Scalable Distributed Control for DC Microgrids Using

Virtual Impedance and Intelligent Load Management

Satyaveer Singh Negi1 , Adel Rawea2,* , Prakash Dwivedi1 , Sourav Bose1 , and Yahya
Alward3
1 Department of Electrical Engineering, National Institute of Technology Uttarakhand, Srinagar Garhwal, India
2 Department of Mechatronics, Faculty of Engineering, Sana’a University, Sana’a, Yemen
3 Electrical Engineering Department, School of Engineering, Gautam Buddha University, Greater Noida, India
* [Link]@[Link]

ABSTRACT

This paper presents a hierarchical control strategy for a grid-connected DC microgrid with intelligent solar lighting systems as a
distributed energy resource and power flow management with the islanded mode operation capability. The proposed system
integrates distributed solar LED lights with individual maximum power point tracking controllers and battery storage (10 kW
aggregate solar capacity with 80 kWh total battery storage), a 50 kW solar photovoltaic plant, and an electronically controlled
dump load (10 kW). A novel virtual impedance-based droop control mechanism is implemented for distributed converters
to enhance current sharing accuracy and system stability. The hierarchical control architecture encompasses primary and
secondary control, with a 60 kW bidirectional interlinking converter facilitating power flow between the 220V DC microgrid and
220V AC grid. In islanded mode the excess power is managed with the help of an electronic load controller for dump loads. The
intelligent load management system enables uses intensity control of solar LED lights, reducing up to 20 percent load during
low power conditions. Battery DOD is maintained to 20 percent to ensure long battery life. Simulation results show better
voltage regulation, better current sharing among distributed converters resulting in stable microgrid voltage. There is seamless
transition between grid-connected and islanded operation with efficient load management under various operating scenarios.

1 Introduction
There is a need of providing electricity in rural areas and with the renewable energy resources such as solar-PV which can be
easily installed near load centers there is a demand for the development of microgrid technologies. DC microgrids offer some
advantages by directly adding DC loads without the need of extra converter and incorporate solar photovoltaic systems with
the grid while minimizing conversion losses. The increase in DC loads, particularly LED lighting systems, DC pumps and
electronic devices, given the inherent DC nature of photovoltaic (PV) systems and battery storage, makes DC microgrids an
obvious choice for modern microgrid systems1 .
There are some challenges in operation of microgrids such as current sharing accuracy, voltage regulation, and load
management during critical operating conditions. Recent comprehensive reviews9 establish virtual impedance-based droop
control as a mature solution for addressing these challenges, particularly in DC microgrids where line impedance variations
significantly affect power sharing performance. Advanced implementations with hybrid energy storage systems17 have
demonstrated successful integration achieving automatic low and high-frequency power sharing through combined virtual
resistance and virtual capacitor droop mechanisms. The investigation of droop control strategies for DC microgrids are done in
several literature1 , emphasizing on state-of-charge balance3 and voltage sharing4 . There are novel approaches proposed for
average voltage sharing control and frequency injection-based decentralized control5 . The hierarchical control framework2 has
been established as a general approach to coordinate droop-controlled microgrids.
Recent studies in hybrid AC-DC microgrids target bidirectional interlinking converter control strategies10, 11 . Lin et al.16
proposed a distributed power management strategy for multi-paralleled bidirectional interlinking converters, demonstrating
improved reliability and power sharing. Recent advances include unified decentralized frameworks using dual active bridge
converters with galvanic isolation21 and robust sliding mode control designs22 that maintain power stability under load
variations. Enhanced normalization methods24 and generalized bidirectional droop control architectures25 further improve power
flow management during mode transitions. However, practical implementation challenges remain regarding computational
complexity and real-time performance under varying load conditions.
Energy management in DC microgrids with distributed solar generation has been extensively studied. Qi et al.14 pre-
sented distributed hybrid energy storage control based on improved MPPT algorithms, while Gugulothu et al.15 proposed
 Figure 1. System Architecture of DC Microgrid with Bidirectional Interlinking Converter

comprehensive energy management strategies for standalone DC microgrids with multiple energy sources. The integration
of electronic load controllers for excess power management has been systematically reviewed23 , revealing methodologies
to enhance generator efficiency and extend lifespan by reducing dump load utilization from 100% to 80% of rated capacity.
The coordination of dump loads13 with battery storage and renewable curtailment represents a critical design consideration
for islanded operation. These works highlight the importance of coordinated control between generation, storage, and load
management systems.
The control architecture proposed in this paper implements a two-level hierarchical structure: primary and secondary
controls, where each control layer is operating at different time scale providing short time and long time load management.
The system addresses the challenge of managing distributed solar LED lightening system with dc-dc converters having MPPT
function and connected to their own battery storage maintaining the stable DC bus voltage and sharing current using virtual
impedance for droop control.
This paper proposes a droop control strategy based on virtual impedance for a DC microgrid with intelligent solar lighting
system with the ability to adjust its power consumption autonomously to support microgrid. The Microgrid system has
distributed solar LED lights with individual MPPT controllers and 24V battery storage, a centralized 50 kW PV plant, a 60 kW
bidirectional interlinking converter for AC-DC coupling and an electronic load controller with 10 kW dump load capacity.
The overall objective of the paper is to include:

1. Implementation of virtual impedance based droop control for smart lightning system such that they share current with
DC Microgrid without communication

2. Implementation of a bidirectional interlinking converter control strategy for seamless transition between grid-connected
and islanded modes

3. Implementation of an electronic load controller for managing excess generation during islanded operation.

4. Implementation of battery management system for distributed batteries in lightening system.

2 System Architecture and Configuration

2.1 Overall System Description
The proposed grid connected DC microgrid architecture is as shown in Fig.1. The system is operating at a DC bus voltage of
220V.
DC Microgrid System Components:

• A 50 kW solar PV plant with MPPT function.

• A distributed solar LED street lightening system with an aggregate power of 10 kW from solar panels, each with
individual boost converter incorporating MPPT function and virtual impedance based droop control

• Total battery energy storage is 80 kWh distributed across LED units where each unit has 2 kWh capacity operating at 24V

• 50 kW of local load.

2/15
 • An Electronic Load Controller (ELC) for 10 kW dump load consisting of DC water pumps and heaters.
Interlinking System:
• 60 kW bidirectional AC-DC interlinking converter (220V DC to 220V AC)
• Dual-mode operation: grid-following (grid-connected) and grid-forming (islanded)
• Power flow controller with synchronization capability
• Protection and mode transition control logic.

2.2 Solar LED Lighting System

Each solar LED light system unit consists of:
• Photovoltaic panel of 500W nominal capacity with bidirectional dc-dc converter and MPPT controller
• Battery storage (24V, 2 kWh) with Battery Management System (BMS)
• LED driver with PWM dimming capability having 200W nominal load
• controller logic with virtual droop control for connection to 220V DC microgrid
The solar LED lightning system operate in multiple modes:
1. Autonomous Mode: During daylight, PV panels charge batteries
2. Grid Injection Mode: Excess power is injected into the DC microgrid through virtual droop-controlled converters
3. Battery Priority Mode: During low SOC conditions, converters prioritize battery charging

2.3 Power Management Architecture

The system operates in two primary modes based on AC grid availability:
Grid-Connected Mode: When the main AC grid is available, the interlinking converter operates in grid-following mode
and measures the net power balance in the DC microgrid:

Pnet = Pgeneration − Pconsumption (1)

where Pgeneration = PPV,50kW + PLED,solar and Pconsumption = Ploads + PLED,consumption .

During grid-connected mode:
• Solar Panels of LED lightning system operate autonomously with local MPPT
• The 50 kW PV plant exports maximum available power to DC bus
• Batteries charge when SoC < 80% and sufficient power is available
• Excess power (Pnet > 0) is exported to AC grid through interlinking converter
• Power deficit (Pnet < 0) is imported from AC grid
• DC bus voltage is regulated at 220V ± 5%
• Electronic Load Controller remains inactive
Islanded Mode: When the AC grid is unavailable or during intentional islanding, the system manages power balance
through hierarchical control:
• Priority 1: Maintain critical loads at 220V DC bus
• Priority 2: Battery charging if SoC < 80% with controlled charge rates
• Priority 3: Electronic Load Controller (ELC) activation for dump load when SoC > 90% or bus voltage exceeds 225V
• Priority 4: Solar PV power curtailment if voltage exceeds 235V despite full ELC activation
• Priority 5: LED load reduction (up to 20%) during severe battery depletion

3/15
2.4 Electronic Load Controller Configuration
The ELC consists of controllable resistive loads for absorbing excess power:

• 3 × DC heater banks: 3 kW, 3 kW, 2 kW (total 8 kW)

• 2 × DC water pumps: 1 kW each (total 2 kW)

• PWM control for heaters enabling fine power control

• ON/OFF control for pumps

• Total dump load capacity: 10 kW

The ELC is active during Islanded Mode subjected to SOC of Battery, power generation and power demand.

2.5 Energy Storage Management

Battery management has following operating conditions.

• The battery DOD is 20%

• For long battery lifespan the state of Charge must not be greater than 80%

• The battery operating voltage range is 20V to 28V

The batttery SOC is estimated as below:

Z t
η
SoC(t) = SoC(0) + Ibat (τ)dτ (2)
Cbat 0
20% ≤ SoC(t) ≤ 90% (3)

where Cbat is battery capacity, Ibat is battery current and η is coulombic efficiency.

Hierarchical Control Strategy

This paper implements a two level hierarchal control strategy. The DC microgrid has intelligent solar lighting system and
bidirectional grid connected interlinking converter. The hierarchical control architecture implements primary and secondary
control. The interlinking converter allows bidirectional power flow between AC and DC buses.

2.6 Primary Control Level

The operating time of primary control is very fast ranging from a few milliseconds to seconds and it regulated the local voltage
and current with the help of droop control using virtual impedance.

2.6.1 Distributed Converter Primary Control with Virtual Impedance

The LED lightning system have bidirectional dc-dc converters. The control logic implements a virtual impedance based droop
control following established methodologies9, 17 for communication-free current sharing. The virtual impedance approach,
proven effective in hybrid energy storage systems17 , enables automatic power sharing across distributed converters through
adaptive impedance modulation.

∗
VDC = VDC − Rd · IDC − Rv · IDC (4)
Irated − IDC
Rv = Kv · (5)
Irated
where VDC∗ is the reference DC voltage, R is the physical droop resistance, R is the virtual impedance, K is the virtual
d v v
impedance coefficient and Irated is the rated current of each converter.
The virtual impedance changes based on loading conditions. It improves the current sharing among parallel-connected
converters. When the converter current is less than its rated current the virtual impedance is increased to raise its output
impedance. This allows better load sharing with other converters.
The control algorithm operates on battery SoC,voltage and current, DC bus voltage, converter output current and solar-PV
MPPT operating voltage and current. The Operating conditions are as follows

4/15
 1. battery charging for SoC less than 20%
2. grid feeding mode for SoC greater than 80%
3. prioritse charging when SoC is less than 50% and VDC > 225V
The virtual impedance changes during three loading regions:
1. during light loading condition the λ < 0.3 and Rv = Kv (1 − λ ) for increasing load sharing
2. during normal loading condition, 0.3 ≤ λ ≤ 0.7 and Rv = Kv (Irated − IDC )/Irated for proportional sharing
3. during heavy loading condition the λ > 0.7 and Rv = 0.1Kv to reduce the voltage drop, where λ = IDC /Irated is the
loading factor.
2.6.2 Centralized PV Plant Control
The 50 kW PV plant operates with centralized MPPT control. During normal operation, the plant extracts maximum power
using perturb-and-observe MPPT algorithm. During islanded mode with high SoC or voltage conditions, power curtailment is
implemented:


PMPP
 if VDC < 225V and SoCavg < 85%
PPV,re f = PMPP · (1 − kcurtail ) if VDC ≥ 225V or SoCavg ≥ 85% (6)

0 if VDC > 235V


where kcurtail is the curtailment factor (0 to 1) determined by voltage and SoC conditions as follows:
• kcurtail = 0: No curtailment under normal MPPT operation and VDC < 225V and SoCavg < 85%
• 0 < kcurtail < 1: Partial curtailment and proportional reduction kcurtail = min((VDC - 225)/(235 - 225), (SoCavg - 85)/(95
- 85)) and 225V < VDC < 235V OR 85% < SoCavg < 95%
• kcurtail = 1: Full curtailment when PV is disconnected and VDC > 235V (critical overvoltage)

2.7 Interlinking Converter Control

The bidirectional interlinking converter provide the grid connectivity to the DC Microgrid and allows power exchange between
the 220V DC microgrid and 220V AC grid. Following recent advances in unified decentralized frameworks21 and robust control
strategies22 , the converter implements dual-mode operation with enhanced normalization methods24 for power flow control.
The control strategy incorporates generalized droop characteristics25 , coupling DC voltage, AC frequency, and active power in
a three-dimensional control plane to enable autonomous bidirectional power flow without communication infrastructure.
The bidirectional converter has 60 kW, 220V DC/220V AC, 50 Hz operating capability and operate in dual modes based on
grid availability. In grid-connected mode, synchronization checks verify voltage within 10% nominal and frequency within 0.5
Hz of 50 Hz. Power management allocates excess power (Pnet > 0) to battery charging (if SoCavg < 80%) then grid export,
while deficit (Pnet < 0) triggers battery discharge (if SoCavg > 50%) and grid import as needed, maintaining unity power factor.
In islanded mode, the converter ceases grid interaction while the ELC handles excess power. Current control implements
grid-following operation while voltage-forming control enables islanded mode, with power limits enforcing rated capacity
constraints.
The converter operates in two distinct modes depending on grid availability.
2.7.1 Grid-Following Control (Grid-Connected Mode)
During grid-connected operation, the interlinking converter operates in grid-following mode. A phase-locked loop (PLL)
synchronizes the converter with the AC grid voltage. The active and reactive current references are generated based on the DC
power balance.
The reference currents in the synchronous reference frame are given by:
P∗
i∗d = (7)
1.5Vd
∗
iq = 0 (8)
ensuring unity power factor operation.
A cascaded control structure is implemented with an inner current control loop and an outer DC voltage or power control
loop. The current control loop bandwidth is designed to be approximately 2 kHz to ensure fast dynamic response.

5/15
2.7.2 Voltage-Forming Control (Islanded Mode)
When the AC grid is unavailable, the interlinking converter transitions to voltage-forming mode and operates as a controlled
voltage source. The AC voltage reference is generated internally and regulated using a droop-based voltage control strategy.
The voltage reference is defined as:
∗
VAC = Vnom − Zdroop IAC (9)
where Zdroop is the virtual impedance implemented on the AC side.
This control strategy enables stable AC voltage generation and allows seamless resynchronization when the grid is restored.

2.8 Electronic Load Controller Management

The ELC provides critical functionality during islanded operation by absorbing excess generation that cannot be stored or
consumed. Following established methodologies for electronic load controllers in standalone generating plants23 , the system
reduces generator operating stress by limiting dump load engagement to 80% of rated capacity, thereby enhancing efficiency
and extending component lifespan. The hierarchical load engagement strategy prevents DC bus overvoltage while allowing
continued operation of renewable generation with reduced thermal cycling.
The ELC manages 10 kW dump load capacity (three heater banks: 3+3+2 kW with PWM control; two pumps: 1+1 kW
with ON/OFF control) with activation thresholds at 230V (overvoltage) and 235V (critical). During islanded operation with
SoCavg > 90% or VDC > 230V , the controller calculates excess power exceeding battery absorption capacity and activates loads
sequentially (H1→H2→H3→P1→P2). Heater duty cycles are proportional to required dump power (clamped 0-1), while
pumps switch ON when remaining requirement exceeds rating. Emergency protection at VDC > 235V activates all dump loads
and triggers PV curtailment.

2.9 Battery Management System

Distributed battery management coordinates the charging and discharging of 50 individual 24V, 2 kWh battery units while
ensuring longevity and safety.
The battery management system coordinates 50 units (24V, 83.33 Ah, 2 kWh each) with SoC limits of 20-90% and
temperature compensation. Charging modes adapt to SoC: (i) bulk charge at 0.3C (25A) for SoC < 20%, (ii) constant current at
0.2C (16.7A) for 20-80% SoC, (iii) constant voltage at 28.8V for 80-90% SoC with current tapering, and (iv) float charge at
27.6V with 0.05C trickle above 90% SoC. Power limiting ensures charging does not exceed available system power, with safety
limits at 30V (overvoltage protection) and 1C maximum current (83.33A). SoC estimation employs coulomb counting with
0.95 coulombic efficiency.

2.10 Secondary Control Level

Secondary control operates at a slower time scale (seconds to minutes) and addresses the deviations caused by primary control
while optimizing system performance. The control structure is illustrated in Fig. 2.
2.10.1 Voltage Restoration
The droop control at primary level causes DC bus voltage deviation from the nominal 220V. Secondary control restores voltage
using a centralized PI controller:

Z
∗ ∗
δVsec = KIV (VDC −VDC,avg )dt + KPV (VDC −VDC,avg ) (10)

where δVsec is the secondary control voltage correction added to primary control references, and KIV , KPV are integral and
proportional gains (typical values: KPV = 0.1, KIV = 2).
The average DC bus voltage VDC,avg is computed from measurements at multiple points:
Nmeas
1
VDC,avg = ∑ VDC,i (11)
Nmeas i=1

2.10.2 Power Flow Optimization

The interlinking converter power reference is optimized based on economic dispatch and grid constraints:

∗
PILC = PDC,load − PDC,gen + Ploss,comp + δ Peconomic (12)
where PILC ∗ is the optimal interlinking converter power, P
DC,load is total DC load power, PDC,gen is total DC generation,
Ploss,comp is loss compensation, and δ Peconomic is economic optimization adjustment based on grid tariffs.

6/15
 Figure 2. Secondary Control Flowchart for Voltage Restoration and Power Optimization

2.10.3 SoC Balancing

Secondary control implements SoC balancing among the 50 distributed battery units:

δ Icharge,i = KSoC (SoCavg − SoCi ) (13)

where δ Icharge,i is the charge current adjustment for battery i, and KSoC is the balancing gain.

3 Load Management
3.1 LED Intensity Control
The intelligent load management system implements an Intensity control of LED based on priority to prevent excessive battery
discharge during critical conditions while maintaining the essential lighting services.
The LED intensity control is governed by:


Inominal
 if SoC > SoCthreshold,1
ILED,i = Inominal · (1 − 0.1 · LMi ) if SoCthreshold,2 < SoC ≤ SoCthreshold,1 (14)

Inominal · (1 − 0.2 · LMi ) if SoC ≤ SoCthreshold,2


where SoCthreshold,1 = 50%, and SoCthreshold,2 = 20%.

3.2 Load Shedding Priority

Load shedding follows a hierarchical priority structure:
1. Local critical load is highest priority and is shed last
2. Battery Charging of Solar LED lighting system has second priority and is shed after all the dump load is removed
3. The dump load is shed first and has least priority

7/15
4 Simulation Results and Analysis
4.1 Simulation Setup
The proposed DC microgrid system is simulated in MATLAB/Simulink environment with detailed component models and
control implementations. The simulation was run with a discrete time solver with time step of 1 microsecond. the simulation
parameters are provided in Tables 1–6. The test scenario encompasses the following sequence of events:

• t = 0–0.2 s: System startup and islanded mode operation with 40 kW DC load. Distributed batteries and solar PV supply
local loads.

• t = 0.2 s: Grid connection event. ILC transitions from disabled to grid-connected mode.

• t = 0.3 s: Load step change from 40 kW to 55 kW while in grid-connected mode.

• t = 0.5 s: Grid disconnection event. System transitions back to islanded mode with 55 kW load.

• t > 0.5 s: Continued islanded operation with adaptive LED load management active.

• The Total 50 Lightening units are distributed into two groups of equal capacity of 5kW PV, 24 V, 40kWh usable battery
capacity and 2kW LED power [Link] battery SOC of Group LED1 is kept at 85% and that of Group LED2 is
70%.

This test sequence validates system performance under various operational conditions including startup, mode transitions,
load variations, and islanding events.

Table 2. AC Grid Interface Parameters

Table 1. DC Microgrid Parameters
Parameter Value
Parameter Value Grid Voltage (L-L RMS) 220 V
Nominal DC Bus Voltage 220 V Grid Frequency 50 Hz
DC Bus Capacitance 10 mF ILC Filter Inductance 2 mH
Maximum Voltage Deviation ±10 V (±4.5%) ILC Filter Capacitance 50 µF
Switching Frequency 10 kHz ILC Damping Resistance 0.5 Ω
ILC Switching Frequency 10 kHz

Table 4. Energy Storage Parameters

Table 3. Generation Parameters
Parameter Value
Parameter Value Battery per Unit 2 kWh (24 V)
Central PV Plant Rating 50 kW Total Battery Capacity (Installed) 100 kWh
Solar per LED Unit 200 W Usable Capacity (20–100% SoC) 80 kWh
Total Distributed Solar 10 kW (50 units) Battery Chemistry Lithium-ion
MPPT Algorithm Perturb & Observe Maximum Charge Rate 0.5C
Maximum Discharge Rate 1C

Table 6. Control Parameters

Table 5. Load Parameters
Parameter Value
Parameter Value Virtual Resistance Rv 1.5 Ω
DC Load 50 kW Droop Resistance Rd 0.5 Ω
LED Load per Unit 80 W Voltage Controller Gain K p 0.1
Total LED Load (50 units) 4 kW Voltage Controller Gain Ki 5
Dump Load Capacity 10 kW Current Controller Gain 10
Secondary Control Period 100 ms

8/15
 (a) DC Bus Voltage Response (b) DC Bus Current Response

Figure 3. DC microgrid performance during mode transitions.

4.2 DC Microgrid Performance

The DC microgrid’s voltage and current regulation performance is evaluated under various operating conditions including
startup, grid connection, load step changes, and islanding transitions, as shown in Figure 3.
Voltage Regulation Performance:
The DC bus voltage as shown in Figure 3. shows steady-state regulation with small variation within ±0.4% (±1 V) from the
nominal operating point. During system startup (0–0.15 s), the voltage rises smoothly with a rise time of 80 ms and overshoot
of 4.2% (235 V peak), settling in 120 ms. This soft-start behavior prevents excessive inrush currents and shows effectiveness of
the primary control’s droop characteristics.
The load changes at t=0.3 s, from 4 kW to 5.5 kW, it causes a voltage deviation of less than 0.4% with little overshoot and
settling time under 20 ms. This voltage regulation shows the effectiveness of the virtual impedance-based droop control in
maintaining voltage of Microgrid despite significant load variations.
Grid disconnection at t=0.5 s results in a minor voltage dip with recovery time of 40 ms. The smooth transition is due to the
battery and distributed generation sharing current without load interruption.
Current Distribution and Load Following:
Figure 3b shows the DC bus current, which clears the power flow and load conditions. During the islanded operation during
steady state (0.15–0.2 s), the bus current stabilizes at 200 A, corresponding to 46 kW, which includes the 4 kW DC load plus
distributed LED loads and system losses (approximately 4.3%).
The load change at t=0.3 s causes a current increase of 20 A (from 205 A to 225 A), it is 9.8% change. The calculated
power increase of 4.6 kW includes the 1.5 kW load change plus increased battery charging and PV power delivery. The current
transient settling time is under 30 ms showing the distributed converters’ coordinated operation via. the droop mechanism
without needing communication.
The grid disconnection occur at t=0.5 s, the DC bus current increases slightly to maintains a regulation of (±2 A variation).
This increase indicates that the local generation and battery storage continue to supply the entire load in islanded mode. The
transient during mode switching is well within limits and shows the effectiveness of the bidirectional interlinking converter
control. The DC microgrid achieves the following quantified performance:
• Steady-state voltage regulation: ±0.4% (230 ± 1 V)
• Load step response: <0.4% voltage deviation, <20 ms settling time
• Grid disconnection: –0.9% voltage dip, 40 ms recovery
• Current regulation: ±2 A variation in steady-state
• All transients are well within limits
The results shows that the virtual impedance-based droop control is maintaining required voltage regulation while allowing
contribution from distributed LED lightening systems without communication. The fast transient response is due to the

9/15
hierarchical control where the primary control is handling the fast dynamics of milliseconds while the secondary control handles
the steady-state variations from few milliseconds upto seconds.

4.3 AC Grid Interface Performance

The performance of 60 kW interlinking converter is shown using the voltage and current waveforms while operating in Grid
connected mode and Islanded mode in Figure 4.

(a) ILC AC Voltage Waveforms (b) ILC AC Current Waveforms

Figure 4. Three-phase AC performance of interlinking converter.

Three-Phase Voltage Quality:

Figure 4a shows the three-phase AC voltage waveforms at the ILC terminals. The operating point with 220 V RMS (310 V
peak) nominal voltage. The three phases has a 2% amplitude variation showing proper ILC three-phase inverter control.
During the load change at t=0.3 s, the AC voltage shows a variation of 3.2% (<10 V peak variation) with recovery within
one AC cycle (20 ms at 50 Hz). This fast response is due to the effective AC voltage regulation loop despite sudden load
changes.
The grid disconnection occurs at t=0.5 s that results in a voltage dip of 6.5% (from 310 V to 290 V peak) with 30 ms
recovery time. The seamless transition without voltage or frequency distortion is due to the better islanding detection and
seamless transfer capability. Three-Phase Current Performance:
Figure 4b shows the three-phase AC current waveforms, which defines the power transfer and converter loading characteris-
tics. During grid connected mode (t=0.2–0.5 s), the steady-state current is 71 A RMS per phase, which at 220 V three-phase is
equal to to 46.8 kW
The 50 kW active power export shown in Figure 6, including the converter losses of approximately 6.4% confirms this load
current.
The load change at t=0.3 s results in a current increase from 71 A to 79 A RMS, representing an 11.3% increase. The power
increase of approximately 5.3 kW three-phase corresponding to 79 A is a result of good load-following capability. The current
overshoot is less than <5% with a settling time of 30 ms, it confirms a well damped current control loop.
Current Transient Analysis:
The three-phase current variations are below 3%.
Performance Summary:
The AC side of the microgrid shows:

• Three-phase voltage balance as there is less than <2% variation

• On load change, the respons is 11.3% current increase, less than 5% overshoot with a 30 ms settling time

• Phase current are balanced with a 3% variation

10/15
 Figure 5. ILC transient analysis

(a) Active Power Flow (b) Reactive Power

Figure 6. Active and reactive power

4.4 Power Flow Analysis

Figure 6 shows variations in active and reactive power through the interlinking converter for startup period, during grid transition
from islanded to on-grid and also the significant load change.
Active Power Management:
Figure 6a shows the active power flow through the ILC during all the operating modes. During the initial islanded
period (0–0.2 s), the system is operating on Solar LED lightening system’s PV, local battery and main solar generation with

11/15
approximately 40 kW internal DC bus power flow supporting local loads.
After grid connection at t=0.2 s, the system exports around 50 kW to the AC side, the grid-support capability is shown
where excess solar generation is fed to the Ac side. This Power export is due to the successful coordination between the
centralized 50 kW solar PV plant and the distributed 10 kW solar LED Lightening units.
The load change at t=0.3 s from 4 kW to 5.5 kW is managed by increased grid power export from 50 kW to 55 kW. The
power response shows an overshoot of approximately 8% (peak 59 kW) with settling time of 30 ms. This fast, well-damped
response is attributed to effective power control loop tuning and reflects the ability to accommodate sudden load changes.
The power regulation is maintained at ±1 kW variation (±2% of setpoint), which indicates better control and low
disturbances from the distributed LED lightening sources.
After the grid disconnection at t=0.5 s the active power is supplied by LED Lightening system’s battery and PV generation
with total load being supplied by the DC Microgrid alone.
Reactive Power Control and Power Factor:
Figure 6b shows that the reactive power variation across the load changes and mode transitions is low (±0.1 VAR,
representing ±0.2% variation). During the load change at t=0.3 s, the reactive power increase is only 0.55 VAR, due to the
independent control of active and reactive power components because of proper decoupled dq-axis control implementation.
During Grid Transition there is a short spike of Reactive power due to the presynchronization of ILC causing active power and
reactive power to find new balance for incoming grid connection.

4.5 Adaptive LED Load Management Performance

Figure 7 shows the analysis of the distributed solar-LED-battery system, where all 50 units are taken into two groups on with
Battry SoC 85% and second group with 70% SoC. The LED Power confirms the rule based power dimming for the LED system
based on battery SoC.

(a) LED Unit Group Currents (b) LED Power Consumption

Figure 7. LED system performance

System Configuration Context:

The microgrid includes 50 distributed solar-LED-battery units, each comprising 200 W solar PV panel with MPPT, 80 W
rated LED street light with PWM dimming, 2 kWh battery storage (24 V), and bidirectional DC-DC converter. These 50
units are organized into two groups (Group 1 with 25 units, Group 2 with another 25 units), enabling the overall control
while maintaining distributed operation without centralized communication with the help of Virtual Impedance applied to the
switching control.
Group Current Output Analysis:
Figure 7a presents the combined output currents IS1 and IS2, representing the total current contribution from each 25-unit
group to the DC microgrid bus. During the startup period (0–0.05 s), both groups shows an inrush current of approximately
45 A, corresponding to 1.8 A per unit. Attributed to the DC-DC converter capacitor charging and soft-start dynamics, the short
duration ( 50 ms) and low magnitude indicate that the soft-starter is limiting.
when the the DC voltage is higher (0.05–0.45 s), both groups reduce their output current to low levels (<2 A combined,
<0.1 A per unit), indicating that rule based control is limiting current output. This shows the adaptive load management
responding to battery state-of-charge conditions and DC bus Voltage.

12/15
 At t > 0.45 s, as the Voltage drops due to droop control of main solar PV, the currents increase to IS1 = 15 A (Group 1,
0.6 A per unit) and IS2 = 14 A (Group 2, 0.56 A per unit). At the 220 V DC bus voltage, this corresponds to Group 1: 132 W
per unit and Group 2: 123 W per unit. These currents represent the net power delivered from the solar-LED-battery units to the
microgrid including converter losses.
The current difference between the two groups (1 A average). This is attributed to the Battery Soc of 70% triggering
LED dimming and Battery SoC of another group at 85% maintaining full intensity. thereby limiting current from low SoC
Battery. Distributed control triggers where groups autonomously adjust current sharing based on local SoC without requiring
communication.
LED Power Consumption Analysis:
Figure 7b presents the total LED power consumption for each 25-unit group, directly demonstrating the adaptive dimming
strategy. Steady-state power consumption shows LED1 (Group 1) at 2100 W total (84 W per unit, 105% of 80 W rating,
operating at full/near-full brightness) and LED2 (Group 2) at 1600 W total (64 W per unit, 80% of rating, operating at reduced
brightness for battery conservation).
The power consumption by LED’s between the two groups shows the adaptive strategy: Group 2 / Group 1 ratio = 1600 W /
2100 W = 76.2%, there is a 23.8% dimming reduction in Group 2. This 24% power reduction is the intentional adaptive load
management based on battery state-of-charge conditions, providing: (1) extended battery charge state for Group 2 with low
SoC; (2) system-level load reduction of 300 W (7.5% of rated capacity); and (3) maintaining the reduced lighting rather than
total outage.

5 Discussion
The simulation results shows the effectiveness of the proposed hierarchical control strategy for DC microgrids with distributed
solar LED lighting systems. The virtual impedance-based droop control results in accurate current sharing without needing
communication, it is scaled to 50 distributed units. The steady-state voltage regulation of ±0.4% and fast transient response of
less than <20 ms settling time confirms the effectiveness of the two-layer hierarchical architecture.
The adaptive LED load management strategy results in extending of battery SoC and longer runtime in distributed systems.
The demonstrated 24% power reduction during low DC bus voltage conditions translates to 31.6% runtime extension for
affected groups, while maintaining 80% illumination levels. This approach is better than load shedding schemes that result in
complete service interruption.
The reactive power variation less than 50 VAR at 50 kW active power reduces distribution network losses and voltage
drop. The bidirectional interlinking converter manages power flow in both grid-connected and islanded modes, with seamless
transitions validated through load change and grid disconnection.
The proposed approach offers several advantages for Microgrid: (1) no communication required between distributed units,
reducing system cost and improving reliability; (2) inherent redundancy through distributed control, with no single point of
failure; (3) scalable architecture with 50 units is shown and is extensible to larger systems (4) adaptive load management
extending battery operational time without any human intervention

6 Conclusion and Future Work

This paper presents a hierarchical control strategy for grid-connected DC microgrids with intelligent solar lighting and
bidirectional AC-DC interfacing. The system integrates distributed LED converters. solar PV plant , bidirectional interlinking
converter, and electronic load controller for islanded operation.
Key Contributions:
1. Virtual impedance-based droop control achieving <2% current sharing accuracy among 50 converters without communi-
cation infrastructure
2. Dual-mode interlinking converter control enabling seamless grid-connected to islanded transitions (<100ms)
3. Hierarchical power management coordinating distributed generation, battery storage, dump loads, and intelligent LED
dimming
Future Research Directions:
1. Implement Deep reinforcement learning to autonomously tune virtual impedance and droop gain in real-time.
2. Implementation of tertiary control to implement prediction of Solar PV irradiation data
3. Implementation of Fault detection/isolation algorithms

13/15
 4. Hardware validation on realtime platforms and full-scale prototypes

The proposed architecture demonstrates potential for rural electrification, campus microgrids, and islanded communities
requiring reliable distributed solar integration.

References
1. Dragicevic, T., Lu, X., Vasquez, J. C. & Guerrero, J. M. DC microgrids—Part I: A review of control strategies and
stabilization techniques. IEEE Trans. Power Electron. 31, 4876–4891 (2016).
2. Dragicevic, T., Lu, X., Vasquez, J. C. & Guerrero, J. M. DC microgrids—Part II: A review of power architectures,
applications, and standardization issues. IEEE Trans. Power Electron. 31, 3528–3549 (2016).
3. Guerrero, J. M., Vasquez, J. C., Matas, J., de Vicuna, L. G. & Castilla, M. Hierarchical control of droop-controlled AC and
DC microgrids—A general approach toward standardization. IEEE Trans. Ind. Electron. 58, 158–172 (2011).
4. Peyghami, S., Mokhtari, H. & Blaabjerg, F. Autonomous operation of a hybrid AC/DC microgrid with multiple interlinking
converters. IEEE Trans. Smart Grid 9, 6480–6488 (2018).
5. Shafiee, Q., Guerrero, J. M. & Vasquez, J. C. Distributed secondary control for islanded microgrids—A novel approach.
IEEE Trans. Power Electron. 29, 1018–1031 (2014).
6. He, J. & Li, Y. W. Analysis, design, and implementation of virtual impedance for power electronics interfaced distributed
generation. IEEE Trans. Ind. Appl. 47, 2525–2538 (2011).
7. Lu, X., Guerrero, J. M., Sun, K. & Vasquez, J. C. An improved droop control method for DC microgrids based on low
bandwidth communication with DC bus voltage restoration and enhanced current sharing accuracy. IEEE Trans. Power
Electron. 29, 1800–1812 (2014).
8. Tuladhar, A., Jin, H., Unger, T. & Mauch, K. Control of parallel inverters in distributed AC power systems with
consideration of line impedance effect. IEEE Trans. Ind. Appl. 36, 131–138 (2000).
9. Anand, S., Fernandes, B. G. & Guerrero, J. Distributed control to ensure proportional load sharing and improve voltage
regulation in low-voltage DC microgrids. IEEE Trans. Power Electron. 28, 1900–1913 (2013).
10. Kakigano, H., Miura, Y. & Ise, T. Low-voltage bipolar-type DC microgrid for super high quality distribution. IEEE Trans.
Power Electron. 25, 3066–3075 (2010).
11. Elsayed, A. T., Mohamed, A. A. & Mohammed, O. A. DC microgrids and distribution systems: An overview. Electr.
Power Syst. Res. 119, 407–417 (2015).
12. Kumar, D., Zare, F. & Ghosh, A. DC microgrid technology: System architectures, AC grid interfaces, grounding schemes,
power quality, communication networks, applications, and standardizations aspects. IEEE Access 5, 12230–12256 (2017).
13. Meng, L. et al. Review on control of DC microgrids and multiple microgrid clusters. IEEE J. Emerg. Sel. Topics Power
Electron. 5, 928–948 (2017).
14. Nguyen, T. L., Guillo-Sansano, E., Syed, M. H., Nguyen, V. H., Blair, S. M. & Reguera, L. Multi-agent system with plug
and play feature for distributed secondary control in microgrid controller. in IEEE PES Innov. Smart Grid Technol. Conf.
Europe 1–6 (2018).
15. Han, Y., Li, H., Shen, P., Coelho, E. A. A. & Guerrero, J. M. Review of active and reactive power sharing strategies in
hierarchical controlled microgrids. IEEE Trans. Power Electron. 32, 2427–2451 (2017).
16. Khorsandi, A., Ashourloo, M. & Mokhtari, H. A decentralized control method for a low-voltage DC microgrid. IEEE
Trans. Energy Convers. 29, 793–801 (2014).
17. Nasirian, V., Moayedi, S., Davoudi, A. & Lewis, F. L. Distributed cooperative control of DC microgrids. IEEE Trans.
Power Electron. 30, 2288–2303 (2015).
18. Gao, F., Kang, R., Cao, J. & Yang, T. Primary and secondary control in DC microgrids: A review. J. Mod. Power Syst.
Clean Energy 7, 227–242 (2019).
19. Augustine, S., Mishra, M. K. & Lakshminarasamma, N. Adaptive droop control strategy for load sharing and circulating
current minimization in low-voltage standalone DC microgrid. IEEE Trans. Sustain. Energy 6, 132–141 (2015).
20. Prabhakaran, P., Goyal, Y. & Agarwal, V. Novel nonlinear droop control techniques to overcome the load sharing and
voltage regulation issues in DC microgrid. IEEE Trans. Power Electron. 33, 4477–4487 (2018).

14/15
21. Peng, Y., Shuai, Z., Shen, X., Tu, C., Li, Y. & Cheng, M. Flexible robust control strategy for AC/DC hybrid microgrids
based on event-triggered mechanism. in IEEE Appl. Power Electron. Conf. Expo. 3043–3047 (2019).
22. Parisio, A., Rikos, E. & Glielmo, L. A model predictive control approach to microgrid operation optimization. IEEE Trans.
Control Syst. Technol. 22, 1813–1827 (2014).
23. IEEE Standard for Interconnection and Interoperability of Distributed Energy Resources with Associated Electric Power
Systems Interfaces. IEEE Std 1547-2018 (2018).
24. Liu, J., Zhang, W. & Rizzoni, G. Robust stability analysis of DC microgrids with constant power loads. IEEE Trans. Power
Syst. 33, 851–860 (2018).
25. Malik, S. M., Sun, Y., Huang, W., Ai, X. & Shuai, Z. A generalized droop strategy for interlinking converter in a standalone
hybrid microgrid. Appl. Energy 226, 1056–1063 (2018).

Acknowledgements
The authors acknowledge the support of the National Institute of Technology Uttarakhand for providing computational resources
and simulation facilities.

Competing Interests
The authors declare no competing interests.
Data availability: The simulation data and MATLAB/Simulink models that support the findings of this study are available
from the corresponding author upon reasonable request. Detailed algorithmic implementations are provided in the supplementary
algorithms document.

15/15