IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS, VOL. 14, NO.

7, JULY 2018 2941

Autonomous Power Control and Management

Between Standalone DC Microgrids
Pannala Sanjeev , Student Member, IEEE, Narayana Prasad Padhy, Senior Member, IEEE,
and Pramod Agarwal, Member, IEEE

Abstract—Renewable integrated dc Microgrids (DCMGs) generation and consumption. Thus, effective coordination and
are gaining popularity by feeding remote locations in quali- control plays key role in ADCMGs to meet optimal energy man-
tative and quantitative manner. Reliability of autonomous agement and efficient utilization of resources and storage units
dc microgrids depend on battery capacity and size due
to stochastic behavior of renewables. Over charging and [10].
discharging scenarios compel the microgrid into insecure Control schemes based on centralized controller provide the
zone. Increasing the storage capacity is not an economi- optimal operation among various units by acquiring the informa-
cal solution because of additional maintenance and capital tion from them and manage the data centrally [11]. But system
cost. Thus, interconnecting neighbor microgrids increases
reliability is degraded due to high dependence on central con-
virtual storing and discharging capacity when excess power
and deficit scenario arises respectively in any of the DCMG. troller and communication link. Droop control [12] is a basic
Control strategy plays vital role in regulating the power decentralized control method, which works based on local in-
within and between microgrids. Power control and man- formation but lacks with optimum utilization of resources of
agement technique is developed based on bus signaling microgrid. To overcome above drawbacks, a distributed con-
method to govern sources, storages, and loads to achieve trol strategy based on dc bus signaling method (DCBSM) was
effective coordination and energy management between mi-
crogrids. Proposed scheme is simple and reliable since bus introduced in [13]. But it fails to consider the over charging
voltages are utilized in shifting the modes without having and discharging of battery. In [14], state of charge (SoC) of
dedicated communication lines. Proposed scheme is vali- battery is included in primary level control based on DCBSM.
dated through real time simulation of two autonomous dc Secondary level control is designed for adjusting bus voltage
grids in real time digital simulator (RTDS) and its results are
as per the reference voltage. As the battery alone regulates the
verified by hardware experimentation.
bus voltage, reliability of system degrades. In [15], decoupling
Index Terms—Autonomous dc Microgrid, bus signaling the operating regions in primary level control is proposed using
method, power control and management scheme, renew- dc bus voltage levels. Also, coordination among various storage
able sources, real time simulation.
devices is achieved in secondary level through communication.
However, excess generation is inefficiently managed by using
I. INTRODUCTION dump loads. Multilevel energy management strategy is proposed
in [16], where hybrid storage devices are utilized to suppress
SOLATED dc microgrids are gaining lot of interest from re-
I searchers in recent past [1]–[4] over counter parts available
due to their merits in relieving from complex control, synchro-
both low- and high-frequency components during power varia-
tions. During over charging or discharging conditions, the hybrid
storage devices are poorly managed if the communication fails
nization issues, harmonics, and reactive power [5]. In devel-
among control levels.
oping/undeveloped countries, domestic consumers, data center,
Xu and Chen [17] and Chen et al. [18] proposed the differ-
and telecommunication systems in remote locations are served
ent control modes based on bus voltage deviation for regulating
by local dc grids instead of conventional grid since utility con-
the dc microgrid under variable generation and storage. These
nection is not feasible or uneconomical [6]–[8]. Unlike grid
papers utilize bus voltage for indicating status of dc microgrids.
connected dc microgrid [9], there is no utility available in au-
Both the papers consider the utility grid and assigns slack role
tonomous dc microgrids (ADCMGs) to balance power between
to different sources (i.e., utility grid side converter or storage
converter) in each mode based on conditions of dc microgrid
Manuscript received June 21, 2017; revised October 24, 2017; ac- and utility grid. Distinct control loops are employed under each
cepted November 7, 2017. Date of publication November 14, 2017; date
of current version July 2, 2018. This work was supported by the Depart- mode for optimizing the system performance, which requires
ment of Science and Technology, Government of India towards India-UK frequent switching between control loops that causes switching
“HEAPD” research project no. DST/RCUK/SEGES/2012/09(G). Paper transients and also increases burden on control processor. Be-
no. TII-17-1324. (Corresponding author: Pannala Sanjeev.)
The authors are with the Department of Electrical Engineering, In- sides this, excess power beyond the battery charging rate and
dian Institute of Technology, Roorkee 247667, India (e-mail: sanjeev. grid side converter rating is not explored in [17]. Although it
[email protected]; [email protected]; [email protected]). is considered by [18], but the deviation of bus voltage is more
Color versions of one or more of the figures in this paper are available
online at http://ieeexplore.ieee.org. than 10% of nominal value in islanded mode, which affects
Digital Object Identifier 10.1109/TII.2017.2773507 the sensitive loads connected. Power line signaling method is

1551-3203 © 2017 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.

Authorized licensed use limited to: KIIT University. Downloaded on July 27,2022 at 07:51:30 UTC from IEEE Xplore. Restrictions apply.
2942 IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS, VOL. 14, NO. 7, JULY 2018

storages, which is not cost effective. In [24], interconnection of

ADCMGs is explored by proposing centralized control strategy
to ensure optimal operation of each grid using conventional non-
isolated BDC. System reliability heavily depends on centralized
controller and communication link. This yields to single point
of failure. Moreover, isolation between ADCMGs is not pro-
Fig. 1. Typical interconnection of two ADCMGs. vided. Even though isolation inherently present in [25] due to
flyback converter, but power sharing heavily depends on the gap
between generation and demand that is similar to previous case,
proposed in [19] to overcome problem of limited number of which entails the communication from various units. In [26], a
operating modes based on fixed voltage deviation in DCBSM. new dc–dc converter based on LCL filter is proposed by elimi-
It dispatches the status of batteries and other sources in terms of nating transformers in dual active bridge converter (DABC) for
distinct frequency signals superimposed on bus voltage. Various high power and high voltage transmission systems. Although it
sources can shift their operating modes by extracting the infor- shows significant improvements in reducing the weight, losses
mation from different frequency signals. However, this method and reactive power, but suits for high power and high dc voltage
is not suitable for increased number of storage systems (SSs) applications where the operating frequency is limited. Besides,
and distributed energy resources (DERs) since available carrier it also lacks with galvanic isolation. Extended phase shift control
signal frequencies are limited and also varies based on different is proposed for DABC in [27] to interface the low and medium
converter parameters, which makes proposed scheme cumber- voltage dc grids. Main focus lies on bidirectional power control
some to implement for more number of DERs and SSs. Besides, and fails to consider the conditions of the individual dc grids.
it consumes additional current from battery for dispatching the Design and operation of new BDC enclosing CLLC resonant
various signals. tank for zero voltage and zero current switching is explored in
Above all, beyond the charging and discharging capacity of [28] for dissimilar voltage based dc buses but does not com-
the storage devices in ADCMGs, when power surplus or deficit ment on the status of sources connected to individual dc buses,
arises in the ADCMG, an additional storage system is required which influences the power flow between the dc grids. Lee
[20]. Increasing the SS will elevate system cost due to extra et al. [29] proposed a strategy based on bus states for trans-
maintenance and initial setup cost. Besides, it introduces com- ferring the power between dc microgrids. As the droop control
plexity in control, which indirectly reduces reliability due to is employed for governing the distributed generation in single
dependency on fast processing and communication technology microgrid, scheme fails to extract the maximum power from
between different storage units. In order to avoid the additional renewable sources and demands central communication for op-
storage requirement, interconnection of dc grids is developed timal operation. Two-level tertiary control scheme is proposed
[21] similar to conventional ac systems (one area to other area), in [30] for optimal power sharing among dc microgrids within
which enhances system reliability and make the efficient usage the cluster by regulating reference voltages in dc microgrids.
of resources. Typical structure for interconnecting ADCMGs However, this scheme entails sparse communication and may
is shown in Fig. 1. Each microgrid encloses DERs, SSs, and yield to nonoptimal power sharing between the dc microgrids ei-
various loads. ther due to upper level communication failure or due to tracking
Interconnection of dc microgrids are described in [21]–[30]. errors during dynamic power variation of sources/loads within
Adhikari et al. [21] proposed decentralized power flow con- microgrids [31].
trol between dc microgrid clusters using tie line connection In this paper, a power control and management strategy
based on bus voltage adjustment/ regulation without incorporat- (PCMS) is developed based on DCBSM for individual AD-
ing extra bidirectional dc–dc converter (BDC). Small deviation CMGs and as well as between ADCMGs without any dedi-
in bus voltages may yield unwanted power flow between the cated communication infrastructure. Control and management
dc microgrids and also increases transmission losses. Various of sources, storages and loads in PCMS are executed by using
configurations for interconnecting the dc microgrid clusters are bus voltage information. Similarly both the bus voltages are used
studied in [22] and a control strategy is proposed based on to trigger the DABC between the ADCMGs for power exchange.
hierarchical control for regulating the power through interlink- In addition, PCMS embraces extreme situations like over and
ing BDC between dc microgrids. As it embraces three control under loading conditions of individual ADCMGs, which are
layers (i.e., primary, secondary, and tertiary), which demands handled efficiently by running load shedding and PV derating al-
dedicated and fast communication channels between them for gorithms, respectively. PCMS is simple and communication free
optimal power sharing among dc microgrids that makes sys- power control strategy within and between microgrids so that
tem more costly and also increases burden on control processor. effective utilization and management of resources is achieved
Three approaches (circuit switching, packet switching, and vir- along with enhanced reliability of ADCMGs.
tual packet switching) are proposed in [23] for power sharing Paper is organized as follows: architecture of system is ex-
between dc microgrid clusters by taking analogy from internet plained in Section II. Proposed PCMS for ADCMGs is discussed
architecture. They are analyzed in terms of their abilities and in Section III. Control loops for various converters and isolated
merits over other with four fundamental rules of internet. But bidirectional DC-DC converter (IBDC) are elaborated in Sec-
it increases number of interlinking dc–dc converters and buffer tion IV. Simulation results are discussed in Section V, and are

Authorized licensed use limited to: KIIT University. Downloaded on July 27,2022 at 07:51:30 UTC from IEEE Xplore. Restrictions apply.
SANJEEV et al.: AUTONOMOUS POWER CONTROL AND MANAGEMENT BETWEEN STANDALONE DC MICROGRIDS 2943

Fig. 2. System architecture for interconnection of two ADCMGs.

validated by experimental results in Section VI. Section VII

concludes the paper.

II. SYSTEM STRUCTURE

System contemplated in this paper is as shown in Fig. 2,
which consists of two ADCMGs spatially apart from each other Fig. 3. Proposed PCMS between two ADCMGs.
with considerable line resistance between them. Each ADCMG
consists of one photovoltaic (PV) source and battery as equiv-
alent to group of sources from renewable sources and storage 1) Maximum capacity of PV source is higher than the rated
devices family respectively in order to simplify the analysis battery charging power and load power to serve the loads
for the proposed PCMS between the ADCMGs. As most of dc most of the time through PV power.
loads are of constant power loads (CPLs), which are integrated 2) Nominal bus voltage deviations are within the tolerance
through dc–dc converter. Hence, CPLs can able to maintain band of loads.
fixed power irrespective of variations in dc bus when its voltage
oscillations lie within the sustainable range [15]. PV source is III. POWER CONTROL AND MANAGEMENT STRATEGY
interfaced to dc bus through boost converter and bidirectional
Source and storage units of ADCMGs are operated based
buck-boost dc–dc converter is utilized for connecting the bat-
on bus voltage levels in the grid by making bus voltage as
tery storage. Interconnection of two ADCMGs is realized by
information carrier between the units for proper coordination
considering DABC as interfacing unit, which provides galvanic
and management. Loads are managed depending on the SoC of
isolation and high power feeding capability in both directions
battery and power condition of ADCMG, which is expressed
along with large conversion ratios through high-frequency trans-
in terms of bus voltage deviation. Instantaneous SoC can be
former [27]. Two full H-bridge converters are connected to either
estimated by using coulomb counting method [32]
side of the transformer to produce the high-frequency ac from
dc. Besides, this topology recognized as DABC and widely 1
employed in transferring the power from low to high voltage SoCi = SoC0 − ∫ δibat dt (1)
CN
dc [27].
As DABC is placed nearby low voltage dc grid (i.e., AD- Where, SoCi , SoC0 , CN , and ibat denotes instantaneous SoC,
CMG2) to avoid losses, its bus voltage is directly available to initial SoC, nominal capacity, and input current of battery, re-
DABC. Bus voltage information of other ADCMG (i.e., AD- spectively. δ is loss coefficient, which typically varies in the
CMG1) is taken from the line connecting the two ADCMGs. range of 0.98–1 [33]. SoC of battery can be regulated by limit-
Suppose the line is open, then voltage measured at capacitor ing the charging and discharging currents. Voltage levels envi-
C1 (V  a ) is considered as bus voltage of ADCMG1 for BDC sioned throughout the analysis are in compliance with the paper
triggering as there is no line drop. If the line is closed between [34]. Operation of each ADCMG is bifurcated into five zones in
two ADCMGs then the current measured in the line can be used which each zone will be active depending upon particular bus
to find out the voltage drop (iexp1 × R) in the line. This voltage voltage threshold as shown in Fig. 3. Va and Vb are grid voltages
drop of the line can be further added to the voltage measured of ADCMG1 and ADCMG2. In order to leverage maximum
at high voltage side of the DABC (i.e., Va ) to estimate the ex- power from PV source, it is operated at maximum power point
act voltage of ADCMG1. Hence, the voltage of ADCMG1 can (MPP) using perturb and observe method in all zones excluding
be easily obtained without any communication link. Therefore, zone-5. PCMS is explained with respect to ADCMG1. Region1
ADCMGs can exchange the power among them based on bus and region2 indicates power flow directions from ADCMG1 to
voltages information available at DABC without dedicated com- ADCMG2 and ADCMG2 to ADCMG1, respectively. Bisection
munication systems. The following assumptions are made in the line used to differentiate the regions of power transfer from one
system. ADCMG to other ADCMG with distinct background colors.

Authorized licensed use limited to: KIIT University. Downloaded on July 27,2022 at 07:51:30 UTC from IEEE Xplore. Restrictions apply.
2944 IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS, VOL. 14, NO. 7, JULY 2018

Total power generation and load in each ADCMG is indicated transmit the power from ADCMG2 to ADCMG1 and regulates
by pg and pL , respectively in Fig. 3. the voltage of ADCMG2 at VbH 2 . Transmitted power can push
1) Zone-1 (Balanced power mode): As power generated by the ADCMG1 to operate at different operating points in dissim-
PV source (pPV 1 ) is more or less equal to demand (pL 1 ) ilar zones depending upon power imported (pim p1 ) in region2
in the ADCMG1, which keeps battery in idle state. Small as shown in Fig. 3, which is explained with following equations:
variations of load and source will not trigger the storage ⎧
unit in this mode because predefined voltage limits are ⎪
⎪ VaL 1 : if pPV1 + pbat1 + pim p1 = pL 1
⎨ ∼ pL 1
able to sustain these fluctuations. As there is no fixed VaL 1 < Va < VaH 1 : if pPV1 + pim p1 =
Va = . (5)
source to regulate the bus voltage in this zone, which in ⎪
⎪ V aH 1 : if p PV1 + p im p1 > pL1 ,
⎩
turn allow its voltage to vary between the limits VaH 1 pbat1 = − (pPV1 + pim p1 − pL 1 )
and VaL 1 that are treated as boundaries of this mode
for ADCMG1. Similarly VbH 1 and VbL 1 are boundaries If the first condition in (5) is met by ADCMG1 then it will be
for ADCMG2. Neither of the ADCMGs share power to operated at point h in zone-2 shown in Fig. 3. Similarly, once
other ADCMG in this zone. Status of various units in second and third condition is met, then ADCMG1 can be oper-
ADCMG1 are given by ated between h and i in zone-1 and at i in zone-3, respectively.
In contrast to above points, whenever pim p1 starts diminishing
pPV 1 ∼
= pL 1 ; pbat1 = 0; VaL 1 < Va < VaH 1 (2) then operating point will move from i to g. At any condition if
where pbat1 is instantaneous battery power in ADCMG1. pim p1 becomes zero then operating points shift from i, h to j, k,
2) Zone-2 (Battery discharging mode): As the PV power is respectively.
not sufficient in fulfilling the demand that yields to con- Case2: This case arises when ADCMG2 is not feeding either
tinuous deviation in bus voltage (Va ). Once Va fall below enough power or zero power to ADCMG1. Hence, load shedding
threshold value (VaL 1 ) then storage steps into discharging is done as per the priority order in ADCMG1. Load sheds either
mode from idle state in order to cover the gap between based on decrement of SoC1 below cutoff limit or falling of
supply and demand. Battery clamps the bus voltage at voltage Va below VaL 2 . In order to avoid the load shedding,
same threshold (VaL 1 ) by keeping it in bus regulating nonrenewable source is to be activated during above scenarios.
mode. ADCMG1 is ready to absorb the excess power Here the operating points g and l indicate the load shedding of
from ADCMG2 if available (i.e., operating point h shown ADCMG1 corresponding to below equation order
in Fig. 3), but reverse power transfer is not possible. Ex-
pPV1 + pbat1 + pim p1 < pL 1
treme conditions of battery are taken care in control loop. (6)
pPV1 + pbat1 < pL 1 ; pim p1 = 0.
Equation indicating the conditions of ADCMG1 is given
by
5) Zone-5 (Excess power mode): This zone is further split
pPV1 + pbat1 = pL 1 ; Va = VaL 1 . (3) into two sub cases, where one case deals with export-
3) Zone-3 (Battery charging mode): If the PV source is pro- ing surplus power from ADCMG1 and another case is
ducing excess power than required, then this mode comes without exporting power.
into picture where voltage Va rises continuously due to Case1: Either rise in generation or fall in load beyond the
surplus power and halts at threshold limit (VaH 1 ) by shift- charging rate of battery will enforce the ADCMG into this op-
ing the battery into charging mode. Battery is allowed to erating zone. If the surplus power is not utilized within the
charge until its cutoff limit is met. ADCMG1 cannot feed grid then bus voltage starts rising. Status of other ADCMG is
the power to ADCMG2, but absorb the power from it observed from its bus voltage in order to dispatch the excess
when Vb is at VbH 2 and battery1 in ADCMG1 is not fully power available. If the ADCMG2 is in a state other than ex-
charged or maximum charging rate is not met (shown as cess power case (i.e., Vb ≤ VbH 1 ), then power is transferred
operating point i in Fig. 3). from ADCMG1 to ADCMG2 displayed as region1 in Fig. 3.
If the power is exported from ADCMG1, then bus voltage of
pPV1 > pL 1 ; pbat1 = − (pPV1 − pL 1 ) ; Va = VaH 1 . ADCMG1 (Va ) is regulated at VaH 2 . Therefore, generated re-
(4) newable power is efficiently used. ADCMG2 can be operated at
4) Zone-4 (Power deficit mode): It is an extension of zone- (b, c, d, e), (a, f) similar to case1 and case2 in zone-4 depending
2 and comes into active state when load rises beyond upon its imported power (pim p2 ).
discharging rate of battery. In this mode, battery runs at Case2: This is an extension of previous case where ADCMG2
maximum discharging current limit. There are two sub is not able to absorb the excess power readily available in AD-
cases in this zone, in which first case deals with power CMG1 that increase bus voltage (Va ). On the detection of rise in
import from ADCMG2 whereas in the second case, there bus voltage (Va ) above VaH 2 , PV source will shift its operating
is no power import from ADCMG2. mode from MPP to bus voltage regulation mode and clamps its
Case1: In this case, status of ADCMG2 is checked against bus voltage at VaH 3 . Status of ADCMG1 in zone5 is given by
surplus power mode by observing its bus voltage Vb . Simi- 
larly status of ADCMG1 is known by its bus voltage (Va ). If VaH 2 : if pexp1 > 0
Va = (7)
Vb = VbH 2 and Va ≤ VaH 1 , then DABC will be triggered to VaH 3 : if pexp1 = 0&VPV 1ref = VaH 3

Authorized licensed use limited to: KIIT University. Downloaded on July 27,2022 at 07:51:30 UTC from IEEE Xplore. Restrictions apply.
SANJEEV et al.: AUTONOMOUS POWER CONTROL AND MANAGEMENT BETWEEN STANDALONE DC MICROGRIDS 2945

Fig. 4. Control loop of PV source in ADCMG1.

Fig. 5. Control loop of battery in ADCMG1.
where pexp1 and VPV 1ref represents instantaneous power ex-
ported from ADCMG1 to ADCMG2 and PV voltage reference,
respectively.

IV. CONTROL LOOPS

Converter control plays vital role in executing the proposed
PCMS for switching the source converters between different
modes based on predefined thresholds for triggering. Until and
unless the defined threshold voltage is reached by the dc grid,
particular mode will not be active because the outer loop is sat-
urated and does not provide any reference to inner loop. Con- Fig. 6. Control of BDC between two ADCMGs.
trol of battery1 and PV source1 inside the ADCMG1 is elabo-
rated and the similar control structure utilized in ADCMG2 for which inner loop is common for both the modes. No mode will
battery2 and PV source2. Power flow control of IBDC between be active until the bus voltage crosses its predefined threshold
two dc grids is explained in detail. value due to saturation of controller. Outer voltage loop mainly
tracks bus voltage reference and produces the current reference
A. PV Control Loop as output, and that is fed to inner loop for tracking the reference
PV source always remain at MPP except in zone5 irrespective effectively. In discharging mode, top outer loop gives the posi-
of change in load and power import/export so that maximum tive reference current (Ibtrd1 ) by regulating bus voltage (Va ) at
renewable power is extracted and utilized efficiently. PV control VaL 1 as load dominates, which is fed to discharging rate limiter
loop of ADCMG1 is shown in Fig. 4 and the same is followed and then checked against its cutoff limit (Vbat1L ) to ensure the
for PV source in ADCMG2. PV source can be operated in two optimal utilization and extended battery’s life. Cutoff limit of
modes, one as MPP mode and other as bus voltage regulation SoC1 = 20% is expressed in terminal voltage (Vbat1L ) [33]. In
mode. First one consist of two loops in which outer loop is this mode, bottom outer loop is inactive. During charging mode,
mainly for tracking MPP voltage (VM PP ) through perturb and bus voltage is regulated at VaH 1 through bottom outer loop and
observe (P&O) method and provides voltage reference as input produces negative current as reference to inner loop since bus
to inner loop. Inner loop works at faster speed for tracking the voltage always try to rise above the threshold limit (VaH 1 ). Neg-
given reference through PI controller and produce the duty cycle ative reference current (Ibtrc1 ) is checked against charging cur-
(δPV 1 ) as its output. Second mode encloses only one loop and rent limit and full charge (i.e., SoC1 = 90%) equivalent voltage
comes into picture when case2 of zone-5 occurs. Bus voltage Vbat1H to avoid high charging rate and over charging condi-
is regulated at its reference (VaH 3 ) through PI controller by tion respectively. Inner control loop track the reference current
adjusting duty cycle. Both the modes use the PWM comparator (Ibtref 1 ) by changing the duty cycle (δbtr1 ). In both modes, duty
to generate the required pulses for switches inside the converter cycle is fed to the PWM comparator for synthesizing the pulses
by feeding duty cycle to it. Switching between two modes is for battery converter.
done selectively by observing the bus voltage. If the Va increase
above the upper threshold VaH 2 , then PV enter into the bus C. Bidirectional Power Control Between ADCMGs
regulation mode, otherwise PV keeps running in MPP mode.
DABC is employed as BDC for transferring the power be-
Although battery1 is fully charged, but PV is not pushed into
tween ADCMGs. Mainly it works on conventional phase shift
regulation mode since excess power transfer may takes place
method [35]. Power transfer takes place either from ADCMG1
from ADCMG1 to ADCMG2 for the effective utilization of
to ADCMG2 or from ADCMG2 to ADCMG1. BDC control
renewable energy, which increases reliability of the system.
is shown in Fig. 6. It consists of two loops in which one is
outer voltage loop and other is inner current loop. BDC comes
B. Battery Control Loops to active state when one of two ADCMGs are possessing the
Battery is operated either in charging or discharging mode surplus power and other grid is able to absorb. Power transfer
based on its cut in thresholds. Battery control loop of AD- from ADCMG2 to ADCMG1 is treated as positive convention
CMG1 is shown in Fig. 5. Each mode employs two loops in and reverse action as negative. If excess energy is available in

Authorized licensed use limited to: KIIT University. Downloaded on July 27,2022 at 07:51:30 UTC from IEEE Xplore. Restrictions apply.
2946 IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS, VOL. 14, NO. 7, JULY 2018

TABLE I
SIMULATION PARAMETERS

Components Parameters ADCMG1 ADCMG2

PV Capacity Maximum power 4.5 kW 750 W

@1000 W/m2
Battery Capacity 200 Ah 100 Ah
Nominal voltage 96 V 24 V
Nominal grid voltage Rated voltage 380 V 48 V
Voltage Thresholds Vx H 3 410 54 V
Vx H 2 400 52 V
Vx H 1 390 50 V
Fig. 7. Load shedding control. Vx L 1 370 46 V
Vx L 2 360 44 V
DC load Fixed Load 1 kW 200 W
ADCMG1 and other grid status is examined by its voltage. If Variable Load 2 kW 300 W
other grid is ready (i.e., Vb ≤ VbH 1 ), then it starts exporting the Line parameters Resistance(R) 0.15 Ω
Inductance (L) 0.24 mH
power by regulating the bus voltage Va at VaH 2 through outer
voltage controller and provide the negative reference current ∗ x = a for ADCMG1 and x = b for ADCMG2.
(Ib dc2 ), which is passed through limiter for avoiding high rush
current than absorbing capacity. Then, reference current (Ib dcr )
is tracked by BDC current (ib dc ) using inner current controller where VaT is threshold voltage values for all microgrids con-
and generates the required negative phase shift. Similarly while nected to ADCMG1 side, Vai∗ , αai , and Iai are output volt-
transferring the power from ADCMG2 to ADCMG1, a positive age reference set point, droop coefficient, and output current of
reference current (Ib dc1 ) is generated by regulating the voltage ith ADCMG, respectively (i = 1, 2, . . . m). Similar convention
Vb at VbH 2 through outer loop and is followed by inner loop for is followed to (9) (where j = 1, 2, . . . n). Droop coefficients
synthesizing the positive phase shift. While obtaining the status αai , αb j are given by
of ADCMGs by scanning its bus voltage, actual bus voltages VaTm ax − VaTm in
are passed through the low pass filter (LPF) to avoid the false αai = (10)
Paim ax − Paim in
switching.
VbTm ax − VbTm in
αb j = (11)
D. Load Control Pb jm ax − Pb jm in

Constant power load is developed in this paper by feeding Where VaT m ax , VaT m in -Maximum and minimum voltage
resistive load through the buck converter. Buck converter con- variation allowed around particular threshold value in common
trol is similar to conventional control scheme [36]. Load shed- bus of ADCMG1. Paim ax and Paim in are maximum power in-
ding is performed based on dc bus voltage and battery status jected and absorbed to/from common bus of ADCMG1 by ith
as shown in Fig. 7. If the battery voltage falls below cutoff ADCMG of same voltage type, respectively. Same convention
value VbatL 2 (equivalent to SoC = 30%) then load shedding is is followed to (11).
initiated to preserve the battery to feed the essential loads ef- This scheme may not be effective and efficient for more than
fectively. Similarly if the bus voltage decreases below the lower two dissimilar voltages based microgrids since complexity in-
threshold (VaL 2 ) then load shedding is activated because power creases with increased number of different voltages-based mi-
deficit exceeds discharging rate of battery. Simultaneously if crogrids. Therefore, this scheme is economical and reliable for
both occurs during worst scenario, then also it triggers the load multiple similar ADCMGs (i.e., Va and Vb voltage-based grids),
shedding. Load can be recovered when bus voltage and battery which are very nearer. Complete analysis of multiple ADCMGs
voltage become higher than their limits (i.e., VaL 1 , VbatL 1 {i.e., will be carried out in future research work.
SoC = 50%}). Actual bus voltage and battery terminal voltage
are fed through the LPF to avoid the unwanted switching of V. SIMULATION RESULTS
load shedding process during sudden increment in load power Simulation of the system shown in Fig. 2 is carried out on real-
or decrement in source power. time digital simulator (RTDS) platform to validate the developed
The proposed PCMS can be extended to multiple similar PCMS. System is implemented in RSCAD/RTDS environment.
ADCMGs operating at same nominal voltages (i.e., Va and Vb ) Specifications of system components such as PV, battery, and
by utilizing the droop control strategy [12], [13]. It is assumed load in both the ADCMGs are detailed in Table I. Two practical
that same nominal voltage based multiple microgrids can be dc grid voltage ratings are chosen to prove the applicability of
placed on identical grid voltage side and its cutoff voltages PCMS. Load is classified as fixed and variable loads based on its
are same as proposed. Power sharing among similar microgrids existence throughout certain period. They represent the essential
connected to ADCMG1 and ADCMG2 are given by and nonessential loads indirectly. Nonessential loads/variable
Vai∗ = VaT − αai Iai (8) loads are shed according to the requirement and other type is
fed continuously. Proposed PCMS is explored under various
Vb∗j = VbT − αb j Ib j (9) operating scenarios of ADCMGs including extreme conditions

Authorized licensed use limited to: KIIT University. Downloaded on July 27,2022 at 07:51:30 UTC from IEEE Xplore. Restrictions apply.
SANJEEV et al.: AUTONOMOUS POWER CONTROL AND MANAGEMENT BETWEEN STANDALONE DC MICROGRIDS 2947

single window. Cutoff limits for each zone are selected with
difference of 10 V [34] and bus voltage deviation for all zones
lie within the ±5% of bus nominal voltage. It is observed from
the Fig. 8 that generated power is equal to load demand from 0 to
t1 that indicates the operating zone-1, during which bus voltage
is allowed to vary between boundaries. At time t1 , increase in
PV power (pPV 1 ) causes the excess power compared to local
demand (pL 1 ), which drive the battery into charging mode from
idle state in previous case and starts regulating the bus voltage
(Va ) at VaH 1 , which exhibit the zone-3 of ADCMG1. Sudden
decrease in variable load appears at t2 that impel the battery
charging current to exceed its saturation limit, which causes rise
of Va since surplus power is not consumed by other ADCMG or
local load. Thus, PV move into bus regulation mode from MPP
mode on the detection of rise in Va above VaH 2 that is shown
as zone-5 in Fig. 8. PV controls the bus voltage Va at VaH 3
until t3 where Va falls below VaH 2 due to fall of PV power with
decay in irradiation. Again zone-3 appears from t3 to t4 . At time
t4 , rapid increase in load causes the zone-2 to appear between
instants t4 and t5 where the demand dominates the generation
and battery compensate the difference by shifting from charging
to discharging mode instantaneously at t4 as shown in Fig. 8(e).
Further increment in load and fall of irradiation occurs at t5
that compel battery to discharge more and make bus voltage to
fall below its lower cutoff limit (VaL 2 ). Thus, load shedding is
performed at t6 to bring back ADCMG1 under normal operating
zones. Time interval between t5 and t6 reveals the operating
zone-4. Once load is shed then it retains the operating zone-2.
Similarly, ADCMG2 is operated with voltage thresholds men-
tioned in Table I. ADCMG2 follows same sequence of operating
zones as that of ADCMG1 as shown in Fig. 9 except Zone-4
in which load sheds when battery terminal voltage falls below
its cutoff limit (i.e., SoC1 = 30%). Although it follows same
sequence of zones as ADCMG1 for operation, but power vari-
ations inside ADCMG2 is different than ADCMG1 to include
wide cases. Aforementioned analysis reveals the effectiveness of
the control scheme even under extreme conditions of individual
ADCMGs.
Fig. 8. Operating zones of ADCMG1. (a) Bus voltage. (b) Irradiation.
(c) PV output power. (d) Battery terminal voltage. (e) Battery output
power. (f) Load power. B. Inter DC Grid Power Flow
Simulation of inter dc grid power flow using developed PCMS
of battery and grid like over charging and discharging, overload is shown in Fig. 10. From 0 to t8 , ADCMG1 is operated in
and underload scenarios of ADCMG. Operation of individual zone-3 where the generation dominates the power demand and
ADCMGs is illustrated using PCMS under different zones and it battery is charging with surplus power as shown in Fig. 10(e).
is followed by bidirectional power transfer between ADCMGs Simultaneously ADCMG2 is in zone-2 with predominant load.
considering aforementioned conditions in zone-4 and zone-5. As Further increase in load occurs at t7 and clamp the bus voltage Vb
explained in the previous section, bus voltages information of at VbL 1 till t8 . Sudden rise in PV power (pPV 1 ) inside ADCMG1
ADCMGs are directly available to BDC for switching between at t8 causes the battery to charge at its saturation current limit due
different modes. Typical line length considered here is 1 km and to surplus power, which pushes the bus voltage (Va ) to increase
its parameters are listed in Table I. further. Once Va reaches VaH 2 , then BDC control checks the
status of other grid (i.e., ADCMG2). As Vb ≤ VbH 1 , then BDC
is allowed to deliver the power from ADCMG1 to ADCMG2
A. Individual DC Microgrids by controlling the BDC output current as shown in Fig. 10(c)–
Various operating regions of ADCMG1 is shown in Fig. 8. Fig. 10(d). Although power is transferred from ADCMG1 to
Battery terminal voltage is kept just above the lower cutoff limit ADCMG2, but ADCMG2 remains to be in zone2 since total
to compile all zones including extreme conditions within the available power within ADCMG2 including the power imported

Authorized licensed use limited to: KIIT University. Downloaded on July 27,2022 at 07:51:30 UTC from IEEE Xplore. Restrictions apply.
2948 IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS, VOL. 14, NO. 7, JULY 2018

Fig. 10. Operation of BDC using PCMS. (a)ADCMG1 voltage. (b) AD-
CMG2 voltage. (c) Power exported from ADCMG1. (d) Power exported
from ADCMG2. (e) Powers within ADCMG1. (f) Powers within ADCMG2.

Fig. 9. Operating zones of ADCMG2 (a) Bus voltage. (b) Irradiation. (c)
PV output power. (d) Battery terminal voltage. (e) Battery output power.
(f) Load power.
ADCMG2 to ADCMG1 (i.e., pexp2 ) due to rise of Vb above
threshold VbH 2 and at the same time Va is below VaH 1 .
In order to validate the effectiveness of PCMS, a compari-
son is made against configurations from [24], [25] as config-I
from ADCMG1 is less than the demand. Once load falls at time
and config-II, respectively in the Table II. It can be observed
t9 in ADCMG2, then it jumps into zone-3 from zone-2 as total
that proposed configuration finds advantages in terms of in-
available power including power imported exceeds the local
creased reliability, flexible control without communication and
demand as shown in Fig. 10(b) and Fig. 10(f). At instant t10 ,
complexity, effective utilization of resources, and storages by
sudden fall of pPV 1 and increment of load inside the ADCMG1
enclosing the extreme conditions of battery and ADCMGs.
occur simultaneously. This results in ceasing of power export
(pexp1 ) from ADCMG1 as pPV 1 is sufficient enough to feed
the load pL 1 . Meanwhile load pL 2 decreases, which leads the VI. EXPERIMENTATION
ADCMG2 to remain in zone-3 till the instant t11 . Variations In order to validate the simulation results, a prototype model
in bus voltage Va is seen between t10 and t11 due to minor is developed. Bus voltage of ADCMG2 remains same as that
fluctuations in pPV 1 while tracking MPP through P&O method. of simulation, whereas ADCMG1 bus voltage is considered as
pPV 1 and pL 2 decreases simultaneously at time t11 that produces 100 V for experimentation and corresponding voltages thresh-
surplus power in ADCMG2 beyond its battery charging rate. olds are VaH 2 = 110, VaH 1 = 105, VaL 1 = 95, VaL 2 = 90.
Therefore, BDC control is activated to export the power from Two programmable dc sources of rating 200 V, 40 A are config-

Authorized licensed use limited to: KIIT University. Downloaded on July 27,2022 at 07:51:30 UTC from IEEE Xplore. Restrictions apply.
SANJEEV et al.: AUTONOMOUS POWER CONTROL AND MANAGEMENT BETWEEN STANDALONE DC MICROGRIDS 2949

TABLE II Experimentation is started by operating ADCMG1 and AD-

COMPARISON OF PCMS BASED SYSTEM.
CMG2 in zone-3 and zone-2, respectively, where battery1
is charging with excess power available in ADCMG1 and
Parameter Config-I Config-II Proposed system battery2 is discharging to meet the demand in ADCMG2 as
[24] [25]
shown in Fig. 11. At time t1 , increase in PV1 current produces
Communication Required Required Not required excess power more than charging rate of battery1 in ADCMG1
Control complexity High Medium Low
Isolation Not Provided Provided Provided
that leads to rise in Va . Mean time ADCMG2 is in battery dis-
Voltage stress High High Low charging mode (i.e., zone-2), which can be clearly observable
of BDC from its bus voltage (i.e., Vb at VbL 1 ) in Fig. 11. Once Va rises
Reliability Low∗ Low∗ High
Power sharing Flexible Not flexible Flexible
above VaH 2 , then power export takes place from ADCMG1 to
between DCMGs ADCMG2 between t1 and t3 . ADCMG2 is shifted to zone-1 at t1
since generation along with imported power equals the demand.
∗– with respect to communication based control scheme As load2 falls at t2 , ADCMG2 is shifted into zone-3 by pushing
the battery into charging mode due to surplus power. Increment
in load1 at t3 , pushes the ADCMG1 to operate in zone-1 and it
stops exporting the power to ADCMG2. At the same time AD-
CMG2 is shifted into zone-2 since there is no power exchange.
Increase in PV2 current at t4 forces the ADCMG2 to operate
in zone-3 where excess power is used to charge battery2. At t5 ,
decay in PV1 current causes power deficit in ADCMG1, which
is compensated by discharging the battery1 (i.e., zone-2). As
PV2 current increases at t6 , this results in surplus power, which
exceeds the charging rate of battery2 in ADCMG2, that compel
it to export the excess power to ADCMG1.

VII. CONCLUSION
A PCMS is developed based on bus signaling technique for
inter dc grid power flow in case of ADCMGs to increase the sys-
tem reliability and efficient utilization of resources. Two practi-
cal dc grid voltages (380 V, 48 V) are considered for evaluating
the performance of developed scheme in simulation. PCMS is
explored under normal and extreme scenarios including the over
and under loading conditions of ADCMGs, further more with
over charging and discharging of battery. It can be observed
from above analysis that proposed PCMS is stable, efficient and
effective in realizing communication independent control even
under dynamic power variations during the power exchange.
This statement is also justified with experimental results ob-
tained through prototype model developed in the laboratory
with reduced voltage of ADCMG1. Proposed system provides
isolation and also enhances system reliability. Application po-
tential of system suits low and medium voltage customers such
Fig. 11. Experimental results of inter dc grid power flow: Bus voltages as domestic consumers, data centers, and telecommunication
(V a , V b ); current from ADCMG1 to ADCMG2 (ie x p 1 ); current from AD- systems in isolated locations where utility connection is not
CMG2 to ADCMG1 (ie x p 2 ); PV source currents (iP V 1 , iP V 2 ); Battery
source currents (ib a t1 , ib a t2 ); Load currents (iL 1 , iL 2 ). present or feasible.

ured to mimic the behavior of PV sources. Resistances of range REFERENCES

0-75 Ω, 5 A are considered for loading the dc grids. Two series [1] T. Dragicevi, X. Lu, J. C. Vasquez, and J. M. Guerrero, “DC Microgrids –
connected lead acid batteries with each of rating 12 V, 60 Ah Part II: A review of power architectures, applications and standardization
issues,” IEEE Trans. Power Electron., vol. 31, no. 5, pp. 3528–3549,
are used in ADCMG2 and similar rated batteries of four num- May 2016.
bers are connected in series within ADCMG1. Bus voltages are [2] Q. Yang, L. Jiang, H. Zhao, and H. Zeng, “Autonomous voltage regulation
sensed by AD202JN isolation amplifier and all source currents and current sharing in islanded Multi-inverter DC microgrid,” IEEE Trans.
Smart Grid, to be published.
are sensed through Hall effect Telcon sensors. Proposed scheme [3] J. Torreglosa, P. Garcia, L. Fernandez, and F. Jurado, “Predictive control for
is implemented in RTDS to communicate with prototype model the energy management of a fuel Cell-Battery-Supercapacitor tramway,”
via digital and analog input/outputs. IEEE Trans. Ind. Informat., vol. 10, no. 1, pp. 276–285, Feb. 2013.

Authorized licensed use limited to: KIIT University. Downloaded on July 27,2022 at 07:51:30 UTC from IEEE Xplore. Restrictions apply.
2950 IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS, VOL. 14, NO. 7, JULY 2018

[4] L. Herrera, W. Zhang, and J. Wang, “Stability analysis and controller [30] S. Moayedi and A. Davoudi, “Distributed tertiary control of DC micro-
design of DC microgrids with constant power loads,” IEEE Trans. Smart gridclusters,” IEEE Trans. Power Electron., vol. 31, no. 2, pp. 1717–1733,
Grid, vol. 8, no. 2, pp. 881–888, Mar. 2017. Feb. 2015.
[5] D. E. Olivares et al., “Trends in microgrid control,” IEEE Trans. Smart [31] M. Yazdanian and A. Mehrizi-Sani, “Distributed control techniques in
Grid, vol. 5, no. 4, pp. 1905–1919, Jul. 2014. microgrids,” IEEE Trans. Smart Grid, vol. 5, no. 6, pp. 2901–2909,
[6] G. Karina, V. Vagelis, S. Alan, and B. Gabriel, “Optimizing energy sav- Nov. 2014.
ings from “Direct-DC” in US residential buildings,” Lawrence Berkeley [32] S. Piller, M. Perrin, and A. Jossen, “Methods for state-of-charge de-
National Laboratory, Berkeley, CA, USA: Tech. Rep., Oct. 2011. termination and their applications,” J. Power Sources, vol. 96, no. 1,
[7] Schneider Electric, Indogreen, Telecom Towers Case Study, Rueil- pp. 113–120, 2001.
Malmaison, France: Tech. Rep. 2014. [33] M. Coleman, C. K. Lee, C. Zhu, and W. G. Hurley, “State-of-charge
[8] EmergeAlliance, 380 Vdc Architectures for the Modern Data Center, San determination from EMF voltage estimation: Using impedance, terminal
Ramon, CA, USA: Tech. Rep. 2013. voltage, and current for lead-acid and lithium-ion batteries,” IEEE Trans.
[9] G. S. Seo, J. W. Shin, B. H. Cho, and K. C. Lee, “Digitally controlled cur- Ind. Electron., vol. 54, no. 5, pp. 2550–2557, Oct. 2007.
rent sensorless photovoltaic micro-converter for DC distribution,” IEEE [34] K. Sun, L. Zhang, Y. Xing, and J. M. Guerrero, “A distributed control
Trans. Ind. Informat., vol. 10, no. 1, pp. 117–126, Feb. 2014. strategy based on DC bus signaling for modular photovoltaic genera-
[10] T. L. Vandoorn, B. Meersman, L. Degroote, B. Renders, and L. Vande- tion systems with battery energy storage,” IEEE Trans. Power Electron.,
velde, “A control strategy for islanded microgrids with DC-Link voltage vol. 26, no. 10, pp. 3032–3045, Oct. 2011.
control,” IEEE Trans. Power Del. vol. 26, no. 2, pp. 703–713, Apr. 2011. [35] C. Mi, H. Bai, C. Wang, and S. Gargies, “Operation, design and control of
[11] X. Liu, P. Wang, and P. C. Loh, “A hybrid AC / DC microgrid and its dual H-bridge-based isolated bidirectional DC–DC converter,” IET Power
coordination control,” IEEE Trans. Smart Grid, vol. 2, no. 2, pp. 278–286, Electron., vol. 6, no. 7, pp. 2846–2852, 2011.
Jun. 2011. [36] N. Mohan and T. Undeland, Power Electronics: Converters, Applications,
[12] J. John, F. Mwasilu, J. Lee, and J. Jung, “AC-microgrids versus DC- and Design. Hoboken, NJ, USA: John, 2007.
microgrids with distributed energy resources: A review,” Renew. Sustain.
Energy Rev. J., vol. 24, pp. 387–405, 2013.
[13] J. Schönberger, R. Duke, and S. D. Round, “DC-bus signaling: A dis-
tributed control strategy for a hybrid renewable nanogrid,” IEEE Trans. Pannala Sanjeev (S’16) received B.Tech. and
Ind. Electron., vol. 53, no. 5, pp. 1453–1460, Oct. 2006. M.Tech. degrees in electrical engineering from
[14] D. Wu, F. Tang, T. Dragicevic, J. M. Guerrero, and J. C. Vasquez, “Coor- CVR College of Engineering, Hyderabad, In-
dinated control based on Bus-Signaling and virtual inertia for DC islanded dia, and National Institute of Technology Bhopal,
microgrids,” IEEE Trans. Smart Grid, vol. 6, no. 6, pp. 1–12, Nov. 2015. Bhopal, India, in 2011 and 2013, respectively.
[15] C. Jin, P. Wang, J. Xiao, Y. Tang, and F. H. Choo, “Implementation He is currently working toward the Ph.D. de-
of hierarchical control in DC microgrids,” IEEE Trans. Ind. Electron., gree at Indian Institute of Technology Roorkee,
vol. 61, no. 8, pp. 4032–4042, Aug. 2014. Roorkee, India, under DST sponsored HEAPD
[16] J. Xiao, P. Wang, and L. Setyawan, “Multilevel energy management system Project.
for hybridization of energy storages in DC microgrids,” IEEE Trans. Smart He was an Assistant Professor with Mukesh
Grid, vol. 7, no. 2, pp. 847–856, Mar. 2016. Patel School of Technology Management and
[17] L. Xu and D. Chen, “Control and operation of a DC microgrid with variable Engineering, NMIMS University, Mumbai, India, from 2013 to 2014. He
generation and energy storage,” IEEE Trans. Power Del., vol. 26, no. 4, is currently a Research Scholar under DST sponsored HEAPD Project.
pp. 2513–2522, Oct. 2011. His area of interest includes dc microgrids, Hybrid energy storage sys-
[18] D. Chen, L. Xu, and L. Yao, “DC voltage variation based autonomous tems, and Renewable energy source integration.
control of DC microgrids,” IEEE Trans. Power Del., vol. 28, no. 2, Mr. Sanjeev was the recipient of the best paper award at IEEE Na-
pp. 637–648, Apr. 2013. tional conference ICAER held in Punjab University, October 2013.
[19] T. Dragicevi, J. M. Guerrero, and J. C. Vasquez, “A distributed con-
trol strategy for coordination of an autonomous LVDC microgrid based
on power-line signaling,” IEEE Trans. Ind. Electron., vol. 61, no. 7,
Narayana Prasad Padhy (SM’09) received the
pp. 3313–3326, Jul. 2014.
Ph.D. degree in power systems engineering from
[20] T. Dragicevic, J. M. Guerrero, J. C. Vasquez, and D. Skrlec, “Supervi-
Anna University, Chennai, India, in 1997.
sory control of an adaptive-droop regulated DC microgrid with battery
He is currently a Professor and NEEPCO
management capability,” IEEE Trans. Power Electron., vol. 29, no. 2,
Chair Professor in the Department of Electri-
pp. 695–706, Feb. 2014.
cal Engineering, Indian Institute of Technology
[21] S. Adhikari, Q. Xu, Y. Tang, and P. Wang, “Decentralized control of DC
Roorkee, Roorkee, India. He is the National lead
microgrid clusters,” in Proc. Int. Future Energy Electron. Conf. ECCE
of the HEAPD project under India-UK Smart En-
Conf. Asia, 2017, pp. 567–572.
ergy Grid and Energy Storage research initia-
[22] J. Ma, M. Zhu, X. Cai, and Y. W. Li, “Configuration and operation of DC
tives. His research interests are power system
microgrid cluster linked through DC-DC converter,” in Proc. IEEE Conf.
analysis, optimization, demand side manage-
Ind. Electron. Appl., 2016, pp. 2565–2570.
ment, and smart grid.
[23] A. Werth, M. Tokoro, and K. Tanaka, “Bottom-up and recursive intercon-
Dr. Padhy has received the Boyscast Fellowship and the Humboldt
nection for multi-layer DC microgrids,” in Proc. Int. Conf. Environ. Elect.
Experienced research Fellowship during 2005 and 2009, respectively.
Eng., 2016, pp. 1–6.
He is also a Fellow of the Indian National Academy of Engineers and the
[24] M. Kumar, S. C. Srivastava, S. N. Singh, and M. Ramamoorty, “Devel-
Institution of Engineering and Technology.
opment of a control strategy for interconnection of islanded direct current
microgrids,” IET Renew. Power Gener., vol. 9, no. 3, pp. 284–296, 2015.
[25] S. Konar and A. Ghosh, “Interconnection of islanded DC microgrids,” in
Proc. IEEE PES Asia-Pac. Power Energy Eng. Conf., 2015, pp. 1–5. Pramod Agarwal (M’99) received the B.E.,
[26] D. Jovcic and L. Zhang, “LCL DC/DC converter for DC grids,” IEEE M.E., and Ph.D. degrees in electrical engi-
Trans. Power Del., vol. 28, no. 4, pp. 2071–2079, Oct. 2013. neering from the Indian Institute of Technology
[27] B. Zhao, Q. Yu, and W. Sun, “Extended-phase-shift control of isolated Roorkee, Roorkee, India, in 1983, 1985, and
bidirectional DC-DC converter for power distribution in microgrid,” IEEE 1995, respectively.
Trans. Power Electron., vol. 27, no. 11, pp. 4667–4680, Nov. 2012. In 1985, he joined the Indian Institute of
[28] W. Chen, P. Rong, and Z. Lu, “Snubberless bidirectional DC-DC converter Technology Roorkee as a Lecturer, where he is
with new CLLC resonant tank featuring minimized switching loss,” IEEE currently a Professor with the Department of lec-
Trans. Ind. Electron., vol. 57, no. 9, pp. 3075–3086, Sep. 2010. trical Engineering. During 1999–2000, he was a
[29] M. Lee, W. Choi, H. Kim, and B. H. Cho, “Operation schemes of in- Postdoctoral Fellow with the Ecole de Technolo-
terconnected DC microgrids through an isolated bi-directional DC-DC gie Supérieure, University of Quebec, Montreal,
converter,” in Proc. IEEE Appl. Power Electron. Conf. Expo., 2015, QC, Canada. His research interest includes power electronics and smart
pp. 2940–2945. grid.

Authorized licensed use limited to: KIIT University. Downloaded on July 27,2022 at 07:51:30 UTC from IEEE Xplore. Restrictions apply.