Adaptive Virtual Inertia Control Method Based on

Droop Control
Lingyu Sun Pengcheng Wang Feng Jiang
Polytechnic Institute College of Electrical Engineering College of Electrical Engineering
Zhejiang University Zhejiang University Zhejiang University
Hangzhou, China Hangzhou, China HangZhou, China
[email protected] [email protected] [email protected]

Wei Zhang Ming Zhong Yuqiang Wang

Inner Mongolia Power Research Inner Mongolia Power Research Inner Mongolia Power Research
Institute Institute Institute
Inner Mongolia Enterprise Key Inner Mongolia Enterprise Key Inner Mongolia Enterprise Key
Laboratory of Smart Grid Simulation of Laboratory of Smart Grid Simulation of Laboratory of Smart Grid Simulation of
Electrical Power Electrical Power System Electrical Power System
Hohhot, China Hohhot, China Hohhot, China
[email protected] [email protected] [email protected]

Abstract—With the increasing proportion of renewable proposed. By adding virtual inertia and damping to the
energy in microgrids, traditional droop control without inertia inverter control, the inverter can simulate the response
2021 IEEE 30th International Symposium on Industrial Electronics (ISIE) | 978-1-7281-9023-5/21/$31.00 ©2021 IEEE | DOI: 10.1109/ISIE45552.2021.9576329

can hardly meet the needs of inverter control. Load changes can characteristics of the traditional synchronous machine
cause sudden changes in the frequency of the microgrid system, [6].However, if the virtual inertia is set too large, the response
which can easily trigger the protection device of the system. The speed of the system will slow down, and it is easy to cause
stable operation of the microgrid is affected. At the same time, system power oscillation. If the virtual inertia is set too small,
if there is a large inertia in the microgrid, it will reduce the sudden changes in frequency will caused in microgrid[7].
response speed of the system and easily cause oscillations. In this
paper, an adaptive inertia control method is proposed. Then, the An low-pass filter is added to the traditional droop control,
stability analysis of the control strategy is carried out. Finally, which has an equivalent control effect with the VSG[5]. By
simulations results verify the correctness and effectiveness of the adjusting the droop coefficient in real time, the virtual inertia
proposed method. of the inverter is dynamically changed to improve the output
characteristics of the inverter in different scenarios. However,
Keywords—droop control, virtual inertia, three-phase inverter, in this type of method, the real-time frequency in the system
rate of change of frequency (ROCOF), virtual synchronous needs to be acquired by the phase-locked loop(PLL) in real
generator (VSG)，microgrid time, which requires high speed and accuracy of the phase-
locked loop [8-10].
I. INTRODUCTION
As the proportion of renewable energy in the traditional The frequency change is divided into two states for control
power system increases, it has a negative impact on the stable in[11]. The value of the virtual inertia is set according to the
operation of the power system. In the traditional power frequency offset direction. The value of virtual inertia is set
generation system, the synchronous machine is the main according to the microgrid operation mode[12]. A small
power generation unit. Renewable energy in the power system inertia is used in inverters in the grid-connected mode to
is connected through the inverter. Since there is no inertia in suppress frequency oscillations. In the islanding mode, larger
traditional inverters, the changes of the load will cause the inertia is used to avoid sudden changes in frequency. However,
sudden changes of frequency in the system[1]. The change of these methods lack the analysis of the influence of inertia on
frequency will trigger the protection device in the microgrid the ROCOF, and cannot directly limit the value of the ROCOF.
system, which will have a negative impact on the stable In this paper, an adaptive virtual inertia control method
operation of the system. based on droop control is proposed. The small signal model of
In the microgrid, the inverter adopts droop control, which the inverter with droop control is established. The influence of
has the characteristics of no communication, strong flexibility the change of the time constant in the filter on the stability of
and high redundancy[2-4]. However, traditional droop control the control strategy is analyzed. By designing the parameters
has no inertia. Load changes and power fluctuations will cause of the adaptive virtual inertia, ROCOF in the microgrid is
sudden changes in frequency. Since there are frequency limited. Finally, simulations verify the effectiveness of the
change protection devices in the microgrid, when the proposed control strategy.
frequency change rate exceeds the set threshold, the power
II. COMPARISON OF DROOP CONTROL AND VSG
generation unit will disconnect from the grid [5].
A. Operation principle of droop control
In order to solve the problem of lack of inertia in inverter
control, the concept of virtual synchronous generator(VSG) is The droop control is developed based on the theory of
power flow control. The amplitude and frequency of the
This work was funded by Inner Mongolia Autonomous Region Science output voltage are adjusted by calculating the active and
and Technology Project: 2019GG373, Inner Mongolia Electric Power Group reactive power of the inverter [13].
(Limited) Company Science and Technology Project (issue number:
(Neidian Technology [2020] No. 6)). In addition, we would like to express The circuit topology from the inverter to the AC bus can
our special gratitude to the Microgrid Electrical Equipment and Manufacture be simplified as shown in Fig. 1.
Platform (MEEMP) in the Polytechnic Institute of Zhejiang University, on
which many valuable tests in our paper were performed.

978-1-7281-9023-5/21/$31.00 ©2021 IEEE

Authorized licensed use limited to: Universidad Nacional Autonoma De Mexico (UNAM). Downloaded on September 23,2024 at 18:55:50 UTC from IEEE Xplore. Restrictions apply.
  
d ref   

 J 0
 dt
 Pm  Pe  Dm ref  

  


V  Vref  nv Q  Qref 
Fig. 1. Simplified circuit topology from inverter to AC bus J is the virtual inertia coefficient of the VSG; ω0 is the
rated angular frequency of the inverter; ωref is the reference
In Fig. 1, E is the output voltage amplitude; ϕ is the phase value of the inverter output angular frequency; ω is the actual
angle difference between the inverter output voltage and the output angular frequency of the inverter; Pm is the mechanical
AC bus voltage; E ∠ ϕ is the inverter output open circuit power; Pe is the electromagnetic power; Dm is the damping
voltage; U∠0 is the AC bus voltage; Z∠θ=jωL+R is the coefficient; Vref is the reference value of the output voltage
output line impedance; ω is the output angular frequency; P is amplitude; V is the actual amplitude of the inverter output
the active power of the inverter; Q is the reactive power of the voltage; nv is the reactive power control droop coefficient; Q
inverter. is the output reactive power; Qref is the reference value of the
When the impedance between the inverter and the AC bus reactive power.
is approximately inductive, the active and reactive power The control block diagram of the VSG is shown in Fig. 2.
output from the inverter to the bus can be expressed as

 UE
 P  Z
   
Q  U ( E  U )
 Z

It can be obtained from (1) that the active power output by

the inverter is proportional to the phase angle, and the reactive
power is proportional to the amplitude of the output voltage.
The traditional droop control can be expressed as[14,15]. Fig. 2. Power control structure of VSG

  ref  mP C. Droop control with low pass filter

    The traditional droop control has a fast response speed, but
V  Vref  nQ due to the lack of virtual inertia, its transient characteristics
can easily lead to large frequency changes. Adding a low-pass
ωref and Vref are the reference values of the angular filter to the droop control can adjust the response speed of the
frequency and amplitude of the inverter output voltage; ω and droop control. The active droop control with a low-pass filter
V are the angular frequency and amplitude of the inverter can be expressed as
output voltage; m and n are the droop coefficients of the active
and reactive power control. The values of m and n are m
determined according to the maximum output power and the     ref  P 
maximum offset of the output voltage. The calculation method  s 1
of m and n is shown in formula (3).
Where: τ is the time constant of the active power low-pass
filter.   is the time constant of the reactive power low-pass
 max  min filter.
m  Pmax

   
 n  max  Vmin
V
 Qmax

ωmax and ωmin are the maximum and minimum values of

the output angular frequency of the inverter ; Vmax and Vmin are
the maximum and minimum output voltage of the inverter;
Pmax and Qmax are the maximum active power and maximum
reactive power of the inverter.
B. Operation principle of VSG Fig. 3. Droop control structure with low-pass filter
Compared with droop control, virtual inertia and damping Changing the form of formula (5), we can get
are added to the VSG control. Therefore, the instantaneous
response speed of the inverter can be adjusted. The control
equations of active power and reactive power of VSG can be  d ref   
expressed as [12] 
m dt
 P
1
m
 
ref    

Authorized licensed use limited to: Universidad Nacional Autonoma De Mexico (UNAM). Downloaded on September 23,2024 at 18:55:50 UTC from IEEE Xplore. Restrictions apply.
  1 Fig. 5. Improved adaptive inertia droop control structure
Let J  , D  , P  Pm  Pe , formula (6) can be
m0 m The time-domain expression of droop control with low-
expressed as pass filter is shown in equation (9).


d ref   P d 
 J 0
dt
m 
 Pe  D ref       
dt
   mP  

Therefore, the active power droop control equation with a Where: =ref - .
low-pass filter is similar to the active power control equation
of the VSG. The inertia of the inverter can be changed by From ordinary differential equation (9), the relationship
adjusting the parameters of the low-pass filter in the droop between transient output angular frequency and output power
control. can be obtained, as shown in equation (10).

t t
III. ADAPTIVE VIRTUAL INERTIA CONTROL  
   0  e   mP  (1  e  )  
In the microgrid system, the design of virtual inertia
determines the transient response effect of the system and the
stability of the system. Where: 0 is the initial value of the frequency offset.
If the virtual inertia setting in the droop control with a low-
Differentiate  to get the relevant influencing factors of
pass filter is too large, it will cause the system's dynamic
response speed to be slow, and may cause power oscillations; ROCOF.
If the virtual inertia is set too small, it will cause large ROCOF,
which will affect the stable operation of the system. Therefore, d   mP  t    t
  e  e  
while suppressing the ROCOF of the system, power dt  
oscillation should be avoided. Based on the traditional droop
control strategy, this paper proposes an adaptive virtual inertia Combining formula (8) and formula (11) can be obtained
control strategy.
The topology and control structure of the three-phase d  mP 0 -
t

inverter are shown in Fig. 4.  (  )e   

dt  0  KP  0 +KP

Analyzing equation (12) can get the following conclusions

mP m
 0   
 0  KP K

 0 0
 0   
 0 +KP  0

Combining (13) and (14) can be obtained

Fig. 4. Three-phase inverter droop control structure with low-pass filter
 0 mP  0 m
  (  )  
Power adjustment is introduced in the low-pass filter 0  0  KP  0 +KP K
shown in (5). The specific adjustment method is shown in
equation (8). So the value range of ROCOF can be obtained

   0  KP   0 d  m
     
0 dt K
In (7), τ0 is the initial value of the time constant, and K is
the adaptive inertia coefficient. The improved adaptive inertia The design of the parameters m, K and ω0 can realize the
droop control structure is shown in Fig. 5. control of ROCOF in the microgrid system. The inverter
adaptively adjusts the value of inertia according to the output
power to improve the stability of the microgrid system. As the
output power change rate increases, the frequency change rate
gradually increases, and the absolute value approaches m/K
and 0 /τ. The specific parameter design method is
explained in IV.

Authorized licensed use limited to: Universidad Nacional Autonoma De Mexico (UNAM). Downloaded on September 23,2024 at 18:55:50 UTC from IEEE Xplore. Restrictions apply.
 IV. STABILITY ANALYSIS AND PARAMETER DESIGN From equation (21), the characteristic equation of control
can be obtained as
A. Output power modeling and stability analysis of the
inverter
 s3  As2  Bs  C  0  
The equivalent circuit of the inverter connected to the
microgrid is shown in Fig. 6. In the figure, output voltage of
the inverter is E∠δ, the line impedance value is XL=R+jωL=X In the formula(22), A, B, C are respectively
∠θ, the AC bus current flowing into the inverter is I, and the
AC bus voltage is U∠0. 2 mU cos 
 A   
 X

mU cos  nU cos  1
 B=   2  
X 2X 

mU cos  mnU 2 E cos 2  mnU 2 E sin 2 

Fig. 6. Equivalent circuit of inverter connected to microgrid  C   
 2X 2X 2 2X 2
The apparent power output of the inverter can be
expressed as Under different time constants τ, the distribution of control
roots can be calculated by Matlab. The parameters used are
* shown in Table Ⅰ.
 E   U 0 
 S  U  I *  U 0     
 X   TABLE I. ACTIVE POWER DROOP STABILITY ANALYSIS PARAMETERS

Symbol Parameters Value

Then the active power and reactive power output by the U Output voltage of inverter 311V
inverter can be expressed as (18). E Voltage of AC bus 311V
Rated angular frequency
ω 314rad/s
of voltage
 1
 P  [( EU cos   U 2 ) cos   EU sin  sin  ] L Line inductance 0.5mH

 X  m
Active power droop
0.000628
 coefficient
Q  1
[( EU cos   U 2 )sin   EU sin  cos  ] n
Reactive power droop
0.00311
 X coefficient
Time constant of low pass
τ -1~1
filter
When the line impedance is approximately inductive, it
can be approximated as sinθ=1 and cosθ=0. (18) can be
simplified to

 EU
 P  X sin 
  2  
Q  EU cos   U
 X

In order to analyze the dynamic performance of the system,

a small-signal modeling of the droop-controlled inverter Fig. 7. Drooping characteristic root distribution with variable time constant
system is carried out. The small signal disturbance equation of
the inverter output instantaneous power in formula (19) is: As shown in Fig. 7, as the time constant τ increases, a pair
of characteristic roots gradually move to the right of the
 U coordinate axis. Therefore, under this condition, τ should be
 p  X [sin   e  E cos    ] less than 0.08 to ensure the stable operation of the system.
   
 q  U [cos   e  E sin    ] B. Design of Control parameter
 X In droop control, the value of the droop coefficient m
determines the active power sharing effect and the deviation
The small disturbance calculation is performed on the from the frequency. As shown in Fig. 8. The larger the m
droop control equation added with the low-pass filter, and it design, the better the power between different parallel
can be obtained inverters. But the static frequency difference is relatively large;
The smaller the m design, the smaller the static frequency
 m U difference. However, the sharing effect between parallel
    ( s  1) s  X (sin   e  E cos    ) inverters is worse.
   
 e   n  U (cos   e  E sin    )
  s 1 X

Authorized licensed use limited to: Universidad Nacional Autonoma De Mexico (UNAM). Downloaded on September 23,2024 at 18:55:50 UTC from IEEE Xplore. Restrictions apply.
 L Filter inductance 0.5mH
C Filter capacitor 20μF
m Active power droop coefficient 0.000628
n Reactive power droop coefficient 0.00311
τ0 Initial value of time constant 0.02
K Adaptive inertia coefficient 0.0002
Ri Line resistance 0.1Ω
Li Line inductance 0.5mH
Z load1 Load1 10kW
Z load2 Load2 5kW
Z load3 Load3 5kW
In the initial state, all the switches are in the off state, and
Fig. 8. Droop control characteristic curve the inverters 1 and 2 are in the no-load start state; At 0s, close
switch S1, and put in 10kW load 1. At 1s, inverter b starts to
It can be seen from equation (16) that ROCOF is
perform pre-synchronization until the voltage amplitude,
determined by m and K. In this method, the design of m
frequency and phase of inverter a and b are consistent. At 2.35
determines the power sharing effect of the parallel system and
seconds, inverter 2 completes the pre-synchronization and
the static frequency deviation
closes switch SP at the same time, and inverter 2 is put into
Then by designing the K value, the maximum ROCOF operation in the system. At 4s, the switches S2 and S3 are
during the dynamic change of the load is limited. The design closed at the same time, and the load in the microgrid becomes
of K restricts the maximum value of ROCOF. 20kW. At 7s, the switches S1 and S2 are turned off, and the
load running in the system becomes 5kW.
The specific design method of m and K is shown in
formula(26) and (27). Frequency Hz

）
max  min
 m  
Pmax

m
 K  
2 rmax Pre-synchronization Parallel Add 10kW load Reduce 15kW load
Time (s)

Where: ωmax is the upper limit of angular frequency, ωmin Fig. 10. Frequency variation waveform of islanded microgrid
is the lower limit of angular frequency. Pmax is the maximum
output power. rmax is the maximum ROCOF of the inverter. ROCOF(Hz/s)
Reduce 15kW load
At the same time, the design of K should verify the Add 10kW load

stability analysis. If the stability condition is not met, K should

be reduced until the stability boundary is reached.
Parallel

V. SIMULATION RESULTS
In order to verify the above control strategy, the simulation
software PLECS is used to build a microgrid simulation model ROCOF Hz/s
Time (s)
ROCOF Hz/s
（

with two inverters in parallel. The simulation structure

diagram is shown in Fig. 6. The specific parameters of each
device in the island microgrid simulation system are shown in Traditional droop
control with LPF Adaptive virtual interia
Traditional droop
control with LPF
Adaptive virtual interia

Table Ⅱ. Time (s)

Fig. 11. ROCOF of islanded microgrid

Fig. 10 shows the change of AC bus voltage frequency in

the microgrid. The frequency varies from 49.5 Hz to 50.5 Hz,
which meets the design index. Fig. 11 shows the ROCOF of
the AC bus of the microgrid. The blue line in the figure is the
ROCOF of adaptive power droop control, and the red line is
ROCOF of the traditional droop control with low-pass filter.
It can be seen that the maximum ROCOF in the adaptive
droop control is 1.8, and the maximum ROCOF of the
Fig. 9. Simulation system structure of island microgrid traditional droop control with a low-pass filter is 5.2.

TABLE II. SIMULATION PARAMETER OF ISLANDED MICROGRID

SYSTEM

Symbol Parameters Value

U DC DC voltage 800V
E AC voltage amplitudes 311V
f Rated voltage frequency 50Hz

Authorized licensed use limited to: Universidad Nacional Autonoma De Mexico (UNAM). Downloaded on September 23,2024 at 18:55:50 UTC from IEEE Xplore. Restrictions apply.
 Active power w
established, and stability analysis is carried out. Finally, the

（
）
simulation results verify the effectiveness of the proposed
control strategy.
REFERENCES
[1] J. Driesen and K. Visscher, "Virtual synchronous generators," 2008
IEEE Power and Energy Society General Meeting - Conversion and
Delivery of Electrical Energy in the 21st Century, Pittsburgh, PA, USA,
2008, pp. 1-3.
Pre-synchronization Parallel Add 10kW load Reduce 15kW load
Time (s) [2] M. Gao, M. Chen, C. Wang and Z. Qian, "An Accurate Power-Sharing
Control Method Based on Circulating-Current Power Phasor Model in
Fig. 12. Output power of each inverter Voltage-Source Inverter Parallel-Operation System," in IEEE
Transactions on Power Electronics, vol. 33, no. 5, pp. 4458-4476, May
Current (A) 2018.
[3] W. Yao, M. Chen, J. Matas, J. M. Guerrero and Z. Qian, "Design and
Analysis of the Droop Control Method for Parallel Inverters
Considering the Impact of the Complex Impedance on the Power
Sharing," in IEEE Transactions on Industrial Electronics, vol. 58, no.
2, pp. 576-588, Feb. 2011.
[4] M. Gao, C. Zhang, M. Qiu, M. Chen and A. Levy, "Design of control
(s)
system for smooth mode transfer in smart microgrid application," 2016
IEEE Applied Power Electronics Conference and Exposition (APEC),
Time (s) Long Beach, CA, USA, 2016, pp. 138-142.
Current (A) Current (A)
[5] Soni N, Doolia S, Chandorkar M C. Improvement of Transient
Response in Microgrids Using Virtual Inertia[J]. IEEE Transactions on
Power Delivery, 2013, 28(3): 1830–1838 .
Waveform at parallel time (s) Load current waveform [6] Q. Zhong and G. Weiss, "Synchronverters: Inverters That Mimic
Time (s) Synchronous Generators," in IEEE Transactions on Industrial
Electronics, vol. 58, no. 4, pp. 1259-1267, April 2011.
(a)
[7] Meegahapola L, Flynn D. Impact on transient and frequency stability
Current (A) for a power system at very high wind penetration[C]//IEEE PES
General Meeting.
[8] Rolim L, Alves M, Assis T M L, et al. Virtual inertia impact on
microgrid voltage and frequency control[C]//2018 Simposio Brasileiro
de Sistemas Eletricos (SBSE).
[9] Alipoor J, Miura Y, Ise T. Power System Stabilization Using Virtual
Synchronous Generator With Alternating Moment of Inertia[J]. IEEE
Journal of Emerging and Selected Topics in Power Electronics, 2015,
Time (s) 3(2): 451–458.
Current (A) Current (A)

[10] Julius N, Nderu J N, Irungu G K. Frequency Control and Virtual Inertia

Emulation Techniques for Grid Connected Wind Energy Conversion
Waveform at Systems - A Review[C]//2019 IEEE AFRICON.
parallel time Load current waveform
Time (s) [11] Hou X, Sun Y, Zhang X, et al. Improvement of Frequency Regulation
in VSG-Based AC Microgrid Via Adaptive Virtual Inertia[J]. IEEE
(b) Transactions on Power Electronics, 2020, 35(2): 1589–1602.
Fig. 13. Output current waveform of each inverter [12] Meng X, Liu J, Liu Z. A Generalized Droop Control for Grid-
Supporting Inverter Based on Comparison Between Traditional Droop
In Fig. 13 and Fig. 14, during parallel connection and load Control and Virtual Synchronous Generator Control[J]. IEEE
Transactions on Power Electronics, 2019, 34(6): 5416–5438.
changes, the response time of the proposed control strategy
enters a steady state within 0.05s. [13] Guerrero J M, Luis Garcia De Vicuna, Matas J, et al. Output impedance
design of parallel-connected UPS inverters with wireless load-sharing
control[J]. IEEE Transactions on Industrial Electronics, 2005, 52(4):
VI. CONCLUSION 1126–1135.
This paper proposes an adaptive inertia control method [14] Guerrero J M, Matas J, Vicuna L G D V D, et al. Wireless-Control
based on droop control. By introducing the real-time output Strategy for Parallel Operation of Distributed-Generation Inverters[J].
power of the inverter into the control of the time constant, the IEEE Transactions on Industrial Electronics, 2006, 53(5): 1461–1470.
adaptive adjustment of virtual inertia can be realized. The [15] Guerrero J M, Matas J, Vicuna L G de, et al. Decentralized Control for
Parallel Operation of Distributed Generation Inverters Using Resistive
adaptive virtual inertia control method based on droop control Output Impedance[J]. IEEE Transactions on Industrial Electronics,
can effectively suppress ROCOF and improve the stability of 2007, 54(2): 994–1004.
the system. A small-signal model of inverter output power is

Authorized licensed use limited to: Universidad Nacional Autonoma De Mexico (UNAM). Downloaded on September 23,2024 at 18:55:50 UTC from IEEE Xplore. Restrictions apply.