paper DC-DC converter is shown in Fig. 3, with Figure. 3 Circuit diagram of DC-DC converter with
Solar and battery inputs is considered for power battery backup
transfer from PV and Battery to load/Grid and its
working is explained in three modes.
-conducting. The rest of the operation is same as that
Mode I: of mode I. This mode continues till 𝑖𝑝𝑟𝑖 becomes
a) Operation of the converter when inductor current equal to (−𝑖𝐿 ).
is positive: when SW1 is turned on, the PV array
voltage 𝑣𝑝𝑣 is appeared across inductor L and the Mode II:
inductor current,𝑖𝐿 increases. During this period, the (a) Operation of the converter when inductor current
voltage appeared across the primary winding of the is positive: This mode begins when SW1 is off and
transformer is, SW2 is on. At the starting, 𝑖𝐿 is positive and as SW1
is off 𝑖𝑝𝑟𝑖 is zero. Since 𝑖𝐿 >𝑖𝑝𝑟𝑖 , the diode D2 starts
𝑉𝑝𝑟𝑖 = (𝑉𝑏𝑎𝑡𝑡 + 𝑉𝑝𝑣 − 𝑉𝑐1 ) (2) conducting. The voltage appeared across L is, 𝑣𝐿 =
−𝑣𝑏𝑎𝑡𝑡 and hence 𝑖𝐿 starts decreasing. The voltage
appeared across the primary winding of the
Wherein 𝑣𝑏 is the battery voltage and 𝑣𝑐1 is the
transformer is, 𝑣𝑝𝑟𝑖 = 𝑣𝑐1 and hence 𝑖𝑝𝑟𝑖 becomes
voltage across the capacitor, C1. Hence the primary
current of the transformer, 𝑖𝑝𝑟𝑖 increases and the negative and starts falling thereby discharging C1.
The current flowing through the secondary winding
capacitor, C1 gets charged. The current flowing
of the transformer 𝑖𝑠𝑒𝑐 reverses and the diode D4
through the secondary winding of the transformer
gets on. The capacitor C3 is getting charged and the
𝑖𝑠𝑒𝑐 also increases. The diode, D3 is forward biased
voltage across C3 is𝑣𝑐3 = 𝑛𝑣𝑐1 . During this mode,
and the capacitor, C2 gets charged. If ‘n’ is the turn
𝑖𝐿 > (−𝑖𝑝𝑟𝑖 ) and diode D2 is forward biased. This
ratio of the transformer. The voltage across C2 is
given by, mode continues till 𝑖𝐿 becomes equal to−𝑖𝑝𝑟𝑖 .
(b) Operation of the converter when inductor current
𝑉𝑐2 = 𝑛(𝑣𝑏𝑎𝑡𝑡 + 𝑣𝑝𝑣 − 𝑣𝑐1 ) (3) is negative: When𝑖𝑝𝑟𝑖 becomes greater than −𝑖𝐿 the
diode D1 is reverse biased and the switch, SW1 starts
b) Operation of the converter when inductor current conducting. The rest of the operation is same as that
is negative: This mode begins when SW1 is turned on of mode I.
and SW2 is turned off. At the time of starting, 𝑖𝐿 is
negative and, 𝑖𝑝𝑟𝑖 is zero. Hence the diode D1 starts Mode III:
(a) Operation of the converter when inductor current
is positive: When 𝑖𝐿 becomes smaller than −𝑖𝑝𝑟𝑖 , the
Np Ns


Rse diode D2 is reverse biased and the switch, SW2 starts


Np
 conducting. The rest of the operation remains the
same as that of mode II.
(b) Operation of the converter when inductor current
NS is negative: This mode begins when SW1 is turned off
Ns vdc
N p I ph Np
Rsh and SW2 is turned on. During this mode, both 𝑖𝐿 and
𝑖𝑝𝑟𝑖 are negative and the switch SW2 conducts. The
negative current in the primary winding of the
 transformer results in negative current in the
secondary winding of the transformer. Hence the
Figure. 2 Equivalent circuit of PV Cell

International Journal of Intelligent Engineering and Systems, Vol.12, No.3, 2019 DOI: 10.22266/ijies2019.0630.23
Received: May 20, 2017 232

diode D4 is forward biased and the capacitor C3 gets battery charging/discharging, using same converter
charged. During operation in this mode, 𝑣𝐿 = and in view of the same configuration. In this, the
−𝑣𝑏𝑎𝑡𝑡 , 𝑣𝑝𝑟𝑖 = −𝑣𝑐1 and 𝑣𝑐3 = n𝑣𝑐1 . The voltage, modified MPPT control strategy [6,7] is used and
𝑣𝐿 across the inductor, L can be expressed as operating region of this MPPT technique is as shown
mentioned in the Eq. (4) in Fig. 4(b). The configuration is more realistic for
residential DC microgrid application as every house
𝑉𝐿 = 𝑉𝑝𝑣 (4) has its own solar PV and battery installation and they
need to connect to a centralized DC bus to feed its
When the switch SW1 is ON generated power, and if any DC loads exist with them
they connect directly to DC bus, and for ac loads they
𝑉𝐿 = 𝑉𝑏𝑎𝑡𝑡 usually connect it to the utility grid, DC appliances
(loads) utilization reduces power losses in utility grid,
When the switch SW1 is OFF also DC microgrid can feed customers in Islanded
mode when there is a fault on AC side, thus,
Therefore, the average voltage drop across the uninterrupted power is supplied to the customer.
inductor is, Decentralized control of DGs can be possible by
droop control of DC-DC converters connected to DC-
𝑉𝐿 = 𝐷𝑉𝑝𝑣 − (1 − 𝐷)𝑉𝑏𝑎𝑡𝑡 (5) bus, based on the m-value specified for each DG, the
load is shared between different DGs as per the droop
Where in D is the duty ratio of the switch SW1. coefficient ‘m’, expressions for how it shares the load
Equating average is discussed in section 2.3.2.

1−𝐷
𝑉𝑝𝑣 = ( 𝐷
) 𝑉𝑏𝑎𝑡𝑡 (6) Operating Region
Pmpp
PV Power (watts)

From Eq. (6) it can be inferred that the PV voltage

Maximum power point
can be controlled by manipulating D as battery
voltage 𝑣𝑏𝑎𝑡𝑡 can be assumed to be a stiff source.
Therefore, MPPT operation of PV array can be
achieved through a proper manipulation D. The
average output voltage of the above circuit is, 𝑣𝑑𝑐 is
given by,

𝑉𝑑𝑐 = (𝑉𝑐2 + 𝑉𝑐3 ) (7) Vmpp

Voltage (volts)

2.3 MPPT Control (a)

2.3.1. Existing MPPT control Maximum power point(MPP)

Pmpp
PV Power (watts)

In previous work [6], when battery is fully

charged and PV power generation is more than load
power requirement, PV is forced to operate in non Operating Region
dp/d

MPPT mode where mode switching, i.e., reference

v<0
0

voltage to MPPT is changed to maintain voltage

>
v
/d
dp

regulation, thus, there’s a switching between MPPT

to non-MPPT mode. The region of operation of this
MPPT is in and around maximum power point based
on perturb and observe algorithm, i.e., the operating Vmpp
voltage is changing around actual VMPP (voltage at Voltage (volts)

maximum power point) and the operating region on (b)

PV characteristics is as shown in Fig. 4 (a). The Figure. 4 Operating voltage: (a) conventional MPPT and
proposed control strategy unifies DC bus voltage (b) modified MPPT
control, Maximum power point (MPPT) control and

International Journal of Intelligent Engineering and Systems, Vol.12, No.3, 2019 DOI: 10.22266/ijies2019.0630.23
Received: May 20, 2017 233

2.3.2. Modified MPPT control  PV source pumps maximum available power

to microgrid up till DC microgrid is grid
In this control technique, MPP is achieved by (utility) connected or as long as there is a
controlling the operating point with respect to the sufficient load connected when it is in
slope of the right side portion of PV curve which has islanded operation.
a negative slope region on PV characteristics. The  The PV sources can export its maximum
slope of the curve is changing from MPP (maximum available power to the microgrid as long as
power point) to Voltage axis where power is zero. the DC-microgrid is grid-connected or there
Thus, if the slope is chosen as the control variable, by is sufficient load when it is islanded.
changing slope, the operating point can be changed,
 No need of Mode transition of controller
in this way MPP can be reached, at the maximum
during PV operation.
power point, the slope becomes zero (𝑑𝑝/𝑑𝑣 =0).
 As Droop control is decentralized in nature,
no communication system is required
Proposed control strategy caters following benefits:
between DGs (Distributed generators)
 𝑑𝑝/𝑑𝑣 Value does not change with
 PV source regulates DC bus voltage
Irradiation and Temperature.
automatically and takes part in load sharing
along with other PV sources.
Block-3

Block-2 If(B>A)
If(B>A)
Block-1 vmpp I Lmax A O/P=A
vPV X
If v PVref o/p=A
else
else if I Lref
MPPT
Algorithm i PV Y
(Y>0)
o/p=X; o/p  PI I L* B
if(B<=C)
(B<=C)
O/P= o/p 
i PV Z 
Controller-I o/p=C
Shutdown
else 
PI
Controller-II D
VPVr else v PV ILmin C
else
o/p=B
o/p=Z; O/P=B IL

Block-4 Ib max 
vb Battery  I Lmax
Current
Ib limit
Estimator
Ib min

 ILmin
v dcmeas IPV
K 

vbatt
(a)

Block-1 Block-2

If(B>A)
Ib max  IL max A If(B>A)
O/P=A
vb Battery  IL max o/p=A
elseif
else I Lref
Current I L* B
if(B<=C)
(B<=C)

d
Ib limit
Ib min O/P= o/p
o/p=C PI
Estimator  Shutdown
else  Controller-II
IL min IL min C
else
o/p=B
 O/P=B IL

0  dp 
 
vref   dv  ref
 m Block -3
-

v dcmeas

vPV dp/ dv 
PI
Controller-I
I L*
i PV generator dp
dv
(b)
Figure. 5 (a) existing control structure of transformer coupled DC-DC converter and (b) existing control structure of
transformer coupled DC-DC converter

International Journal of Intelligent Engineering and Systems, Vol.12, No.3, 2019 DOI: 10.22266/ijies2019.0630.23
Received: May 20, 2017 234

v pv 1 i

f
1  sT Switching Signal
Low-pass filter
  dp
dv
* 
Detector block

i pv 1 v I

f
L
1  sT
Low-pass filter 

Figure. 6 Existing control structure of transformer coupled DC-DC converter

the droop coefficient ‘m’ with difference between dc

3. Control strategy bus voltage 𝑉𝑑𝑐 and its reference value 𝑉𝑑𝑐_𝑟𝑒𝑓 to
In this section, the control structure of PV System produce (𝑑𝑝/𝑑𝑣)𝑟𝑒𝑓 for 𝑑𝑝/𝑑𝑣 regulator as shown
is explained in detail. First subsection explains the below.
control of DC-DC converter switches SW1 and SW2, Where (dp/dv)* represents the nominal value of
which are complementary in nature. Second sub- dp/dv i.e., slope of PV curve and this value is set to
section discusses control of inverter in such a way ZERO to unify Maximum Power Point Tracking
that its output voltage and frequency is regulated as (MPPT) and voltage regulation control with same
per load requirements. It is controlled by PWM controller.
control based on anti-islanding control, it will work The linearized model of the a given PV based ith DG
on Islanded or grid-connected mode. is given as
Here, classical P&O based MPPT in Fig.5 (a) is
𝑑𝑝𝑟𝑒𝑓 𝑑𝑝∗
replaced by coordinated 𝑑𝑝/𝑑𝑣 based MPPT block- = − 𝑚(𝑣𝑑𝑐 − 𝑣𝑑𝑐_𝑟𝑒𝑓 ) (9)
𝑑𝑣 𝑑𝑣
3 in Fig.5 (b) which coordinates MPPT and DC bus
voltage regulation in a seamless way of switching. 𝑑𝑝 𝑃 𝑑𝑝
𝑃̂𝑝𝑣_𝑖 = 𝑓̂ (𝑑𝑣 ) = 𝑦(𝑣max _𝑖 ) 𝑑𝑣 + 𝑃max _𝑖 (10)
Two DGs are considered to emphasize load sharing 𝑖 𝑝𝑣_𝑜𝑐𝑖 𝑖
between DGs.
Fig.5 (b) shows a modified control strategy, in
which working of 𝑑𝑝/𝑑𝑣 control technique is
explained below. Pmax

𝑑𝑝 𝑑𝑖 𝑖 −𝑖
𝑑𝑣
= 𝑣𝑝𝑣 𝑑𝑣 + 𝑖𝑝𝑣 = 𝑣𝑝𝑣 (𝑣𝑝𝑣 −𝑣𝑓 ) (8)
𝑝𝑣 𝑓

The above expression is implemented in

PV Output (W)

Simulink as shown in Fig6, where ‘𝑖𝑓 ’ and 𝑣𝑓 are

filtered values of ‘ 𝑖𝑝𝑣 ’ and 𝑣𝑝𝑣 ,there should be Slope m
limitation for 𝑑𝑝/𝑑𝑣 to avoid overstepping, a
detector provided, which block the 𝑑𝑝/𝑑𝑣
calculation, when overstepped and adds’ 𝛥𝐼𝐿 ’ to
‘𝐼𝐿 ’.The value ‘m’ decides the power sharing between
DGs, but as PV-curve is nonlinear, it is to be
linearized by line joining Maximum power point
(MPP) and zero power point (ZPP) as show in Fig. 7
[23]. Here hat in Eq. (10) and Solid line of Fig. 7 y(v pv _ oci ) -dp/dv value  y (v ) 
pv

represents linearizing PV power characteristics 𝑉 − ̂pv − 𝑑𝑝/𝑑𝑣 curves for PV

Figure. 7 𝑃𝑝𝑣 − 𝑑𝑝/𝑑𝑣 and P
𝑑𝑝/𝑑𝑣 controller as shown in Fig. 5(b) multiplies panel

International Journal of Intelligent Engineering and Systems, Vol.12, No.3, 2019 DOI: 10.22266/ijies2019.0630.23
Received: May 20, 2017 235

Subscript ‘i’ represents ith DG source. Thus load Here, 𝑝𝑝𝑣 is the power generated by solar PV panel
sharing droop parameter m can be identified as from which 𝑑𝑝/𝑑𝑣 i.e. 𝑦𝑖 (𝑣𝑝𝑣𝑖 ) is calculated, given
𝑃𝑚𝑎𝑥 _1 𝑑𝑝
in Eq. (17), solve for 𝑑𝑝/𝑑𝑣 and substituting 𝑣𝑝𝑣 =
+𝑃𝑚𝑎𝑥 _1
𝑃̂𝑝𝑣_1 𝑦(𝑣𝑝𝑣𝑜𝑐1 )𝑑𝑣1 𝑣𝑜𝑐,𝑛 , to get Eq. (18). ‘𝐺’ is the irradiance at a given
= (11)
𝑃̂𝑝𝑣_𝑖 𝑃𝑚𝑎𝑥 _𝑖 𝑑𝑝
+𝑃𝑚𝑎𝑥 _𝑖 point of time and ‘𝐺𝑛 ’ is the nominal irradiation,
𝑦(𝑣𝑝𝑣 )𝑑𝑣𝑖
𝑜𝑐𝑖 generally 1000w/𝑚2 .
But Eq. (12) ‘𝑑𝑝/𝑑𝑣’ can be represented as −𝑚𝑖 𝛥𝑉, Case Study:
Where, 𝛥𝑉 = 𝑉𝑑𝑐 − 𝑉𝑑𝑐𝑟𝑒𝑓 , 𝑃𝑚𝑎𝑥 _𝑖 represents the
maximum power generated by ith DG Case-I: If 𝑃𝑚𝑝𝑝𝑡 > 𝑃𝑙𝑜𝑎𝑑 i.e., PV panel delivers
Substituting the above value in Eq. (12) and rewritten power which is in operation at MPPT is more than
the same as follows, load required, the surplus power 𝑃𝑚𝑝𝑝𝑡 − 𝑃𝑙𝑜𝑎𝑑 is
𝑃𝑚𝑎𝑥 _1
absorbed by the battery, then total power delivered by
𝑚 .𝛥𝑉+𝑃𝑚𝑎𝑥 _1
𝑃̂𝑝𝑣_1 𝑦(𝑣𝑝𝑣𝑜𝑐1 ) 1 PV panel 𝑃𝑚𝑝𝑝𝑡 = 𝑃𝑙𝑜𝑎𝑑 + 𝑃𝑏𝑎𝑡𝑡 , where 𝑃𝑏𝑎𝑡𝑡 is the
𝑃̂𝑝𝑣_𝑖
= 𝑃𝑚𝑎𝑥 _𝑖 (12)
𝑚𝑖 .𝛥𝑉+𝑃𝑚𝑎𝑥 _𝑖 power absorbed by the battery, from the above
𝑦(𝑣𝑝𝑣
𝑜𝑐𝑖
)
decentralized control structure each DG, 𝐼𝐿 ∗ of
converters become positive representing battery is in
The main purpose of droop controller is proportional charging mode. In this case, the DC bus voltage is
load sharing as shown below, maintained at reference value and PV operating at
̂ pv_1
MPP which is similar to grid-connected operation,
P Pmax _1
̂ pv_i = (13) here, 𝑑𝑝/𝑑𝑣 regulator in Fig. 6(b) gives ZERO
P Pmax _i
reference to operate PV source work in MPP.
Case-II: If 𝑃𝑚𝑝𝑝𝑡 > 𝑃𝑙𝑜𝑎𝑑 , and battery reached
Substituting Eq. (16) in Eq. (15) will give the
expression of 𝑚𝑖 which is slope of PV curve (or) its maximum charging limit 𝑃𝑏𝑚𝑎𝑥 ,as battery should
droop control parameter responsible for proportional not be overcharged, the PV panel is forced to operate
load sharing between different DGs, at reduced power i.e., in non-MPPT mode in order to
maintain voltage regulation at DC-bus. But here
𝑦𝑖 (𝑣𝑝𝑣_𝑜𝑐𝑖 ) battery no more able to store excess power, this
𝑚𝑖 = 𝑣 (14) excess power charges the DC link capacitors thus
𝑑𝑐_𝑚𝑎𝑥 −𝑣𝑑𝑐_𝑟𝑒𝑓
voltage across DC bus increases more than
Where, Vdc_max, Vdc_ref are the upper voltage limit and 𝑣𝑑𝑐_𝑟𝑒𝑓 ,and subsequently the droop controller shown
reference voltages of the DC microgrid respectively. in block-3 of Fig. 8 generated negative 𝑑𝑝/𝑑𝑣
The expression for ipv is given in Eq. (1), multiplying command to render PV source to operate in such a
it with Vpv to get the expression for PV power given way, so as to balance generation and load. If once
in Eq. (11) load increases in voltage regulation mode, the droop
controller makes PV operating point move back
𝑝𝑝𝑣 = 𝑣𝑝𝑣 𝑖𝑝𝑣 (15) towards MPP.
Case-III: Here a decentralized control strategy is
𝑑𝑝 applied i.e., two DGs of similar rating as shown in
𝑦𝑖 (𝑣𝑝𝑣𝑖 ) = 𝑑𝑣
𝑝𝑣 Fig.4 are connected parallel with its individual
𝑑𝑖𝑝𝑣 control strategies across the load whose rating is 1KW.
= 𝑖𝑝𝑣 + 𝑣𝑝𝑣 𝑑𝑣 (16)
𝑝𝑣 A sudden change in load causes change in load shared
by each generator is almost similar. It is the
Yi (Vpv_oci), is the value of dp/dv when PV voltage responsibility of controllers associated with converter
reaches its nominal open circuit value of ith DG, to select the mode of operation and logical decision
i.e. (𝑣𝑝𝑣 = 𝑣𝑜𝑐,𝑛 ) , where suffix ‘n’ represents making Block-2. Block-1 is used to process the data
nominal value. regarding the availability of PV power and SOC of
battery and necessary signals are generated for safe
𝐺 𝑉𝑜𝑐,𝑛𝑖
𝑦𝑖 (𝑣𝑝𝑣𝑜𝑐𝑖 ) = 𝑁𝑝𝑖 𝐼𝑠𝑐,𝑛𝑖 (𝐺 − 1 − ) (17) running of DC-DC converter.
𝑛 𝛼𝑣𝑡 Case-IV: In this case, DG shifts its operation from
Islanded to Grid-connected mode and a sudden load
change causes a corresponding change in grid current,

International Journal of Intelligent Engineering and Systems, Vol.12, No.3, 2019 DOI: 10.22266/ijies2019.0630.23
Received: May 20, 2017 236

but inverter side current remains unaltered, means, Table 1. PV panel Parameters at STC
any changes in load can be supported by grid and PV PV Panel Parameters Value
always operate in MPPT mode in grid-connected case. 𝑁 28
𝑁𝑠 , 𝑁𝑝 1,2
4. Simulated performance
𝑉𝑀𝑃𝑃 34.5V
The Proposed Isolated PV system is simulated on
MATLAB/SIMULINK platform in order to 𝐼𝑀𝑃𝑃 7.4A
determine the feasibility of the control scheme. PV 𝑃𝑀𝑃𝑃 525W
module parameters are shown in Table 1, and PV 𝑉𝑂𝐶 40V
panel has two strings with each string taking one PV
𝐼𝑆𝐶 8.7A
module connected panels. At 1000 W/m and 25 C
(STC), Maximum Power developed by single module
is 250W.so, the maximum power of the PV power Table 2. Standalone System Parameters
delivered by PV panel at STC is 525W.i.e. 2X34.5 Parameter Value
X7.4=525W. power rating 500VA
The simulated performance obtained under Transformer turns ratio, n 6
various test cases is presented in this section. The Capacitors Cpv=2000 µF,
DC-DC converter inverter, battery and load Cb = 1000 µF
Cf = 8µF,
parameters are specified in Table 2. The cases (I-IV)
C1 = 470 µF,
of operation discussed in control strategy are C2 = C3 = 1000 µF
simulated and results are shown below, for the Inductors Lb=1 mH,L =3 mH,
shutdown case which happens for standalone systems Lf = 2.5 mH
only. This case is not discussed here, but this can be Switching frequency, Fs =15 kHz
done by withdrawing pulses to SW1 and SW2 in Fig. MPPT v-dp/dv power point
4(b). tracking
Battery rating 36V,7Ah
Simulated Results for the above case study:
i)𝑃𝑙𝑜𝑎𝑑 < 𝑃𝑚𝑝𝑝𝑡 : A load of 480W is maintained on Table 3. Comparison between existing and proposed
AC side, PV power with MPPT is obtained as 500W, control strategies.
and remaining 20W is delivered to battery. Here Parameter Modified Existing
Method Method
battery current𝑖𝑏𝑚𝑖𝑛 < 𝑖𝑏 < 𝑖𝑏𝑚𝑎𝑥 , battery limits are
between 1A and 3A for maximum and minimum Reference Voltage 400V 410V
currents to maintain battery state of charge (SOC)
Change in Voltage 398-403V 405-430V
between 20% and 90%. due to load change
ii) 𝑃𝑙𝑜𝑎𝑑 < 𝑃𝑚𝑝𝑝𝑡 : A load of 480W is maintained on
AC side, PV power with MPPT is obtained as 500W, Overshoot Absent present
but at 0.4 sec load is changed from 480W to 300W,
and battery current is 𝑖𝑏 > 𝑖𝑏𝑚𝑎𝑥 .In this work 𝑖𝑏𝑚𝑎𝑥
set to 0.7A thus battery cannot draw further current, control strategy and new control strategy is shown in
PV is forced to operate at non-MPPT mode and Fig. 4(a) and 4(b) respectively are shown in the Fig.
feeding reduced power to the load and this will
14, which depicts the contribution of this work.
continue up to 0.7sec and again the load is raised to Fig.13 shows the output voltage and current at load
480W. Thus Changes in PV outputs, battery and DC
in Islanded mode operation, and power delivered to
bus voltage is shown in the following figures. the load is shown in Fig. 12 and Input PV power
Here change in DC link voltage in Fig. 10 is
change in Islanded mode for load change is shown in
minimum, which indicates that controller has worked Fig.16. Load change for this time is created between
to operate at voltage regulation mode when there is a 0.4sec to 0.6sec. In Existing method, reference
change in load and battery is unable to absorb excess voltage is given to DC reference value and is
power delivered by PV when operated under MPPT. maintained at 410V, when load is changed at 0.4sec,
Here, waveforms are started with delay at zero as the
its voltage is raised to 430V i.e., between 0.4 to 0.6sec
mean values are considered for plots. Voltage whereas proposed control strategy controls DC link
regulation improvement comparison of existing voltage accurately near to its reference value.

International Journal of Intelligent Engineering and Systems, Vol.12, No.3, 2019 DOI: 10.22266/ijies2019.0630.23
Received: May 20, 2017 237

Detailed comparison analysis of existing and 1000W with a total load current of 1000/230=4.34A
proposed control strategies are specified in Table 3. (considering R-Load), rating of individual distributed
iii) Load sharing between two DGs are shown in generator shown in Fig. 3 is 500W, assuming droop
Fig.18 by considering different droop coefficients coefficient of DG i.e., m1=50, and for DG2 is
and its details are shown in Table 4. Considering two m2=150, thus the load is shared as per droop values.
DGs with decentralized control sharing a load of

Figure. 8 PV voltage current and power with load maintained constant in Islanded operation

Figure. 9 Battery voltage current and power with load maintained constant in Islanded mode

Figure. 10 DC bus voltage with load maintained constant in islanded operation

International Journal of Intelligent Engineering and Systems, Vol.12, No.3, 2019 DOI: 10.22266/ijies2019.0630.23
Received: May 20, 2017 238

Figure. 11 PV voltage and current waveforms for load change

Figure. 12 Load power output waveform

(a)

International Journal of Intelligent Engineering and Systems, Vol.12, No.3, 2019 DOI: 10.22266/ijies2019.0630.23
Received: May 20, 2017 239

(b)
Figure.13 AC side voltage and current waveforms

Figure. 14 DC link voltage when load is suddenly changed from 480 to 300 change in PV power for change in load with
existing and new MPPT control strategy

Figure. 15 Battery current when load is suddenly changed from 480 at 0.4sec to 300 at 0.7sec in islanded operation.

International Journal of Intelligent Engineering and Systems, Vol.12, No.3, 2019 DOI: 10.22266/ijies2019.0630.23
Received: May 20, 2017 240

Figure. 16 PV power when load is suddenly changed from 480W to 300W in islanded operation

Figure. 16 Change in AC power delivered to load due to sudden load change from 0.4sec to 0.7sec

Figure. 17 Change in AC load Current due to sudden load change from 0.4sec to 0.7sec

Fig. 18 (a) depicts the load sharing between DGs with change in load between 0.6 and 0.8sec.But in
of different rating with different droop coefficients both the waveforms there is a sudden raise in current
and Fig. 18 (b), for load sharing between DGs of at 0.3sec as DC bus voltage crossed its reference
same rating and with same coefficients m1=m2=150 value which activates -negative dp/dv command
respectively. We can notice the difference of droop signal to regulate DC bus voltage at 400V, thus
contributions of the distributed generators clearly increased voltage causes increased load sharing.

International Journal of Intelligent Engineering and Systems, Vol.12, No.3, 2019 DOI: 10.22266/ijies2019.0630.23
Received: May 20, 2017 241

(a)

(b)
Figure. 18 Load sharing: (a) load current sharing between two DGs of similar rating (500W each) but with different
droop coefficients and (b) load sharing between two DGs of similar rating and same droop coefficients

Figure. 19 Inverter and grid Voltage, Load (Iload), Grid (Igrid), Inverter (Iinv) current waveforms

International Journal of Intelligent Engineering and Systems, Vol.12, No.3, 2019 DOI: 10.22266/ijies2019.0630.23
Received: May 20, 2017 242

Table 4. Load sharing between Distributed generating Acknowledgments

sources.
Parameter Modified Existing We thank Dr.M Syam sundar, Assistant professor in
Method Method English, PVP Siddhartha Institute of Technology, for
Load Sharing ILDG1=1.716A comments that greatly improved the manuscript.
when for DG1 ILDG2=0.9A Not Considered
is m1=50 and for IL=ILDG1+ILD
DG2 is G2=2.616
References
m2=150,when (60% of4.34) [1] S.Adhikari and F. Li, “Coordinated V-f and P-Q
load changes Control of Solar Photovoltaic Generators With
from 1000W to MPPT and Battery Storage in Microgrids”,
600W IEEE Trans. Smart Grid, Vol. 5, No. 3, pp.
1270–1281, 2014.
Load Sharing ILDG1=1.31A
when both DG1 ILDG2=1.31A Not Considered [2] M. Miyatake, M. Veerachary, F. Toriumi, N.
& DG2 is IL=ILDG1+ILD Fujii, and H. Ko, “Maximum power point
m1=m2=150,wh G2=2.616 tracking of multiple photovoltaic arrays: A PSO
en load changes (60% of4.34) approach”, IEEE Trans. Aerosp. Electron. Syst.,
from 1000W to Vol. 47, No. 1, pp. 367–380, 2011.
600W [3] J. Zeng, W. Qiao, and L. Qu, “An isolated
multiport DC-DC converter for simultaneous
power management of multiple renewable
Thus the control structure shown in Fig. 5(b) energy sources”, In: Proc. of 2012 IEEE Energy
responsible not only for Maximum power point Convers. Congr. Expo. ECCE 2012, pp. 3741–
tracking (MPPT), Battery State of charge (SOC) 3748, 2012.
control of battery and DC voltage regulation. [4] T.-F. Wu, C.-H. Chang, Z.-R. Liu, and T.-H. Yu,
iv) Fig. 19, shows Inverter/Grid voltage “Single-stage converters for photovoltaic
waveforms, from 0 to 0.5 sec DG is operated in powered lighting systems with MPPT and
Islanded mode, at 0.5 sec, is synchronized to grid by charging features”, In: Proc. of IEEE Appl.
means of synchronizing switch. Load is changed is Power Electron. Conf. Expo. - APEC, Vol. 2, pp.
between 1 to 1.2 sec such that load maximum current 1149–1155, 1998.
value decreased from 3A to 2A peak and [5] D.D. C. Lu and V. G. Agelidis, “Photovoltaic-
corresponding change in load current and grid battery-powered DC bus system for common
currents are shown. When load current decreases, the portable electronic devices”, IEEE Trans. Power
DG feeds its excess current i.e., 𝑖𝑔𝑟𝑖𝑑 = 𝑖𝑑𝑔 − Electron., Vol. 24, No. 3, pp. 849–855, 2009.
𝑖𝑙𝑜𝑎𝑑 − 𝑖𝑏𝑎𝑡𝑡 to grid. [6] D. Debnath and K. Chatterjee, “Two-Stage Solar
Photovoltaic-Based Stand-Alone Scheme
Having Battery as Energy Storage Element for
5. Conclusion
Rural Deployment”, IEEE Transactions on
Solar and Battery based Transformer coupled Industrial Electronics, Vol. 62, No. 7, pp. 4148-
dual input DC-DC converter is considered and 4157, 2015.
improvised its DC side control strategy in such a way [7] D. Debnath and K. Chatterjee, “Transformer
that, MPPT, Battery SOC control, Voltage coupled multi-input two stage standalone solar
Regulation and Load sharing in one control strategy. photovoltaic scheme for rural areas”, In: Proc. of
Thus, avoided the necessity of mode switching and IECON 2013 - 39th Annual Conference of the
improvised the DC bus voltage profile at DC bus IEEE Industrial Electronics Society, pp. 7028-
compared to existing method, when system is 7033, 2013.
subjected to load changes in Islanded condition. This [8] S. H. Hosseini, F. Nejabatkhah, S. Danyali, and
proposed control strategy not only does the voltage S. a Kh, “Grid-Connected Three-Input PV / FC /
regulation, but also does load sharing based on droop Battery Power System with Active Power Filter
control technique and also works with both Islanded Capability”, In: Proc. of IEEE Conf., pp. 1–7,
and grid-connected modes of operation. Future scope 2011.
of this work includes extending this droop control [9] B. Papari, C. S. Edrington, I. Bhattacharya, and
technique for AC-DC microgrid with necessary G. Radman, “Effective Energy Management of
power control strategies. Hybrid AC-DC Microgrids with Storage

International Journal of Intelligent Engineering and Systems, Vol.12, No.3, 2019 DOI: 10.22266/ijies2019.0630.23
Received: May 20, 2017 243

Devices”, IEEE Trans. Smart Grid, Vol. 3053, Advanced Angle Droop Control", IEEE
No. c, pp. 1–10, 2017. Transactions on Smart Grid, Vol. 9, No. 6, pp.
[10] X. Lu, J. M. Guerrero, K. Sun, and J. C. Vasquez, 6461-6469, 2017.
“An improved droop control method for dc [20] Y. Han, H. Li, P. Shen, E. A. A. Coelho, and J.
microgrids based on low bandwidth M. Guerrero, "Review of Active and Reactive
communication with dc bus voltage restoration Power Sharing Strategies in Hierarchical
and enhanced current sharing accuracy”, IEEE Controlled Microgrids", IEEE Transactions on
Trans. Power Electron., Vol. 29, No. 4, pp. Power Electronics, Vol. 32, No. 3, pp. 2427-
1800–1812, 2014. 2451, 2017.
[11] Shivam and R. Dahiya,” Intelligent Distributed [21] B. Mangu and B. G. Fernandes, "Multi-input
Control Techniques for Effective Current transformer coupled DC-DC converter for PV-
Sharing and Voltage Regulation in DC wind based stand-alone single-phase power
Distributed Systems”, Arabian Journal for generating system", In: Proc. of 2014 IEEE
Science and Engineering, Vol. 42, No. 2, pp. Energy Conversion Congress and Exposition, pp.
5071-6081, 2017. 5288-5295, 2014.
[12] J. Rajagopalan, K. Xing, Y. Guo, F. C. Lee, and [22] B. Mangu, S. Akshatha, D. Suryanarayana, and
B. Manners, "Modeling and dynamic analysis of B. G. Fernandes, "Grid-Connected PV-Wind-
paralleled DC/DC converters with master-slave Battery-Based Multi-Input Transformer-
current control", In: Proc. of Applied Power Coupled Bidirectional DC-DC Converter for
Electronics Conference. APEC '96, Vol. 2, pp. Household Applications", IEEE Journal of
678-684, 1996. Emerging and Selected Topics in Power
[13] V. Nasirian, S. Moayedi, A. Davoudi, and F. L. Electronics, Vol. 4, No. 3, pp. 1086-1095, 2016.
Lewis, "Distributed Cooperative Control of DC [23] H. Cai, J. Xiang, W. Wei, and M. Z. Q. Chen,
Microgrids", IEEE Transactions on Power "V–dp/dv Droop Control for PV Sources in DC
Electronics, Vol. 30, No. 4, pp. 2288-2303, 2015. Microgrids", IEEE Transactions on Power
[14] V. Nasirian, S. Moayedi, A. Davoudi, and F. L. Electronics, Vol. 33, No. 9, pp. 7708-7720, 2018.
Lewis, "Distributed Cooperative Control of DC [24] M. G. Molina and P. E. Mercado, "Modeling and
Microgrids", IEEE Transactions on Power control of grid-connected photovoltaic energy
Electronics, Vol. 30, No. 4, pp. 2288-2303, 2015. conversion system used as a dispersed
[15] R. K. Sharma and S. Mishra, "Dynamic Power generator", In: Proc. of 2008 IEEE/PES
Management and Control of a PV PEM Fuel- Transmission and Distribution Conference and
Cell-Based Standalone ac/dc Microgrid Using Exposition: Latin America, pp. 1-8, 2008.
Hybrid Energy Storage", IEEE Transactions on [25] J. A. Gow and C. D. Manning, "Development of
Industry Applications, Vol. 54, No. 1, pp. 526- a photovoltaic array model for use in power-
538, 2018. electronics simulation studies", In: IEEE
[16] Y. Gu, W. Li, and X. He, "Frequency- Proceedings - Electric Power Applications, Vol.
Coordinating Virtual Impedance for 146, No. 2, pp. 193-200, 1999.
Autonomous Power Management of DC
Microgrid", IEEE Transactions on Power
Electronics, Vol. 30, No. 4, pp. 2328-2337, 2015.
[17] M. Nasir, H. A. Khan, A. Hussain, L. Mateen,
and N. A. Zaffar, "Solar PV-Based Scalable DC
Microgrid for Rural Electrification in
Developing Regions", IEEE Transactions on
Sustainable Energy, Vol. 9, No. 1, pp. 390-399,
2018.
[18] W. Yuan, J. Yang, Y. Sun, H. Han, X. Hou, and
M. Su, "A novel operation mode for PV-storage
independent microgrids with MPPT based droop
control", In: Proc. of 2017 IEEE 3rd
International Future Energy Electronics
Conference and ECCE Asia, pp. 936-941, 2017.
[19] B. John, A. Ghosh, and F. Zare, "Load Sharing
in Medium Voltage Islanded Microgrids with

International Journal of Intelligent Engineering and Systems, Vol.12, No.3, 2019 DOI: 10.22266/ijies2019.0630.23