Received 15 September 2023; revised 30 December 2023; accepted 27 January 2024.

Date of publication 31 January 2024; date of current version 13 February 2024.

Digital Object Identifier 10.1109/OAJPE.2024.3360854

Systematic Decoupling Grid-Forming Control

for Utility-Scale Inverter-Based Distributed
Energy Resources in Weak
Distribution Grids
LINA HE (Senior Member, IEEE), AND SHIWEN YU (Graduate Student Member, IEEE)
Department of Electrical and Computer Engineering, University of Illinois at Chicago, Chicago, IL 60607 USA
CORRESPONDING AUTHOR: L. HE ([email protected])

ABSTRACT Existing grid-forming inverter control schemes for distributed energy resources (DERs) pri-
marily rely on active power (P)-frequency (f) and reactive power (Q)-voltage (V) droop mechanisms that are
tailored for highly inductive transmission grids. However, in weak distribution grids where P and Q are highly
coupled due to their resistive network characteristics, these control schemes cannot provide independent and
accurate f and V regulation. This will further deteriorate the dynamic and stability performance, potentially
resulting in inverter and load tripping during disturbances. To address this challenge, this paper proposes an
innovative decoupling grid-forming control scheme, which is designed based on a systematic perspective that
considers the inherent coupling characteristic of the entire distribution grid. The small-signal stability of the
proposed controller is analyzed by varying controller parameters and the grid strength. The effectiveness of
this controller is comprehensively verified using both MATLAB and OPAL-RT platforms by comparing it
with existing grid-forming control strategies. The results show that the proposed controller can effectively
decouple P and Q regulation in weak distribution grids. It enables DERs to provide independent, accurate, and
autonomous f and V regulation, thus improving grid stability and dynamics. The proposed control strategy is
cost-effective, communication-free, and can be easily commercialized due to its straightforward and robust
circuit design.

INDEX TERMS Decoupling control, grid-forming, utility-scale DERs, weak distribution grids, frequency
and voltage regulation.

NOMENCLATURE I. INTRODUCTION
Acronyms
DER Distributed energy resource. U NDER strong clean energy incentives, such as ‘‘100%
carbon-free electricity by 2035’’ of the U.S. power
sector [1], the conventional electricity network is rapidly
P Active power.
f Frequency. transitioning into a more distributed structure to facilitate
Q Reactive power. the widespread integration of renewable DERs [2]. These
V Voltage. DERs are usually integrated into distribution grids via invert-
PV Photovoltaic. ers, which introduce numerous power electronic components
PLL Phase-locked loop. and greatly decrease the system inertia [3]. The resulting
EV Electric vehicle. low inertia can significantly reduce the strength of the grid,
FRT Fault ride-through. making it vulnerable to voltage instability and frequency
PCC Point of common coupling. oscillations. Operational challenges and instability issues of
SCR Short circuit ratio. inverter-based resources in weak transmission grids have
ZDER Output impedance of DER. been documented globally. Notable examples include reports
p.u. Per unit. from the ERCOT [4] and AEMO [5]. These issues can result
2024 The Authors. This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.
For more information, see https://creativecommons.org/licenses/by-nc-nd/4.0/
VOLUME 11, 2024 27
in severe consequences, such as device damage and even extensively utilized to address challenges such as harmonics,
widespread power outages, exemplified by the 2016 South power sharing, and system instability due to the integration
Australia blackout [5] and the 2019 United Kingdom black- of DERs. A coordinated virtual impedance control scheme
out [6]. Similar risks are present in weak distribution grids using sequence-based droop control and virtual impedance
with high penetration of DERs and less rotating mass [7], [8]. is presented to enhance the integration of electric vehicles
In response, the IEEE 1547 Standard provides critical (EVs) with the power grid [21]. It can effectively improve
guidelines for DER integration, emphasizing the need for harmonics mitigation and power quality under both balanced
effective f and V regulation in distribution grids [9]. Mean- and unbalanced events. In [22], a control method for AC/DC
while, there is a notable surge in research focusing on f hybrid microgrids based on dual converters is discussed. This
and V regulation in distribution grids. In [10], the effects of control aims to enhance grid-tied performance, operational
employing thermostatically controlled loads for f regulation transitions, and power sharing in islanded modes. Droop con-
in distribution grids are studied, which reveals increased risks trol can be integrated into a hierarchical control framework
in V variations and transformer aging. Multiple f and V regu- that addresses reactive and harmonic power sharing, espe-
lation techniques such as grid-forming and grid-supporting cially in islanded microgrids [23], [24], [25], [26]. In [27],
converter controls are comprehensively reviewed in [11]. a controller using droop control and virtual resistance is
It particularly focuses on modified control strategies for reported. It aims to provide advanced services such as fault
inductive, resistive, and general line types. In [12], the use ride-through (FRT) [28]. However, these studies focus on
of energy storage systems for f regulation in power sys- the inverter side system viewed from the point of common
tems is investigated based on a bi-level optimization model coupling (PCC) and neglect the influence of the grid-side
with chance constraints and a decentralized algorithm. How- characteristics on controller dynamics. Factors such as the
ever, economic factors are not extensively covered. In [13], grid strength and coupling issue of P and Q are not taken into
an online V regulation method for active distribution grids is account.
discussed using a deep neural encoding-decoding approach. Besides, the other two existing grid-forming control meth-
It is designed to overcome the complexities of traditional ods exhibit inherent limitations. For instance, virtual syn-
mixed-integer quadratic programming by using an offline chronous generator control schemes replicate the inertia and
encoding process that simplifies the problem into quadratic droop characteristics of synchronous generators to achieve f
programming and an online decoding process using a deep and V control [19]. However, this approach does not fully
neural network. In [14], a step control reactive power strategy exploit the rapid response capabilities of inverters and may
for V regulation is presented for unbalanced power systems adversely affect system dynamics. Additionally, virtual oscil-
with residential photovoltaics (PVs). It can effectively man- lator control schemes have emerged in recent years, aiming
age V unbalance and overvoltage. However, the long-term to emulate a nonlinear oscillator with a natural frequency
integration challenges should be comprehensively addressed. aligned with the grid frequency. These schemes are designed
To avoid these risks, it is urgent for DERs, especially to provide support for both frequency and voltage, although
those at the utility-scale, to have grid-forming capabilities. they are still in the early stages of development [20]. Although
These capabilities play a crucial role in regulating the f these control methods exhibit distinct dynamic performances
and V of the power grid to ensure its stability, reliabil- owing to their unique characteristics, they share a common-
ity, and resilience [15]. Grid-forming control schemes have ality in regulating inverter output with a similar steady-state
been demonstrated to enhance damping in comparison to response, as they all adhere to the droop law [29]. It is known
traditional grid-following control schemes. This enhance- that droop-based control is traditionally designed for highly
ment becomes increasingly vital as the penetration level of inductive transmission grids where f and V are decoupled due
inverters increases, contributing to sustained reliable system to a high reactance-to-resistance (X/R) ratio [15]. However,
operation and optimal dynamic performance. this design can lead to instability in scenarios with wide
Existing commercial DERs usually operate as grid- variations in grid impedance X/R ratios and associated power
following sources that act as passive P and Q controlled coupling issues. In weak distribution grids with resistive
sources. They merely regulate their power output by fol- characteristics and coupled P and Q, existing grid-forming
lowing the grid angle/frequency via a phase-locked loop control schemes based on the droop mechanism encounter
(PLL), yet they fall short in providing constructive f and V challenges in achieving independent and effective f and V
support for weak grids [16], [17]. In contrast, DERs that regulation [30].
employ grid-forming control schemes, such as droop con- Addressing the coupling of P and Q is crucial for effective
trol [18], virtual synchronous generator control [19], and f and V regulation, as well as for resolving stability and
virtual oscillator control [20], act as autonomous voltage dynamic issues in DER control. The work in [31] and [32]
sources. They are independent of the PLL and can actively uses modified droop functions that consider both P and Q in
regulate their output f and V, which enables them to oper- voltage and/or frequency regulations. However, they do not
ate as synchronous generators and stabilize the system. sufficiently address the decoupling of P and Q based on grid
Among these grid-forming control methods, droop control is characteristics in their controller designs [31], [32]. Specif-

28 VOLUME 11, 2024

He and Yu: Systematic Decoupling Grid-Forming Control

ically, the study in [31] applied a P-Q-V droop controller, grid impedance compared to the DER system and cannot
which uses both P and Q to regulate the output voltage. How- be neglected in inverter controller design. Consequently, the
ever, the coupling coefficient of the P and Q in this study is ZDER -based decoupling control schemes discussed in [33]
assumed to be a constant, i.e., 1, which is unrealistic for most and [34] are unable to enable utility-scale DERs to provide
distribution grids, thus failing to accurately reflect real-world independent and effective f and V regulation. The reason is
coupling characteristics [15]. In addition, it neglects the cou- that these schemes treat the power grid as an ideal Thevenin
pling of P and Q in f control. Another droop controller based equivalent circuit and neglect the inherent coupling issue of
on P-V and Q-(-δ) is discussed in [33]. It aims to ensure weak distribution grids. As a result, this oversight might exac-
stable parallel operation of diverse inverter types, facilitating erbate vulnerabilities of such weak grids, posing threats to
accurate load sharing and improved stability. However, it is system stability and resilience. Thus, addressing the coupling
important to note that this control method is tailored for a issue of P and Q on the grid side is essential to prevent severe
passive load rather than a power grid, which may limit its instability in power systems, particularly those facing large
responsiveness to changes in grid dynamics. grid impedance variations due to diverse grid structures and
Additional approaches involve DER output impedance strengths.
(ZDER )-based decoupling control methods, including vir- To overcome these challenges, this paper proposes a novel,
tual impedance-based and virtual frame-based grid-forming comprehensive, and systematic decoupling grid-forming con-
control methods [34], [35]. The virtual impedance-based trol scheme based on the widely used droop control. The
control integrates a large virtual inductor into the control droop control is more appealing than other grid-forming
loop to compensate for the resistance of the outlet distribu- control methods due to its basic simplicity and fast-response
tion line connecting the inverter to the PCC [34]. However, characteristics. The proposed control considers not only the
this method will compromise the stability margin and may inverter and its associated components but also the power
decrease the system damping. The virtual frame control uses grid with intricate power coupling characteristics. Unlike
a transformation frame in the inverter control to build a virtual existing decoupling methods that primarily focus on the
f and V frame, aiming to decouple the output P and Q of circuit of the inverter and its associated components, our
the DER [35]. Nevertheless, these methods primarily address approach rigorously incorporates the grid-side dynamics,
the coupling issue at the DER system level, including the addressing the challenges introduced by the high R/X ratio
filter and outlet line, without considering the coupling of P of weak distribution grids. This holistic design philosophy
and Q on the grid side. This oversight can lead to severe is based on an innovative systematic decoupling of P and
instability issues in real-world power systems with significant Q, enhancing the dynamic performance and stability of the
grid impedance variations due to varying grid structures and power grid with distributed energy resources. It introduces
strength. a significant advancement in grid-forming control strate-
Recent studies have shed light on how grid-side character- gies, providing a more robust and adaptive solution for
istics significantly impact control dynamics [36], [37], [38], next-generation distribution grids with a high penetration
[39]. In [36], the exploration of control stability under varying of DERs.
grid impedance conditions highlights the instability risks of Specifically, the proposed control strategy aims to decou-
grid-connected inverters, particularly in scenarios of high grid ple P and Q of weak distribution grids in real time by
impedance. Reference [37] analyzes the small-signal stability extracting their coupling characteristics online and designing
of power inverters and the impedance characteristics of power an adaptive grid-forming decoupler. This allows DERs to
systems. The impedance behavior of two-stage photovoltaic provide independent f and V regulation, thereby enhancing
inverters is investigated in [38], demonstrating how power system resilience. The proposed grid-forming controller is
inverter dynamics can lead to instability, especially in weak cost-effective and easily commercializable due to its simple
grids with a high R/X ratio. Additionally, [39] explores the and robust circuit design. Additionally, as a grid-forming
sensitivity of inverter-based resources to grid line impedance, control, the strategy operates autonomously on local mea-
especially in residential applications. This study underscores surements, eliminating the need for communication. It is
the necessity for sophisticated control strategies that take into universally applicable to all DERs, reducing communication
account the coupling relationship between P and Q of the grid costs and minimizing vulnerability to cyber-attacks. The per-
side. formance of the proposed control technique is thoroughly
Special consideration should be given to weak distribution validated using a detailed weak distribution grid model
grids with complex grid-side characteristics. A weak distribu- based on the IEEE 14 bus system through comprehensive
tion grid is defined as having a lower short circuit ratio (SCR) MATLAB and OPAL-RT simulation. The results demon-
of less than 3 and an X/R ratio of less than 5 [40]. These grids strate that the proposed control can independently regulate
typically exhibit low inertia and are susceptible to voltage and f and V by decoupling P and Q online. It exhibits supe-
frequency instability. Furthermore, weak distribution grids rior stability, steady-state, and dynamic performance when
are characterized by a low X/R ratio, resulting in a signifi- compared to existing grid-forming control strategies. To the
cant coupling issue of P and Q. They usually have a larger best knowledge of the authors, this paper represents the first

VOLUME 11, 2024 29

 and (4) [12].
EV ϕ
P= (3)
Z
E (E − V )
Q= (4)
Z
As seen, when the magnitudes of the PCC voltage and line
FIGURE 1. Traditional equivalent grid-connected DER system for impedance are fixed, a linear relationship is observed both
control design. between P and ϕ and between Q and E. This linearity facili-
tates the applicability of droop characteristics in grid-forming
inverter control within transmission systems. Specifically,
comprehensive exploration of systematic decoupling control these relationships can be represented as (5) and (6) [14] and
for inverters in weak distribution grids. [18].
The organization of the remaining sections in this paper
is outlined as follows. Section II derives the coupling char- f ∗ = f0 − m(Pmea − P0 ) (5)
acteristics of the power grid and introduces the systematic E ∗ = E0 − n(Qmea − Q0 ) (6)
decoupling grid-forming controller. Section III formulates the
small-signal model of the proposed controller, along with where f ∗ and E ∗ represent the generated references of fre-
detailing the parameter design with consideration of the sta- quency and amplitude of the inverter terminal voltage, and
bility analysis. Section IV provides the simulation scenarios f0 and E0 denote their rated values; Pmea and Qmea are the
and corresponding results. Finally, Section V serves as the measurements of the output active and reactive power; P0
conclusion for the paper. and Q0 are the setpoints based on the dispatched active and
reactive power, indicating the output power of the inverter
II. PROPOSED SYSTEMATIC DECOUPLING at nominal frequency f0 and nominal voltage magnitude E0 ;
GRID-FORMING CONTROL m and n represent the positive coefficients of P-f and Q-E
The proposed systematic decoupling grid-forming control is droop slopes, respectively. Note that the P-f droop scheme
designed based on the coupling characteristics of the weak is used instead of the P-ϕ droop scheme, as the f refer-
distribution grid, as derived in this section. ence can be easily fixed at nominal values, while the initial
phase angle of the PCC voltage cannot be known by the
A. TRADITIONAL DROOP CONTROL MECHANISM inverter [41].
Traditional inverter control design considers the power grid as However, the assumptions inherent in droop control are
an ideal voltage source at the PCC, as shown in Fig. 1 [17], limited to highly inductive transmission grids. In medium
[18], [19], [20]. or low-voltage distribution grids, which are predominantly
The P and Q transmitted from the inverter to the grid can resistive and weak, these assumptions lose their validity.
be represented by the voltages at the inverter terminal and the As a result, traditional droop-based control becomes inca-
PCC, along with the impedance between these two points, pable of independently regulating f and V by adjusting P
as shown in (1) and (2). and Q respectively. The underlying reasons are outlined
EVsinϕ E(E − Vcosϕ) below:
P= sin θ + cos θ (1)
Z Z 1) For a distribution line, its power flow equations (1)
EVsinϕ E(E − Vcosϕ) and (2) cannot be simplified to (3) and (4) due to its low
Q=− cos θ + sin θ (2)
Z Z X/R ratio. As a result, the output P and Q of inverters
where P and Q denote the output active and reactive power of become coupled and associated with the f and V.
the DER, respectively; E and V denote the magnitudes of the 2) A weak distribution grid typically exhibits a high grid
inverter terminal voltage and PCC voltage, respectively. The impedance, so it cannot be treated as an ideal voltage
parameter ϕ represents the phase angle difference between source, as depicted in Fig. 1. Moreover, the resistive
the inverter terminal voltage and PCC voltage (obtained from dominant characteristics of the high grid impedance
ϕ1 − ϕ2 ); while Z ̸ θ represents the output impedance of the result in interdependence between f and V control due
DER system, comprising the impedances of the filter and to the coupling of P and Q. Therefore, the weak dis-
outlet line. tribution grid and its associated impacts should not be
In high-voltage transmission grids, the impedance Z ̸ θ in overlooked in inverter control design.
Fig. 1 is typically highly inductive (θ ≈ 90◦ ) due to the In weak distribution grids, the coupling issue can lead to
high X/R ratio of the transmission line. In addition, the phase significant mutual interference between f and V control in
angle difference ϕ is usually negligible. As a result, under DERs. In extreme situations, this interference may result in
the assumptions of sin θ ≈ 1, cos θ ≈ 0, sinϕ ≈ ϕ, and system instability, severe device damage, or even widespread
cosϕ ≈ 1, (1) and (2) can be simplified, resulting in (3) outages.

30 VOLUME 11, 2024

He and Yu: Systematic Decoupling Grid-Forming Control

of the system viewed from the inverter terminal, as illus-

trated in Fig. 2. Different from the model in Fig. 1, this
equivalent circuit showed in Fig. 2 consists of two time-
varying components: a time-varying impedance Zs ̸ θs and
a time-varying voltage source Vg ̸ ϕ2 . These components
reflect the online dynamics of the studied weak system and
can be determined online based on real-time measurements
during system operation. Zs ̸ θs represents the total system
impedance as seen from the inverter terminal. It is the sum
of three parts, including the filter impedance on the system
side, the impedance of the outlet distribution line, and the
equivalent impedance of the weak distribution grid. The first
two parts are readily determined based on the parameters of
the inverter and distribution line provided by DER and utility
operators. The equivalent impedance of the weak distribution
grid can be identified online by monitoring the voltage and
current variations at the PCC during events like load changes,
as shown in (7).
⇀ ⇀
V1 − V2
Zg ̸ θg = ⇀ ⇀
(7)
I1 − I2
⇀ ⇀
where V 1 and I 1 represent the pre-event measurements of
⇀ ⇀
FIGURE 2. Systematic decoupling grid-forming control system.
voltage and current phasors, respectively, while V 2 and I 2
represent the post-event measurements. By applying KVL,
⇀
the equivalent grid voltage Vg ̸ ϕ2 can be deduced using V 2
and Zg ̸ θg , as shown in (8).
B. PROPOSED SYSTEMATIC DECOUPLING
⇀ ⇀
GRID-FORMING CONTROL Vg ̸ ϕ2 = V 2 − Zg ̸ θg ∗ I 2 (8)
To overcome this challenge, this paper proposes a systematic
Based on the developed equivalent time-varying system
decoupling grid-forming control scheme. This scheme aims
circuit, the combined system impedance, including both the
to decouple the inverter output P and Q with consideration
output impedance and grid impedance, is determined in (9).
of the comprehensive coupling characteristic of the power
grid. It enables DERs to independently control f and V in Zs ̸ θs = Zl ̸ θl + Zg ̸ θg (9)
weak distribution grids and improve dynamic performance.
The configuration of the proposed control scheme is shown where
q
in Fig. 2. As depicted, a DER system is connected to a weak Zs = Zl2 + Zg2 (10)
distribution grid via an LCL filter and an outlet distribution
Zl sin θl + Zg sin θg
line. The node between the two inductors of the filter is θs = tan−1 (11)
viewed as the inverter terminal and its voltage and current Zl cos θl + Zg cos θg
measurements are utilized as input signals for the control Due to its resistive nature, the impedance Zs ̸ θs cannot
system (depicted in the blue frame). This control system is be simplified with the assumptions sin θs ≈ 1, cos θs ≈ 0.
composed of three main components: (1) a P and Q calcula- However, in distribution grids, where the phase angle differ-
tion block with low-pass filters; (2) the proposed outer-loop ence ϕs between the inverter terminal voltage (E ̸ ϕ1 ) and the
decoupling grid-forming controller; and (3) an inner-loop grid equivalent voltage (Vg ̸ ϕ2 ) is typically small [41], the
voltage and current controller. assumptions sinϕs ≈ ϕs and cosϕs ≈ 1 remain valid. Thus,
The proposed systematic decoupling grid-forming control based on (1) and (2), the P and Q output of the DER in the
is designed to decouple the entire resistive-dominated system, weak distribution grid can be represented as (12) and (13).
encompassing the weak distribution grid, the DER filter, and
EVg ϕs E(E − Vg )
the outlet distribution line. The coupling characteristics of Ps = sin θs + cos θs (12)
the entire grid are continuously monitored and utilized to Zs Zs
EVg ϕs E(E − Vg )
determine the control parameters in real-time. The derivation Qs = − cos θs + sin θs (13)
of the coupling characteristic and the controller design are Zs Zs
discussed in the subsequent sections. where Ps and Qs are active and reactive power of the DER,
To identify the coupling characteristic of the entire inte- respectively, and Vg represents the voltage magnitude of the
grated grid, this paper develops an equivalent circuit model equivalent voltage source of the grid. From (12) and (13),

VOLUME 11, 2024 31

 during online operation as previously discussed in (9). Simul-
taneously, the droop parameters can be determined based
on the power-sharing capability of the DER with consid-
eration of the system stability, which is discussed in the
following section. This design introduced a novel control
technique that effectively address the limitations inherent in
existing grid-forming control schemes. It empowers DERs to
FIGURE 3. Proposed systematic decoupling grid-forming control. autonomously decouple the P and Q of the entire power grid,
achieving independent grid-forming f and V control. This
capability enables a more precise and responsive regulation
the coupling characteristic of the system canbe succinctly of the grid, thereby enhancing stability and efficiency in
sin θs cos θs various operating conditions. Importantly, this design equips
expressed by the matrix R = . To decouple
− cos θs sin θs DERs to dynamically adapt to the time-varying conditions of
the Ps and Qs , equation (11) is multiplied by the inverse of R, the power grid, ensuring continue stable and efficient oper-
sin θs − cos θs ation. The resulting enhanced control mechanism facilitates
i.e., R−1 = , resulting in (14).
cos θs sin θs independent and accurate frequency and voltage regula-
   " EVg ϕs # tion, playing a pivotal role in fortifying the reliability and
sin θs −cosθs
  
Pd Ps Zs resilience of modern distribution systems.
= = E(E−V (14)
Qd cos θs sin θs Qs g)
Zs
III. SMALL-SIGNAL STABILITY ANALYSIS AND
where R−1 is the decoupling matrix, and Pd and Qd represent PARAMETER DESIGN
the decoupled Ps and Qs , respectively. Notable, Pd exhibits The small-signal model of the proposed controller is derived
linearity with respect to ϕs , while Qd displays linearity with in this section. This model is utilized to analyze the
respect to E, aligning with the characteristics of the droop small-signal stability and determine the control parameters.
law, similar to (3) and (4). Consequently, the droop control
can be effectively implemented by harnessing Pd and Qd , A. SMALL-SIGNAL MODEL
representing the decoupled Ps and Qs . As a result, this insight The small-signal model is constructed through the lin-
guides the design of the systematic decoupling grid-forming earization of the power transfer functions in (12) and (13).
controller, as shown in Fig. 3, comprising a decoupler and In addition, it incorporates the models of the proposed sys-
a traditional droop controller. Specifically, the decoupler is tematic decoupling grid-forming controller and the low-pass
designed based on the decoupling matrix R−1 . filters used for power measurement. The resulting representa-
As shown in Fig. 3, the inputs to the proposed systematic tion of the small-signal model is expressed by (18) and (19).
decoupling grid-forming controller are the measured P and
1 
Q, i.e., Pmea and Qmea , which are generated from the P and 1Ps = EVg 1ϕ + 1EVg ϕ sinθs + 1E(2E − Vg )cosθs


Q calculation block. The outputs of this controller are the Zs
generated frequency and voltage references for the inner-loop (18)
voltage and current controller. In Fig. 2, the frequency ref- 1 
1Qs = − EVg 1ϕ + 1EVg ϕ cosθs


erence, i.e., f ∗ , is converted into the angle reference ϕ0 for Zs
the inner loop. The generated voltage reference is viewed


+1E 2E − Vg sinθs (19)
as the d-axis voltage reference u∗od of the voltage control loop.
The q-axis voltage reference, i.e., u∗oq , is set to be 0. where the prefix 1 denotes a small deviation in the variables
The proposed controller in Fig. 3 can be represented math- defined in (12) and (13). Incorporating the decoupler in (14),
ematically as (15) and (16), the small-signal model is transformed into (20) and (21).
 1Ps
 
f ∗ = f0 − mD1 (P0 − Pmea ) − mD3 (Q0 − Qmea ) 1Pd = sin θs − cos θs

(15)
1Qs
E ∗ = E0 − nD2 (P0 − Pmea ) − nD4 (Q0 − Qmea ) (16) 1
EVg 1ϕ + 1EVg ϕ


= (20)
The decoupler parameters (D1 ∼ D4 ) are determined based Zs
on the decoupling matrix R−1 , as shown in (17).  1Ps
 
1
1Qd = cos θs sin θs = 1E(2E − Vg ) (21)

1Qs Zs
D1 = sin θs



 D = cos θ
 Subsequently, by modeling the droop control with
2 s
(17) low-pass filters as a first-order inertia element and substitut-


 D 3 = − cos θs ing 1ϕ with 2π 1f /s, the small-signal model of the studied
D4 = sin θs

system can be expressed as (22) and (23).
2π 1f
 
It shows that the decoupler parameters mainly depend on mwf
1f = − EVg + 1EVg ϕ (22)
the system impedance angle, i.e., θs , which can be determined Zs (wf + s) s

32 VOLUME 11, 2024

He and Yu: Systematic Decoupling Grid-Forming Control

FIGURE 4. Root locus diagrams for the small-signal model. (a) Changing m with fixed n, (b) Changing n with fixed m, (c) Changing
SCR with fixed m and n.

nVg ϕ
1E = − (23)
Zs (wf + s)
Where the symbol wf represents the cut-off frequency of
the low-pass filters. Based on (22) and (23), the characteristic
equation can be derived as a third-order expression, as shown
in (24).

a3 s3 + a2 s2 + a1 s + a0 = 0 (24)

where

a3 = Zs2
a2 = 2Zs2 wf + nZs wf (2E − Vg )
a1 = 2πZs wf mEVg + nZs w2f (2E − Vg ) + Zs2 w2f
a0 = 2π Zs w2f mEVg + 2πw2f mnEVg (2E − Vg ) FIGURE 5. A weak distribution system based on IEEE 14 bus
system.
B. PARAMETER DESIGN CONSIDERING SYSTEM
STABILITY
The stability analysis of the proposed control system, includ- corresponding eigenvalues obtained with the specified m and
ing the determination of its parameters, relies on the charac- n are illustrated in green in Fig. 4 (a) and (b).
teristic equation derived from small-signal stability analysis. Furthermore, to investigate the controller stability in the
This process involves the utilization of root locus diagrams face of a heightened weakness in the distribution grid, the
to analyze the system stability and optimize its parameters. SCR of the weak distribution grid is systematically adjusted.
Specifically, the decoupler parameters D1 ∼ D4 are set based As illustrated in Fig. 4 (c), the eigenvalue plot demonstrates
on the derived coupling characteristic of the system as shown the evolution as the SCR decreases continuously from 3.4 to
in (17). The droop coefficients of the DER can be tuned to 0.8 within the studied system. Notably, throughout this range,
achieve the desired power-sharing capability [18]. To investi- all the eigenvalues consistently reside in the left half plane,
gate the impact on the control stability, the droop coefficients indicating that the proposed controller can maintain stability
undergo continuous adjustments, and the corresponding root across different levels of grid weakness.
locus diagrams are generated, as illustrated in Fig. 4 (a)
and (b). Fig. 4 (a) presents the eigenvalue trajectory (in red) IV. CASE STUDIES
with varying droop coefficient m, with arrows indicating the To validate the proposed control scheme, thorough tests
direction of eigenvalue evolution. Similarly, Fig. 4 (b) shows are carried out in a weakened distribution grid using MAT-
how the eigenvalues change with alterations in the droop LAB/Simulink and OPAL-RT. The weak grid model is
coefficient n. In both cases, the eigenvalues remain within constructed based on the IEEE 14-bus system, depicted in
the left half-plane, indicating that droop parameters can be Fig. 5 [42].
fine-tuned to achieve the desired power sharing while main- The 14-bus weak distribution grid is connected to a swing
taining system stability. In this study, the droop coefficients bus (69kV) via a 13.8 kV/69 kV transformer at Bus 1. Voltage
m and n are set based on the P-f droop of 0.05 per unit levels in this grid include 13.8 kV and 4.6 kV. A utility-scale
(p.u.) and the Q-V droop of 0.1 p.u. after decoupling. The DER with a capacity of 3 MVA is connected to Bus 5 via

VOLUME 11, 2024 33

TABLE 1. Simulation cases and specifications.

FIGURE 6. Case I (A): Simulation of f step response.

the proposed control against that of the existing ZDER -based

decoupling grid-forming control under different coupling
characteristics.

A. CASE I: COMPARISON WITH NON-DECOUPLING

DROOP-BASED GRID-FORMING CONTROL
Case I (A): Simulation of f step response. In this case, the
f and V performance of a DER using the proposed control
strategy is contrasted with a traditional control scheme under
an identical condition, specifically a 2.5% f step change at
0.3 s. Their simulation results are shown in Fig. 6.
In Fig. 6 (a), the f step response of the DER using the non-
decoupling grid-forming controller is presented. The results
reveal an oscillatory and underdamped response, character-
a 480 V/13.8 kV transformer. The short-circuit capacity of ized by a substantial valley excursion and a settling time of
the grid is set as 8.5 MVA, resulting in a low SCR of 2.83. 0.7 s. Importantly, the V fails to return to its nominal value in
For our study, we utilize the OPAL-RT real-time simulation the steady state. This indicates that the V of the DER can be
platform, which includes the hardware simulator OP5600, significantly impacted by f disturbances, potentially leading
real-time-based software RT-LAB, and phasor-domain soft- to inverter tripping. Concurrently, the f exhibits considerable
ware ePHASORSIM to facilitate real-time simulation. The oscillation due to the coupled nature of the traditional control.
detailed configuration process of this simulation setup is However, it eventually stabilizes at the nominal value, due to
elaborated in [43]. the swing bus mechanism. In Fig. 6 (b), the f and V perfor-
The step responses to f and V disturbances in the con- mance of the DER using the proposed control is illustrated.
trol loop are conducted to validate the effectiveness of the Notably, after the f reference step change, the V experiences
independent f and V control by evaluating their mutual a better-damped response with a significantly reduced valley
interferences. The decoupling performance of the proposed excursion compared to Fig. 6 (a). The V swiftly returns to
controller is compared with the traditional non-decoupling the nominal value within 0.3 s. This behavior indicates that
droop-based grid-forming control and the existing ZDER - under the regulation of the proposed controller, the V of the
based decoupling grid-forming control. The comparison DER remains stable and unaffected by the f reference step
cases are outlined in TABLE 1. change. Therefore, the proposed control strategy effectively
To evaluate the stability and dynamic performance result- decouples f and V regulations. When compared to the tra-
ing from the decoupled P and Q, both f and V step responses ditional non-decoupling grid-forming control, the proposed
are studied in this paper. It is noted that the f of the DER control enables the DER to achieve independent V control,
in the simulation cases does not instantaneously follow the leading to a better-damped response and enhanced stability
f step-up reference change applied to the controller. This when the DER encounters a frequency disturbances.
behavior aligns with the nature of a realistic power system, Case I(B): Simulation of V step response. This case
where frequency is a global variable that is dominated by involves a comparison of the f and V performance of a DER
the upstream transmission system. In addition, in Case II, the using the proposed control strategy against the traditional
comparison cases are conducted in a system with a varying control scheme under an identical V reference step change of
R/X ratio, aiming to analyze the decoupling performance of 5% at 0.3 s. Their simulation results are presented in Fig. 7.

34 VOLUME 11, 2024

He and Yu: Systematic Decoupling Grid-Forming Control

FIGURE 9. Case II (B): Simulation of V step response of

ZDER -based grid-forming control with R/X ratio as 1.45:1.
FIGURE 7. Case I (B): Simulation of V step response.

B. CASE II: COMPARISON WITH ZDER -BASED

DECOUPLING GRID-FORMING CONTROL
This section presents a comparative analysis of the f and V
step responses between the proposed control and the existing
ZDER -based decoupling grid-forming control. The latter pri-
marily focuses on decoupling the DER system and neglects
the impact of grid coupling issues. To assess the decoupling
performance of these control schemes in the power grid,
different resistance-to-reactance (R/X) ratios (specifically,
1.45:1 and 1.72:1) of the system impedance are utilized.
FIGURE 8. Case II (A): Simulation of f response of ZDER -based It is noteworthy that the performance of the DER using the
grid-forming control with R/X ratio as 1.45:1. proposed control with an R/X ratio of 1.45:1 was previously
elaborated in Case I.
Case II(A): Simulation of f step response (R/X ratio of
In Fig. 7 (a), the V step response using the non-decoupling the entire system is 1.45:1). This case evaluates the f and
droop-based grid-forming control is depicted. The results V performance of the DER using the existing ZDER -based
demonstrate that the V increases to 1.05 p.u. following the decoupling grid-forming control during the same event as
V reference step change. However, this increase is accompa- in Case I, namely a 2.5% f step change at 0.3 s. Fig. 8
nied by a substantial overshoot and a settling time of 0.5 s. illustrates the response of the DER using the existing ZDER -
Simultaneously, the f is significantly influenced by the V based decoupling grid-forming control following the f step
step change, exhibiting considerable fluctuations that could change.
potentially lead to load or inverter tripping. This observation It is seen in Fig. 8 that the f displays severe oscillatory
indicates that the f performance is adversely affected by V and underdamped behavior, similar to that observed in Fig. 6
disturbances. In contrast, Fig. 7 (b) showcases the f and V (a). The V experiences significant excursions and oscillations
performance of the DER when operated under the proposed with a settling time of 0.7 s, failing to return to the nominal
control strategy. The results reveal that the V of the DER value. This outcome reveals that the ZDER -based decoupling
increases to 1.05 p.u. within approximately 0.2 s without control falls short in effectively decoupling f and V regula-
overshoot, demonstrating a fast and well-damped V response tion, as it does not consider the impact of the power grid
akin to a first-order system. Furthermore, it is seen that after on the DER control. In contrast, Fig. 6 (b) showcases the
the volage reference step change, the f excursion is almost proposed control strategy, demonstrating significantly supe-
entirely mitigated, and the resulting settling time is signifi- rior dynamic performance in comparison to the ZDER -based
cantly reduced compared to Fig. 7 (a). decoupling control.
These observations clearly indicate that the dynamic per- Case II(B): Simulation of V step response. (R/X ratio of
formance of the proposed control is substantially superior the entire system is 1.45:1). This case evaluates the f and
to that of the traditional control scheme. In comparison to V performance of the DER using the existing ZDER -based
traditional control, the proposed control can dynamically decoupling grid-forming control under a 5% V reference step
decouple the entire system based on real-time measurements, change at 0.3 s. As shown in Fig. 9, the V increases to
reducing interference between frequency and voltage con- 1.05 p.u. following the step change with a settling time of
trol. As a result, the proposed control strategy enables the 0.4 s. As observed, the f experiences a significant excursion
DER to achieve independent f control and exhibit a better- and underdamped oscillation. Although the settling time is
damped response, thereby improving overall stability during slightly shorter than the 0.5 s observed in Fig. 7 (a), it is
V disturbances. notably longer than the response seen in Fig. 7 (b) with

VOLUME 11, 2024 35

FIGURE 10. Case II (C): Simulation of f step response with R/X
ratio as 1.72:1. FIGURE 11. Case II (D): Simulation of V step response with R/X
ratio as 1.72:1.

the proposed control. This indicates that the existing ZDER - It exhibits improved performance compared to Case II (B)
based control cannot achieve accurate f control with good in Fig. 9, where the DER utilizes the ZDER -based control
dynamic performance under V disturbances. In contrast, the with a power grid R/X ratio of 1.45:1. This enhancement
DER using the proposed control strategy, as depicted in Fig. 7 is attributed to the adjustment of the R/X ratio of the entire
(b), exhibits improved f and V control performance. system impedance closer to the output impedance of the DER.
It is important to highlight that, much like the traditional This adjustment enables the ZDER -based control to perform
non-decoupling droop-based control, the existing ZDER - better under the V disturbance. However, the f step response
based decoupling control shows inferior f and V control in Fig. 11 (a) indicates that this control method is unable
performance compared to the proposed control strategy under to completely mitigate the impact of the V disturbance on
disturbances. Despite its consideration of the coupling issue, f regulation. Fig. 11 (b) demonstrates the simulation results
the ZDER -based decoupling control solely focuses on the using the proposed control. It is seen that there is a reduction
decoupling of the output impedance of the DER system, in the settling time for f with reduced excursion and the f
namely, ZDER , which encompasses only the impedances of returns to the nominal value more quickly.
the filter and outlet line. Furthermore, this ZDER -based con- Furthermore, the f and V step responses of the proposed
trol approach treats the power grid as an ideal voltage source. control in Fig. 10 (b) and 11 (b), closely resemble those
As a result, the ZDER -based decoupling control approach in Fig. 6 (b) and 7 (b). This indicates that the proposed
proves inadequate in weak grids characterized by substan- control strategy can consistently maintain independent and
tial grid impedance, especially when the R/X ratio of ZDER well-damped f and V regulation in a power grid, even with
markedly differs from that of the grid impedance. varying coupling characteristics.
Case II(C): Simulation of f step response (R/X ratio of
the entire system is 1.72:1). To further explore the decou- C. CASE III: COMPARISON WITH ZDER -BASED
pling performance, the R/X ratio of the system impedance is DECOUPLING GRID-FORMING CONTROL UNDER
changed from 1.45:1 to 1.72:1. This case compares the f and UNBALANCED EVENTS AND HARMONICS
V performance of the DER using the proposed control against In this study, we compare the f and V responses of the pro-
the existing ZDER -based control under a 2.5% f reference step posed control with those of the existing ZDER -based control
change. Fig. 10 (a) shows the f step response of the DER using under unbalanced operational conditions. These conditions
the ZDER -based control. The V has a minor excursion with are typically caused by unbalanced loads or unsymmetri-
a settling time of 0.3 s after the f reference step change is cal faults and are known to induce harmonics and stability
applied. However, it does not return to the nominal value after challenges. We simulate these scenarios by adjusting the
the fluctuation, indicating its susceptibility to the f reference voltage magnitude of Phase A at predetermined intervals,
step change. In contrast, Fig. 10 (b) displays the f and V effectively reproducing the dynamics and harmonic impacts
performance of the DER using the proposed control. It is seen of unbalanced operations. The primary objective is to evaluate
that the proposed control effectively eliminates the impact the effectiveness of the proposed decoupler in enhancing f
of f disturbances on the V response, showcasing a notable and V stability. It is important to note that additional control
improvement over the results observed in Fig. 10 (a). function designs, such as the FRT and sequence control of the
Case II(D): Simulation of V step response (R/X ratio of DER, are beyond the scope of this paper.
the entire system is 1.72:1). This case compares the f and Case III(A): Simulation of unbalanced events (R/X ratio
V performance of the DER using the proposed control with of the entire system is 1.45:1). This case investigates the f
that using the existing ZDER -based control under a 5% V step and V performance under a specific scenario: variation in
change. Fig. 11 (a) illustrates the V step response of the DER the voltage amplitude of Phase A. This variation includes
with the ZDER -based control. modifying the V magnitude of Phase A to 0.8 p.u. at 0.3 s,

36 VOLUME 11, 2024

He and Yu: Systematic Decoupling Grid-Forming Control

FIGURE 12. Case III (A): Simulation of an unbalanced event with FIGURE 14. Case III (C): Simulation of a single-phase fault with
R/X ratio as 1.45:1. R/X ratio as 1.72:1.

better f and V dynamics than Fig. 13 (a), characterized by a

reduced excursion and shorter settling time.
Case III(C): Simulation of single-phase failure (R/X ratio
of the entire system is 1.72:1). This case compares the f and
V performance of a DER using the proposed control with that
using the ZDER -based control during a single-phase fault. The
fault is performed by modifying the V magnitude of Phase
A to 0 from 0.3 s to 0.9 s. Fig. 14 (a) displays the f and V
response of the DER using the ZDER -based control.
As seen in Fig. 14 (a), the V experiences a noticeable
FIGURE 13. Case III (B): Simulation of an unbalanced event with
excursion and does not return to the nominal value. This indi-
R/X ratio as 1.72:1.
cates that the ZDER -based control degrades the steady-state
performance during an unbalanced fault. Meanwhile, the f in
1.2 p.u. at 0.6 s, and 1.0 p.u. at 0.9 s. Fig.12 (a) presents the Fig. 14 (a) has a large deviation and a long settling time of
response of the DER when applying the ZDER -based control. around 0.2 s. Conversely, Fig. 14 (b) shows that using the
It reveals an oscillatory f and V response with significant proposed control scheme has a lesser impact on f dynamics
deviations, which may lead to inverter tripping during unbal- and results in a rapid V recovery to the nominal value, effec-
anced conditions. tively mitigating the impact of the unbalanced fault compared
Fig 12 (b) illustrates the response of the DER when uti- to Fig. 14 (a).
lizing the proposed control method. In contrast to Fig. 12 Case III conclusively demonstrates the superior perfor-
(a), Fig. 12 (b) displays enhanced f and V performance of mance of the proposed control over the ZDER -based control
the DER. It is seen that the V response is more effectively in regulating f and V in weak distribution grids. It delivers
damped, rapidly returning to the nominal value unlike the sce- enhanced dynamic responses, steady-state performance, and
nario depicted in Fig. 12 (a). The f shows fewer oscillations superior system stabilization under unbalanced conditions.
compared to the ZDER -based control, indicating that the pro-
posed control scheme has better stabilization capability than V. CONCLUSION
the existing ZDER -based control schemes during unbalanced This paper proposes a robust and unified systematic decou-
events. pling grid-forming control scheme for the DER operating in
Case III(B): Simulation of unbalanced events (R/X ratio weak distribution grids. The primary objective is to signifi-
of the entire system is 1.72:1). This case examines the f and cantly enhance dynamic response and stability performance.
V performance of the DER using the proposed control in In contrast to existing DER grid-forming control methods
comparison to the ZDER -based control under the same event that often treat distribution grids as a simple Thevenin equiv-
as in Case III (A) but with the R/X ratio of the entire system alent circuit, the proposed controller is designed from a
changed to 1.72:1. As observed, Fig. 13 (a) shows improved f system-wide perspective, taking into consideration the cou-
and V performance under the ZDER -based control compared pling characteristics of the entire power grid. The simulation
to Fig. 12 (a) in Case III (A), where the power grid R/X ratio results demonstrate that the developed control strategy can
is 1.45:1. As discussed in Case II (D), this f and V response enable DERs to decouple P and Q and regulate f and V
improvement is attributed to the closer matching of the R/X independently. In conclusion, this proposed control strategy
ratio of the entire system impedance to the output impedance offers several advantages: (1) easy commercialization and
of the DER. Importantly, Fig. 13 (b) shows the f and V per- market promotion due to its simple and robust circuit design;
formance of the DER using the proposed control. It exhibits (2) provision of a unified grid-forming controller for DERs,

VOLUME 11, 2024 37

TABLE 2. Parameters and system settings. [12] Y. Sun, S. Bahrami, V. W. S. Wong, and L. Lampe, ‘‘Chance-constrained
frequency regulation with energy storage systems in distribution net-
works,’’ IEEE Trans. Smart Grid, vol. 11, no. 1, pp. 215–228, Jan. 2020.
[13] L. Sang, Y. Xu, W. Wu, and H. Long, ‘‘Online voltage regulation of active
distribution networks: A deep neural encoding-decoding approach,’’ IEEE
Trans. Power Syst., early access, Sep. 2023.
[14] W. Alabri and D. Jayaweera, ‘‘Voltage regulation in unbalanced power
distribution systems with residential PV systems,’’ Int. J. Electr. Power
Energy Syst., vol. 131, Oct. 2021, Art. no. 107036.
[15] P. Kundur, Power System Stability and Control, 1st ed. New York, NY,
USA: McGraw-Hill, 1994.
[16] L. He, C.-C. Liu, A. Pitto, and D. Cirio, ‘‘Distance protection of AC grid
with HVDC-connected offshore wind generators,’’ IEEE Trans. Power
simplifying various operating modes outlined in IEEE Stan- Del., vol. 29, no. 2, pp. 493–501, Apr. 2014.
dard 1547-2018, particularly in addressing coupled P and Q [17] D. Pattabiraman, R. H. Lasseter., and T. M. Jahns, ‘‘Comparison of grid
issues; and (3) elimination of the need for communication following and grid forming control for a high inverter penetration power
system,’’ in Proc. IEEE Power Energy Soc. Gen. Meeting (PESGM),
links between DERs, reducing costs and minimizing the risk Portland, OR, USA, Aug. 2018, pp. 1–5.
of cyber-attacks. [18] W. Du, Q. Jiang, M. J. Erickson, and R. H. Lasseter, ‘‘Voltage-source
It is important to note that this paper specifically focuses on control of PV inverter in a CERTS microgrid,’’ IEEE Trans. Power Del.,
vol. 29, no. 4, pp. 1726–1734, Aug. 2014.
the systematic decoupling control design of the DER. Coor-
[19] Q.-C. Zhong and G. Weiss, ‘‘Synchronverters: Inverters that mimic
dination among multiple DERs in a grid and the optimization synchronous generators,’’ IEEE Trans. Ind. Electron., vol. 58, no. 4,
of the operation are beyond the scope of this work and will be pp. 1259–1267, Apr. 2011.
explored in future research. Further insight and advancements [20] M. Sinha, F. Dörfler, B. B. Johnson, and S. V. Dhople, ‘‘Virtual oscillator
control subsumes droop control,’’ in Proc. Amer. Control Conf. (ACC),
in this domain will be presented in forthcoming papers. Jul. 2015, pp. 2353–2358.
[21] D. Çelik and M. E. Meral, ‘‘A coordinated virtual impedance control
APPENDIX scheme for three phase four leg inverters of electric vehicle to grid (V2G),’’
Energy, vol. 246, May 2022, Art. no. 123354.
See Table 2. [22] J. He, L. Du, B. Liang, Y. Li, and C. Wang, ‘‘A coupled virtual impedance
for parallel AC/DC converter based power electronics system,’’ IEEE
ACKNOWLEDGMENT Trans. Smart Grid, vol. 10, no. 3, pp. 3387–3400, May 2019.
[23] B. Wei, Y. Gui, S. Trujillo, J. M. Guerrero, J. C. Vásquez, and
The authors would like to thank Shuaiang Rong for editing a A. Marzàbal, ‘‘Distributed average integral secondary control for modular
portion of the manuscript. UPS systems-based microgrids,’’ IEEE Trans. Power Electron., vol. 34,
no. 7, pp. 6922–6936, Jul. 2019.
[24] M. Pham and H. Lee, ‘‘Coordinated virtual resistance and capacitance
REFERENCES control scheme for accurate reactive power sharing and selective harmonic
[1] Act (CLEAN Future Act), Senate House Representatives United States compensation in islanded microgrid,’’ IET Gener., Transmiss. Distrib.,
Amer., Washington, DC, USA, Mar. 2021. vol. 14, no. 22, pp. 5104–5113, Sep. 2020.
[2] United States Distributed Energy Resources Outlook: DER Installations [25] Y. Han, P. Shen, X. Zhao, and J. M. Guerrero, ‘‘Control strategies for
and Forecasts 2016-2025E, Wood Mackenzie, Louisville, KY, USA, islanded microgrid using enhanced hierarchical control structure with
Jun. 2020. multiple current-loop damping schemes,’’ IEEE Trans. Smart Grid, vol. 8,
[3] B. Kroposki et al., ‘‘Achieving a 100% renewable grid: Operating electric no. 3, pp. 1139–1153, May 2017.
power systems with extremely high levels of variable renewable energy,’’ [26] M.-D. Pham and H.-H. Lee, ‘‘Effective coordinated virtual impedance
IEEE Power Energy Mag., vol. 15, no. 2, pp. 61–73, Mar. 2017. control for accurate power sharing in islanded microgrid,’’ IEEE Trans.
[4] E. Rehman, M. G. Miller, J. Schmall, S. H. Huang, and J. Billo, ‘‘Stability Ind. Electron., vol. 68, no. 3, pp. 2279–2288, Mar. 2021.
assessment of high penetration of inverter-based generation in the ERCOT [27] M. Karimi-Ghartemani, S. A. Khajehoddin, P. Piya, and M. Ebrahimi,
grid,’’ in Proc. IEEE Power Energy Soc. Gen. Meeting (PESGM), Atlanta, ‘‘Universal controller for three-phase inverters in a microgrid,’’ IEEE
GA, USA, Aug. 2019, pp. 1–5. J. Emerg. Sel. Topics Power Electron., vol. 4, no. 4, pp. 1342–1353, Dec.
[5] Black System South Australia 28 September 2016-Final Report, Austral. 2016.
Energy Market Operator, Sydney, NSW, Australia, Mar. 2017. [28] I. Sadeghkhani, M. E. H. Golshan, A. Mehrizi-Sani, and J. M. Guerrero,
[6] 9 August 2019 Power Outage Report, Office Gas Electr. Markets (Ofgem), ‘‘Low-voltage ride-through of a droop-based three-phase four-wire grid-
London, U.K., Jan. 2020. connected microgrid,’’ IET Gener. Transm. Distrib., vol. 12, no. 8,
[7] X. Chen, Y. Zhang, S. Wang, J. Chen, and C. Gong, ‘‘Impedance-phased pp. 1906–1914, Apr. 2018.
dynamic control method for grid-connected inverters in a weak grid,’’ IEEE [29] B. Pawar, E. I. Batzelis, S. Chakrabarti, and B. C. Pal, ‘‘Grid-forming
Trans. Power Electron., vol. 32, no. 1, pp. 274–283, Jan. 2017. control for solar PV systems with power reserves,’’ IEEE Trans. Sustain.
[8] I. Hussain, R. K. Agarwal, and B. Singh, ‘‘MLP control algorithm for Energy, vol. 12, no. 4, pp. 1947–1959, Oct. 2021.
adaptable dual-mode single-stage solar PV system tied to three-phase [30] S. Yu, S. Rong, and L. He, ‘‘Performance of grid-forming control of grid-
voltage-weak distribution grid,’’ IEEE Trans. Ind. Informat., vol. 14, no. 6, edge DERs in distribution grids,’’ in Proc. IEEE Power Energy Soc. Gen.
pp. 2530–2538, Jun. 2018. Meeting (PESGM), Denver, CO, USA, Jul. 2022, pp. 1–5.
[9] IEEE Standard for Interconnection and Interoperability of Distributed [31] A. Moawwad, V. Khadkikar, and J. L. Kirtley, ‘‘A new P-Q-V droop control
Energy Resources With Associated Electric Power Systems Interfaces, method for an interline photovoltaic (I-PV) power system,’’ IEEE Trans.
Standard 1547-2018, Revision of IEEE Standard 1547-2003, 2018, Power Del., vol. 28, no. 2, pp. 658–668, Apr. 2013.
pp. 1–138. [32] W. Yao, M. Chen, J. Matas, J. M. Guerrero, and Z.-M. Qian, ‘‘Design and
[10] S. C. Ross, G. Vuylsteke, and J. L. Mathieu, ‘‘Effects of load-based fre- analysis of the droop control method for parallel inverters considering the
quency regulation on distribution network operation,’’ IEEE Trans. Power impact of the complex impedance on the power sharing,’’ IEEE Trans. Ind.
Syst., vol. 34, no. 2, pp. 1569–1578, Mar. 2019. Electron., vol. 58, no. 2, pp. 576–588, Feb. 2011.
[11] J. Rocabert, A. Luna, F. Blaabjerg, and P. Rodríguez, ‘‘Control of power [33] Q.-C. Zhong and Y. Zeng, ‘‘Universal droop control of inverters with
converters in AC microgrids,’’ IEEE Trans. Power Electron., vol. 27, different types of output impedance,’’ IEEE Access, vol. 4, pp. 702–712,
no. 11, pp. 4734–4749, Nov. 2012. 2016.

38 VOLUME 11, 2024

He and Yu: Systematic Decoupling Grid-Forming Control

[34] Z. Peng et al., ‘‘Droop control strategy incorporating coupling compensa- LINA HE (Senior Member, IEEE) received the
tion and virtual impedance for microgrid application,’’ IEEE Trans. Energy B.S. and M.S. degrees from the Huazhong Univer-
Convers., vol. 34, no. 1, pp. 277–291, Mar. 2019. sity of Science and Technology, China, in 2007 and
[35] Y. Li and Y. W. Li, ‘‘Power management of inverter interfaced autonomous 2009, respectively, and the Ph.D. degree in electri-
microgrid based on virtual frequency-voltage frame,’’ IEEE Trans. Smart cal engineering from University College Dublin,
Grid, vol. 2, no. 1, pp. 30–40, Mar. 2011. Ireland, in 2014. From 2014 to 2017, she was
[36] J. Sun, ‘‘Impedance-based stability criterion for grid-connected inverters,’’ the Project Manager and a Senior Consultant at
IEEE Trans. Power Electron., vol. 26, no. 11, pp. 3075–3078, Nov. 2011. Siemens Headquarters, Germany, and Siemens,
[37] K. V. Kkuni, M. Nuhic, and G. Yang, ‘‘Power system stability impact
USA. She is currently an Assistant Professor
assessment for the current limits of grid supporting voltage-source convert-
with the Department of Electrical and Computer
ers,’’ in Proc. IEEE Power Energy Soc. Gen. Meeting (PESGM), Jul. 2021,
pp. 1–5. Engineering, University of Illinois at Chicago, Chicago, IL, USA. Her
[38] E. Zhao, Y. Han, X. Lin, P. Yang, F. Blaabjerg, and A. S. Zalhaf, research interests include modeling, control, and protection of power
‘‘Impedance characteristics investigation and oscillation stability analysis electronics-based power systems, renewable energy integration, HVdc con-
for two-stage PV inverter under weak grid condition,’’ Electr. Power Syst. trol and operation, and wide-area protection and cybersecurity.
Res., vol. 209, Aug. 2022, Art. no. 108053.
[39] S. Bennai, A. B. Abdelghani, I. Slama-Belkhodja, and M. Khalfoun,
‘‘Sensitivity analysis to grid line impedance for grid characterization and
stability assessment investigations,’’ in Proc. IEEE Int. Conf. Electr. Sci.
Technol. Maghreb (CISTEM), vol. 4, Tunis, Tunisia, Oct. 2022, pp. 1–6.
[40] IEEE Guide for Planning DC Links Terminating At AC Locations Hav-
ing Low Short-circuit Capacities, IEEE Standard 1204-1997, Jan. 1997,
pp. 1–216. SHIWEN YU (Graduate Student Member, IEEE)
[41] J. C. Vasquez, J. M. Guerrero, A. Luna, P. Rodriguez, and R. Teodorescu, received the B.S. and M.S. degrees in electrical
‘‘Adaptive droop control applied to voltage-source inverters operating in engineering from the Wuhan University of Tech-
grid-connected and islanded modes,’’ IEEE Trans. Ind. Electron., vol. 56, nology, in 2017 and 2020, respectively. He is
no. 10, pp. 4088–4096, Oct. 2009.
currently pursuing the Ph.D. degree in electri-
[42] IEEE 14 Bus System. Accessed: Feb. 1, 2022. [Online]. Available:
cal engineering with the University of Illinois at
http://labs.ece.uw.edu/pstca/pf14/pg_tca14bus.htm
[43] S. S. Noureen, V. Roy, and S. B. Bayne, ‘‘An overall study of a real-time
Chicago, Chicago, IL, USA. His research interests
simulator and application of RT-LAB using MATLAB simpowersystems,’’ include grid-forming control, power systems sta-
in Proc. IEEE Green Energy Smart Syst. Conf. (IGESSC), Long Beach, CA, bility, and machine learning.
USA, Nov. 2017, pp. 1–5.

VOLUME 11, 2024 39