Received 1 May 2023; revised 22 May 2023; accepted 30 May 2023.

Date of publication 7 June 2023;

date of current version 5 July 2023. The review of this paper was arranged by Associate Editor Min Chen.
Digital Object Identifier 10.1109/OJPEL.2023.3283782

Double-Synchronous-Reference-Frame-Based
Power-Synchronized PLL-Less Grid-Following
Inverters for Unbalanced Grid Faults
NABIL MOHAMMED (Member, IEEE), WEIHUA ZHOU (Member, IEEE),
AND BEHROOZ BAHRANI (Senior Member, IEEE)
Department of Electrical and Computer Systems Engineering, Monash University, Clayton, VIC 3800, Australia
CORRESPONDING AUTHOR: NABIL MOHAMMED (email: [Link]@[Link])
This work was supported by the Monash Grid Innovation Hub and the Australian Renewable Energy Agency (ARENA) under the Advancing Renewable Program
under Grant 2020/ARP007.

ABSTRACT This article proposes a double-synchronous-reference-frame-based power-synchronized grid-

following inverter (DS-PSGFLI). In contrast to existing types of phase-locked-loop (PLL)-based grid-
following inverters (GFLIs) and PLL-less power-synchronized GFLIs (PSGFLIs), the suggested PLL-less
DS-PSGFLI operates seamlessly during unbalanced grid faults in both strong and weak grid connections,
while maintaining stability and avoiding oscillations. Furthermore, unlike previous control strategies for
PSGFLIs that regulate terminal output power, the proposed method regulates the power at the point
of common coupling (PCC). This approach aligns with common PCC control strategies in GFLIs and
grid-forming inverters, thereby avoiding the impacts of voltage drop across the inverter output filter and
high voltage harmonics at the inverter terminal. In the proposed DS-PSGFLI, the positive and negative
sequences of the PCC voltage and current measurements are first extracted. Then, an outer power controller
is employed to regulate the positive active and reactive power at the PCC to the desired references. The
outer power controller generates both the positive internal angle required for grid synchronization and
the positive current reference for the inner current of the positive sequence loop. When unbalanced grid
faults occur, the negative sequence loop adjusts its output signals to mitigate the resulting oscillations at
twice the fundamental frequency. The effectiveness of the proposed control strategy is validated through
simulations and experimental tests, demonstrating its reliable performance across a wide range of test
scenarios.

INDEX TERMS Phase locked loops, positive and negative sequence control, power-synchronized control,
PLL-less inverters, short-circuit-ratio, stability, unbalanced grid voltage, weak grid.

I. INTRODUCTION However, with the reduction in fossil-fuel-based power sta-

The worldwide drive towards clean and affordable energy, tions, the overall inertia of the grid has decreased, thereby
as well as reducing CO2 emissions, has led to a significant giving rise to several operational challenges for GFLIs. These
increase in the integration of renewable energy resources challenges include phase-locked loop (PLL) instability in
into existing power systems in recent decades [1]. Fossil- weak grids and an inability to continue functioning during
fuel-based synchronous power stations are being replaced by abnormal conditions such as unbalanced faults [3], [4].
inverter-based resources (IBRs), with grid-connected inverters For the first challenge, PLL-related instability, it is caused
playing a crucial role in integrating renewable resources such as the operation of traditional control techniques of GFLIs in
as solar photovoltaic and wind [2]. weak grids heavily relies on the PLL used for synchronization
According to current industry trends, the majority of IBRs with AC grids [5]. Over the past two decades, various aspects
currently in operation are grid-following inverters (GFLIs). of PLL dynamics, such as fast tracking and performance under

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see [Link]
474 VOLUME 4, 2023
balanced [1], unbalanced [6], and distorted grid voltages [7] This phenomenon results in unbalanced current injection,
have been extensively studied. Recently, the stability issues of thereby posing challenges for inverter control. To address the
PLL synchronization in GFLIs have been a growing concern. challenges posed by unbalanced grid voltages, various control
Studies suggest that low-frequency instability can occur when techniques have been developed for GFLIs with the aim of
GFLIs are connected to weak AC grids due to their PLLs [8], enabling the injection of balanced current, active power, or
[9]. Small signal analysis indicates that the PLL reshapes the reactive power [26], [27]. Among these techniques, the widely
q − q channel output impedance of GFLIs as a negative re- adopted approach is the positive and negative sequence con-
sistor within its bandwidth [10]. Hence, high PLL bandwidths trol (PNSC) strategy, which facilities independent control of
can result in instability [11]. the positive and negative sequences of the current. By im-
Recent studies have explored various control solutions to plementing this strategy, conventional GFLIs can effectively
improve the stability of PLL-based GFLIs connected to weak operate under unbalanced grid voltages and maintain the in-
AC grids. In [12], an improved vector control strategy is jection of balanced current into the grid [5], [6].
proposed to alleviate the associated negative impacts of the Despite previous efforts to address the challenges of PLL
PLL. Adaptive impedance-based control suggests shaping the instability in weak grids and unbalanced grid voltages, these
output impedance of GFLIs by adjusting the control param- issues remain unresolved. On the one hand, the proposed
eters of the PLL and voltage feed-forward [13]. In [14], an PLL-less GFLIs in [22], [23] demonstrate stable and seam-
adaptive impedance-phased compensation control approach less operation in both strong and weak grids for ideal grid
is proposed to prevent interaction between GFLIs and the conditions with balanced voltage sources. However, when
grid. In [15], a supplementary frequency-based control loop confronted with asymmetrical conditions, these control strate-
is introduced to damp PLL-related oscillations. In [16], a gies fail to deliver balanced current to the grid, resulting in
feed-back linearization controller that eliminates the nonlinear oscillations in the dq reference frame at twice the grid fun-
dynamics of a PLL used in a GFLI is proposed to expand the damental frequency. Moreover, severe unbalanced events can
PLL domain of attraction to the whole plane that is limited to cause the inverter to become unstable. On the other hand,
a small region around the equilibrium point in a conventional although the PNSC strategy ensures stability and performance
PLL. Other methods for improving the synchronization of of the PLL-based GFLIs in strong grids, it remains suscepti-
GFLIs include reducing the PLL bandwidth [10] and feed- ble to PLL-related instability in weak grids. Furthermore, the
forward-control-based methods [17]. PNSC strategy has thus far been exclusively implemented for
To further improve the stability of GFLIs, PLL-less con- the PLL-based GFLIs.
trol strategies have been proposed to eliminate the PLL This article proposes the double-synchronous-reference-
from the control structure. In [18], a direct power control frame-based power-synchronized grid-following inverter
(DPC) approach is presented for grid synchronization. Im- (DS-PSGFLI) for smooth operation of IBRs during unbal-
proved DPC strategies, such as model predictive control-DPC anced grid faults in both strong and weak grid connections.
(MPC-DPC) [19], improved direct power control (GVM- Unlike PLL-based GFLIs, the DS-PSGFLI generates the
DPC) [20], and voltage-modulated DPC (VMDPC) [21], have phase angle required for grid synchronization internally, using
also been developed. More recently, a power-synchronized the outer power controller of the positive sequence, and thus
GFLIs (PSGFLIs) approach is proposed in [22] to reli- does not require a PLL. In the proposed control structure, the
ably operate GFLIs in strong, weak, and even ultra-weak references for the positive current loop are set using the gener-
grids. Additionally, a linear parameter-varying loop-shaping ated current by the outer power controller, while enforcing the
controller for power-synchronized GFLI (LPV-PSGFLI) is references for the negative sequence to zero. The key benefits
proposed in [23]. Despite the effectiveness of the proposed of the proposed DS-PSGFLI include:
control strategies in [22] and [23], the desired performance is 1) independent control of positive and negative sequences,
only guaranteed under normal grid connections. Furthermore, enabling smooth operation in unbalanced grid condi-
both control strategies require inverter terminal measurements tions without being prone to oscillations or instability.
rather than point of common coupling (PCC), resulting in a 2) PLL-less control structure, avoiding PLL-related insta-
significant voltage drop across the inverter output filter, par- bility and operating seamlessly in both strong and very
ticularly in high-power rating inverters. weak grids.
For the second challenge, in order to operate GFLIs dur- 3) power regulation at the PCC, instead of the terminal, is
ing unbalanced faults, grid standards require these devices similar to common PCC control strategies in GFLIs and
to remain connected to the grid for a predetermined period GFMIs, thus avoiding impacts of voltage drops across
of time, which varies depending on the degree of voltage the output filter on the regulated PCC power and elimi-
sag [5]. During this period, GFLIs are required to inject a nating the use of terminal measurements, which include
specific amount of active and reactive power to the grid, in high levels of switching harmonics.
accordance with low voltage ride-through requirements [24], 4) facilitation reliable integration of extensive renewable
[25]. Nevertheless, the presence of unbalanced grid voltages energy resources into existing power systems, while
cause an uncontrollable oscillation at double the grid funda- concurrently addressing vulnerabilities associated with
mental frequency in the synchronous (dq) reference frame. small-signal stability concerns.

VOLUME 4, 2023 475

MOHAMMED ET AL.: DOUBLE-SYNCHRONOUS-REFERENCE-FRAME-BASED POWER-SYNCHRONIZED PLL-LESS GRID-FOLLOWING INVERTERS

In both the PSGFLI and LPV-PSGFLI, the dq reference

frame is aligned with the inverter output current instead of
voltage, leading to a q-component of the current reference
iq−ref = 0 A. Let vt,d and vt,q represent the dq-components
of the inverter terminal voltage, while id and iq represent the
dq-components of the inverter output current. Subsequently,
the terminal voltage and current of the inverter in the dq
reference frame are expressed as

vt,dq = vt,d + jvt,q ,
(1)
idq = id + jiq = id + j0.
Under normal conditions, the terminal active power, Pt , and
reactive power, Qt , of the inverter are calculated using the
instantaneous power theory as follows
FIGURE 1. The conventional PLL-less power-synchronized GFLI [22]. 
Pt = 23 (vt,d id + vt,q iq ),
(2)
Qt = 23 (vt,q id − vt,d iq ).
The structure of the rest of this article is as fol- During unbalanced fault events, the injected Pt and Qt are no
lows: The modeling and limitations of the conventional longer constant due to the negative voltage sequence produced
power-synchronized control strategies for PLL-less GFLIs by the unbalance/fault events. Both Pt and Qt consist of the
are discussed in Section II. The proposed control approach, average values of the inverter terminal active power (P0 ) and
including the stability analysis in the frequency domain based reactive power (Q0 ), as well as additional oscillatory compo-
on the Generalized Nyquist Criterion (GNC), is described in nents (Ps2 , Pc2 , Qs2 , Qc2 ) at twice the fundamental frequency,
Section III. The simulation and experimental results to val- which is 100 Hz for 50 Hz power systems, as shown in (3).
idate the effectiveness of the proposed control scheme are Therefore, Pt and Qt are calculated based on the instantaneous
presented in Section IV and Section V. Finally, the conclu- active and reactive power during unbalanced grid conditions,
sions are drawn in Section VI. as described in [5]

Pt = P0 + Pc2 (2ωt ) + Ps2 (2ωt ),
II. CONVENTIONAL PSGFLI (3)
Qt = Q0 + Qc2 (2ωt ) + Qs2 (2ωt ).
Fig. 1 displays the block diagram of a conventional PSGFLI.
The inverter is powered by a DC source and interfacted with Equation (3) illustrates that asymmetrical faults, commonly
the grid via an inductor filter (Lf ). The grid is represented encountered in power systems due to various factors such
by the equivalent Thévenin model that includes an equivalent as single/double line-to-ground faults and asymmetrical grid
voltage source (vg ) and an equivalent grid impedance (Zg ) voltage, induce uncontrolled oscillations in the dq reference
modeled as a resistive-inductive (Rg , Lg ) circuit. frame of the inverter control loop at twice the grid frequency.
This PSGFLI control strategy, proposed in [22], aims to en- As stated in [5], [28], [29], this can lead to a degradation of
able smooth operation of GFLIs in both weak and strong grids. the performance of the IBRs as the injected active and reactive
Unlike conventional GFLIs, which often face challenges in power to the grid will not remain constant as desired.
weak grid conditions, the PSGFLI is a PLL-less control tech- Moreover, in contrast to the commonly used PCC control
nique that permits full control of active and reactive power strategies, both PSGFLI and LPV-PSGFLI regulate the output
without being prone to instability. Furthermore, instead of power at the inverter terminal instead of the PCC. This leads
the optimization-based design methodology used in the PS- to drawbacks with these control methods. Firstly, the termi-
GFLI [22], the LPV-PSGFLI, proposed in [23], adaptively nal voltage vt is heavily contaminated by inverter switching
tunes the power control loop parameters M11 , M21 , M12 , and harmonics, leading to the need for heavy filtering of the cal-
M22 to maintain the desired closed-loop performance (i.e., culated active/reactive power using a low-pass filter. Secondly,
controller bandwidth) across all operating points. controlling the terminal voltage results in a voltage drop at
Fig. 1 demonstrates that both the PSGFLI and LPV- the PCC due to the impedance of the inverter output filter,
PSGFLI utilize power synchronization concept, where the especially in high-power rated inverters. Finally, considering
power control loop estimates the grid angular frequency (ω) that grid code requirements, including power quality assess-
and d-component of the current reference (id−ref ). Moreover, ments, are typically measured at the PCC rather than at the
the inner current control loop structure in the PSGFLI closely terminal, it is imperative to employ a control strategy that
resembles the widely employed vector-current control found regulates the output power at the PCC in a manner consistent
in conventional GFLIs. However, a notable distinction is the with commonly utilized PCC control strategies for GFLIs.
absence of voltage feed-forward in the PSGFLI and LPV- To further illustrate the limitations of conventional PS-
PSGFLI control scheme. GFLIs shown in Fig. 1, a case study with the simulation

476 VOLUME 4, 2023

TABLE 1. Parameters Used for the Simulation Study

parameters listed in Table 1 is provided in Fig. 2. The inverter

is connected to a weak grid with a short-circuit-ratio (SCR)
of 2.5 and Xg /Rg of 3. It is worth noting that this article
uses the SCR index to categorize the grid strength. The SCR
is represented as the quotient of the short-circuit capacity at
the PCC and the rated capacity of the inverter connected to
the grid, Sg /Sn . Typically, the strength of the power grid is
evaluated as strong when SCR > 3, weak for 2 ≤SCR≤ 3,
and very weak for SCR < 2 [30].
Fig. 2 depicts the simulation results obtained using MAT-
LAB/Simulink, illustrating the connection of the conventional
PSGFLI to the weak grid and the occurrence of an unbal-
anced grid voltage sag at 0.2 s. It can be observed that even
though the PSGFLI tracks the active power set-point after the
reference change, the active and reactive power exhibit oscil-
lations at 100 Hz after the unbalanced voltage sag at 0.2 s [5],
[26]. These oscillations degrade the power quality and may
compromise system reliability as they can result in unwanted
reverse active power flow to the DC link. To address the limi- FIGURE 2. The performance of the conventional PSGFLI in a weak grid
tations arising from unbalanced grid voltages and to leverage during an unbalanced grid voltage event at 0.2 s: (a) grid voltage,
(b) terminal current, (c) dq-components of the filtered terminal voltages,
the benefits offered by the PSGFLI and LPV-PSGFLI strate- (d) dq-components of terminal current, and (e) terminal active and reactive
gies, which encompass seamless operation in both strong and power.
weak grids, this article proposed the PLL-less DS-PSGFLI
that copes with asymmetrical faults. Detailed explanation and
verification of the proposed strategy are presented in the sub-
positive angle (θ + ) and negative angle (θ − ) as
sequent sections.
⎧
⎪
⎪v + = vd+ + jvq+ ,
⎪ dq
⎪
⎨v − = v − + jv − ,
III. PROPOSED DS-PSGFLI dq d q
+ + +,
(4)
A. CONTROL STRUCTURE ⎪idq
⎪ = i + ji
⎪
⎪ d q
The control architecture of the proposed DS-PSGFLI is illus- ⎩i− = i− + ji− .
dq d q
trated in Fig. 3. As seen in the figure, the control strategy is
implemented in the dq reference frame. The first step in im-
plementing the strategy is to extract the positive and negative + ) and positive reac-
Afterwards, the positive active power (Ppcc
components of the PCC voltage (vpcc + , v − ) and current (i+ , +
pcc pcc tive power (Qpcc ) at the PCC are calculated using the positive
−
ipcc ), which are then converted to the dq reference frame using components of the dq voltage and current measurements, as

VOLUME 4, 2023 477

MOHAMMED ET AL.: DOUBLE-SYNCHRONOUS-REFERENCE-FRAME-BASED POWER-SYNCHRONIZED PLL-LESS GRID-FOLLOWING INVERTERS

FIGURE 3. The proposed PLL-less DS-PSGFLI for unbalanced grid faults.

− −
follows (id−ref = 0 and iq−ref = 0). Furthermore, θ − for the negative

+ = 3 (v + i+ + v + i+ ),
Ppcc loop is calculated as θ − = −θ + .
2 d d q q
(5)
+ = 3 (v + i+ − v + i+ ).
Qpcc 2 q d d q
B. ASYMMETRICAL FAULTS
Then, based on the calculated power, the outer power loop During normal operation of the proposed DS-PSGFLI, only
generates the reference current of the positive d-component +
the positive sequence voltage components (vdq ) exist in the
+
(id−ref ) and the positive angle (θ + ) as given below inverter control loop, while the negative sequence voltage
⎧       components are zero (vd− = 0, vq− = 0). However, during
+ + + +
⎪
⎨ ω K11 (s) K12 (s) Pref − Ppcc ωg-ref asymmetrical faults, the negative sequence voltage compo-
= + ,
+
id−ref K21 (s) K22 (s) Qref + +
− Qpcc 0 nents will appear (vd− = 0, vq− = 0), and the amplitude of
⎪
⎩ + +
θ = ωs , these components will directly correspond to the severity of
(6) the asymmetrical faults.
where K11 (s) to K22 (s) are the transfer functions of the The proposed controller during normal and fault operations
+ will only inject a balanced current. As a result, in the event of
power controllers, and ωg-ref is the angular frequency refer-
ence of the grid, equal to 2π 50 rad/s. K11 (s) to K22 (s) can faults, id+ is increased to maintain the desired reference power
be designed/tuned using either the optimization-based design value. Additionally, the controller ensures that iq+ , id− , and iq−
methodology or linear-parameter varying control. In this ar- remain constant and equal to zero. It is important to note that
ticle, the optimization-based design methodology is adopted, the proposed controller is always aligned with iq+ component,
and further details can be found in [22]. which should always be zero.
The inner loop of the proposed DS-PSGFLI consists of two
parallel inner current controllers. The upper current controller C. PCC POWER REGULATION
is responsible for controlling the positive sequence, receiving The control strategy proposed in this article distinguishes
+
its reference id−ref from the outer power loop and setting itself from the approach employed in PSGFLI and LPV-
+
the q-component to zero (iq−ref = 0). On the other hand, PSGFLI, where the terminal output power is regulated. In-
the lower current controller manages the negative sequence, stead, the proposed strategy regulates the inverter power at the
with its current references for both dq-components set to zero PCC, as illustrated in Fig. 3. This control approach, similar to

478 VOLUME 4, 2023

 (strong, weak, and very weak) as the Nyquist plots do not en-
circle the critical point (−1,0). This implies that the proposed
control strategy is expected to operate stably and be robust to
grid impedance variations owing to its wider phase margins,
as verified in the following sections through simulations and
experimental results.

IV. SIMULATION RESULTS

The proposed DS-PSGFLI control strategy, illustrated in
Fig. 3 with the corresponding parameters listed in Table 1,
is evaluated through simulation utilizing MATLAB/Simulink
and PLECS Blockset Packages. The simulation encompasses
various scenarios, including changes in active and reac-
tive power references, as well as unbalanced voltages under
strong, weak, and very weak grids conditions. Moreover, the
performance of the proposed controller is compared with the
established PNSC strategy [5], [26].

A. PERFORMANCE IN STRONG GRID, WEAK GRID, AND

VERY WEAK GRID
Fig. 5 presents the simulation results of the proposed control
strategy in the different grid connections. The inverter is con-
nected to a strong grid with SCR = 15 and Xg /Rg = 7 in the
FIGURE 4. Impedance-based stability analysis of the proposed DS-PSGFLI
in the q-axis for three grid connections (strong grid Zg1 , weak grid Zg2 , and
first case, to a weak grid with SCR = 2 and Xg /Rg = 7 in
very weak grid Zg3 ): (a) grid impedances and inverter output impedance in the second case, and to a very weak and resistive grid with
the q-axis, and (b) stability assessment in the q-axis. SCR = 1 and Xg /Rg = 1 in the third case. The simulation
starts with the inverter operating in a steady state, injecting 2
MW of active power and 1.5 MVar of reactive power. At 0.1 s,
commonly used PCC control strategies, offers several advan- the active power set-point increases to 4 MW. At 0.2 s, a 50%
tages. The proposed method effectively regulates PCC power voltage sag in phase B of the grid voltage is applied. Finally,
to meet desired references by avoiding the impact of voltage at 0.3 s, the reactive power set-point decreases to 0 MVar. The
drops across the output filter on the regulated inverter output simulation results in the strong, weak, and very weak grids
power. It ensures accurate regulation of the PCC voltage and conditions are shown in the left, middle, and right columns of
current. Furthermore, stability is enhanced by employing the Fig. 5, respectively.
PCC voltage as a feed-forward signal within the inner current Fig. 5(a)–(c) depicts the three-phase grid voltage sources in
control loop. the strong, weak, and very weak grids, respectively. Mean-
while, the three-phase output currents of the inverters are
D. STABILITY ANALYSIS displayed in Fig. 5(d)–(f). It is evident that the output currents
The stability of the control strategy depicted in Fig. 3 is exhibit seamless and rapid increase without any overshoot
validated through impedance stability analysis based on the during transients when changes occur in the power references.
GNC in the dq reference frame, as detailed in [31], [32]. The Furthermore, there is a slight reduction in the quality of the
robustness of the proposed control strategy is tested in strong, injected current in the strong grid scenario, as depicted in
weak, and very weak grids considering the system parameters Fig. 5(d). This reduction is attributed to the significantly lower
listed in Table 1. total inductance (Lf + Lg ) in the strong grid compared to the
Fig. 4(a) shows the bode plot of the inverter output weak and very weak grids.
impedance in the q-axis (Zoqq ), which is obtained when the Fig. 5(g)–(i) displays the dq-components of the extracted
+ +
inverter operating point is Pref = 4 MW and Qref = 0 Var. positive and negative sequences of the PCC voltage. In the
Fig. 4 also shows transfer functions for Zg1 , Zg2 , and Zg3 cor- three grid types, it can be observed that the negative sequence
responding to the strong grid with SCR = 15, weak grid with are non-zero after the unbalanced event at 0.2 s. Fig. 5(j)–
SCR = 3, and very weak grid with SCR = 1, respectively. It (l) depicts the dq-components of the extracted and reference
can be seen that the proposed controller exhibits large phase positive sequence current. It can be observed that the in-
margins. For instance, the phase margins in the q axis when verter output current accurately follow the reference. In all
the inverter is connected to the very weak grid is 159◦ < 180◦ . scenarios, the d-components dynamically adjust to track the
Fig. 4(b) presents the impedance-based stability analysis power reference commands, achieving zero steady-state er-
in the q-axis using the Nyquist plot. It is evident that the rors. Meanwhile, the q-component of the positive sequence
proposed control strategy is stable in all three grid connections current is regulated to zero, aligning the inverter controller

VOLUME 4, 2023 479

MOHAMMED ET AL.: DOUBLE-SYNCHRONOUS-REFERENCE-FRAME-BASED POWER-SYNCHRONIZED PLL-LESS GRID-FOLLOWING INVERTERS

FIGURE 5. The simulation results of the proposed DS-PSGFLI upon changes in the power reference commands at 0.1 s and 0.3 s and unbalanced grid
voltage event at 0.2 s. The results, displayed from top to bottom, show the grid voltage, PCC current, dq-components of the positive and negative
voltages, dq-components of positive current, dq-components of negative current, positive active and reactive output power, negative active and reactive
output power, and extracted frequency: The proposed DS-PSGFLI in a strong grid with SCR = 15 and Xg /Rg = 7 (Left column), in a weak grid with SCR = 2
and Xg /Rg = 7 (Middle column), and in a very weak and resistive grid with SCR = 1, Xg /Rg = 1 (Right column).

with the positive q-component of the inverter output current weak, and very weak grid, respectively. It is clear that the pro-
as intended. Fig. 5(m)–(o) shows that the proposed controller posed controller tracks the desired power reference commands
regulates the dq-components of the negative sequence current with zero steady-state errors, irrespective of the grid condi-
to zero in all cases. tions both prior to and during the occurrence of voltage sag
Fig. 5(p)–(r) displays the positive components of the active in phase B. It is important to note that the power waveforms
and reactive power waveforms of the inverter in the strong, shown here are at the PCC, which demonstrates the ability

480 VOLUME 4, 2023

of the proposed controller to regulate the PCC power in-
stead of terminal power, unlike previously proposed PSGFLI
and LPV-PSGFLI in [22], [23]. Additionally, Fig. 5(s)–(u)
shows that the proposed controller successfully regulates the
dq-components of the negative sequence active and reactive
power to zero.
Finally, Fig. 5(v)–(x) presents the extracted frequency by
the proposed control strategy. It can be observed that the
proposed controller remains stable and synchronized with
the grid, even under the large and permanent voltage sag in
the grid voltage at 0.2 s. During transients, this frequency
variable exhibits significant variations. However, it is impor-
tant to note that this frequency does not represent the grid
frequency; rather, it functions as an internal control variable
that promptly responds to changes in the active and reactive
power set-points. This characteristic ensures the achievement
of optimal dynamic performance.

B. REACTIVE POWER SUPPORT

Fig. 6 presents the PCC current, dq-components of the pos-
itive current, and positive active and reactive output power.
For this case study, the inverter is connected to the weak grid
with SCR = 2 and Xg /Rg = 7. At 0.1 s, the unbalanced grid
voltage event is applied. The proposed DS-PSGFLI is shown
to be effective in providing reactive power to the grid when
there is no active power injection from the inverter. These
results are particularly significant in scenarios where there is
a need for reactive power support, but no demand for active
power injection. This highlights the potential benefits of the
proposed method in improving power factor and decreasing
voltage fluctuations, thus improving the stability and perfor-
mance of the grid.

C. IMPACTS OF GRID VOLTAGE HARMONICS

In contrast to other tests, this section evaluates the efficacy FIGURE 6. The simulation results of the proposed DS-PSGFLI in a weak
grid with SCR = 2 and Xg /Rg = 7 upon injecting reactive power only and
of the proposed controller in the presence of a non-ideal grid unbalanced grid voltage event at 0.2 s: (a) PCC current, (b) dq-components
voltage source, taking into account power quality concerns. of the positive current, and (c) positive active and reactive output power.
The grid voltage source is contaminated with five different
harmonics, namely the 3rd , 5th , 7th , 9th , and 11th , all with an
amplitude of 3% of the fundamental voltage.
The simulation results of the proposed controller when con- Overall, these results demonstrate that the proposed controller
nected to this non-ideal grid voltage conditions are presented performs robustly by injecting the desired power references
in Fig. 7, where the inverter is connected to the weak grid while taking into account possible power quality issues.
with SCR = 2 and Xg /Rg = 7. The test is carried out with
variations in active power set-point at t = 0.1 s, applying
50% voltage sag in phase B at t = 0.2 s, and by injecting D. IMPACTS OF THE INVERTER OUTPUT FILTER
the aforementioned harmonics into the grid voltage source at In this case study, the dynamic performance of the DS-
t = 0.3 s. PSGFLI is being investigated with the use of an LC output
Fig. 7(a) displays the grid voltage source waveforms, which filter instead of the L filter in the previous section. The value of
are contaminated by the 3rd , 5th , 7th , 9th , and 11th harmonics. L remains the same as listed in Table 1, but a 1 mF capacitor is
As a result, the voltage waveforms are no longer sinusoidal added to create the output LC filter. After the implementation
for t ≥ 0.3 s. Fig. 7(b) presents the PCC current. Fig. 7(c) and of LC output filters, the proportional and integral parameters
(d) illustrates the dq-components of the inverter voltage and of the PI controllers used in the inner current controller loops
currents. Finally, Fig. 7(e) indicates that the proposed con- (both positive and negative) are adjusted. Specifically, the
troller accurately tracks the desired active and reactive power values listed in Table 1, 0.95 and 100, are modified to 0.19
set-points and slightly impacted by the harmonic distortions. and 5, respectively, for the LC filter.

VOLUME 4, 2023 481

MOHAMMED ET AL.: DOUBLE-SYNCHRONOUS-REFERENCE-FRAME-BASED POWER-SYNCHRONIZED PLL-LESS GRID-FOLLOWING INVERTERS

FIGURE 8. The simulation results of the proposed DS-PSGFLI with an LC

output filter in a weak grid with SCR = 2 and Xg /Rg = 7 upon changes in
the power reference commands at 0.1 s and 0.3 s and unbalanced grid
voltage event at 0.2 s: (a) PCC current, (b) dq-components of the positive
current, and (c) positive active and reactive output power.

The DS-PSGFLI is connected to the weak grid with

SCR = 2 and Xg /Rg = 7. Fig. 8 presents the obtained sim-
ulation results when the inverter is equipped with the LC filter
and the power reference commands are changed at t = 0.1 s
and t = 0.3 s, along with an unbalanced grid voltage event at
t = 0.2 s. The results show that the inverter exhibits robust dy-
namic performance and accurately follows the desired power
references with the LC filter.

E. COMPARISON WITH CONVENTIONAL PNSC-GFLI

FIGURE 7. The simulation results of the proposed DS-PSGFLI in a weak Fig. 9 compares the proposed DS-PSGFLI strategy with the
grid with SCR = 2 and Xg /Rg = 7 upon changes in the active power well-known conventional PNSC-GFLI [5], [26] in a weak
reference command at 0.1 s, unbalanced grid voltage event at 0.2 s, and
grid with SCR = 2.5 and Xg /Rg = 3. On the one hand, the
harmonics injection into the grid voltage source at 0.3 s: (a) Grid voltage,
(b) PCC current, (c) dq-components of the PCC voltage, (d) dq-components proposed DS-PSGFLI shows robust performance and is still
of the positive current, (e) and positive active and reactive output power. capable of operating stably before and after applying the fault.

482 VOLUME 4, 2023

FIGURE 9. Performance comparison upon changes in the power reference commands at 0.1, 0.3 and 0.4 s and unbalanced grid voltage event at t = 0.2 s.
The simulation results corresponds to a weak grid with SCR = 2.5 and Xg /Rg = 3: (a)–(c) the proposed DS-PSGFLI, and (d)–(f) the conventional PNSC-GFLI.

TABLE 2. Parameters of the Experimental Test-Bed

FIGURE 10. The experimental test-bed.

On the other hand, as shown in Fig. 9(d) to (f), the conven- A. TEST 1: PERFORMANCE IN VERY WEAK GRID
tional PNSC-GFLI strategy is driven to instability after the Fig. 11(a), (d), and (g) and (b), (e), and (k) compare the
grid voltage sag is applied at 0.2 s. This instability is due to performance of the conventional PSGFLI and the proposed
the reduction of the SCR after the fault event in the weak grid. DS-PSGFLI strategies under the permanent asymmetrical
It is important to note that the results presented in Fig. 9(f) do voltage sag of 50% in phase B in the very weak grid. Initially,
not represent static errors in the inverter output power. Instead, both inverters operate in a steady state with active power refer-
they depict heavily filtered waveforms obtained through the ence commands of Pref = 200 W and reactive power reference
use of a second-order low-pass filter with a 5 Hz cut-off commands of Qref = 100 Var. At t1 = 0.3 s, the active power
frequency. reference command is stepped from 200 to 500 W, followed
by a step change in the reactive power reference command
from 100 to 250 Var at t2 = 0.6 s.
V. EXPERIMENTAL RESULTS As demonstrated in Fig. 11(a), the conventional PSGFLI in-
The experiments are conducted using a scaled-scale test-bed jects an unbalanced current into the grid. Additionally, despite
to validate the proposed control strategy shown in Fig. 3. tracking the power set-points when changed, the conventional
A photograph of the experimental test-bed is presented in PSGFLI exhibits high oscillations in the inverter output power
Fig. 10. The test-bed includes a SiC-based two-level Imperix and frequency, as seen in Fig. 11(d) and (g) respectively. In
three-phase inverter powered by a DC source, an output L fil- contrast to the conventional PSGFLI, the proposed controller
ter, a series RL impedance that represents the grid impedance, tracks the power set-points and injects a balanced current into
and a grid simulator. The inverter is controlled by a Boom- the grid, without any oscillations in the inverter output power
Box Imperix controller that is programmed directly from or frequency, as seen in Fig. 11(e) and (h), respectively.
MATLAB/Simulink. The experiment parameters are listed in It is noteworthy that Fig. 11(a) and (b) depict a slight modi-
Table 2. The performance of the proposed control strategy fication in the magnitude of the PCC current after adjusting
is evaluated in both strong and very weak grid connections. the reactive power set-point at t2 . This is due to the small
A third test, with the same very weak grid condition, is amount of reactive current required for injection at t ≥ 2 since
conducted to demonstrate the limitations of the conventional the inverter is already injecting active and reactive current,
PSGFLI under unbalanced faults. where the active power is twice the value of the final reactive

VOLUME 4, 2023 483

MOHAMMED ET AL.: DOUBLE-SYNCHRONOUS-REFERENCE-FRAME-BASED POWER-SYNCHRONIZED PLL-LESS GRID-FOLLOWING INVERTERS

FIGURE 11. The experimental results of the PCC current waveform, output power waveforms, and the extracted frequency upon changes in the power
reference commands: (a), (d), (g) the conventional PSGFLI in a very weak grid with SCR = 1, (b), (e), (h) the proposed DS-PSGFLI in a very weak grid with
SCR = 1, and (c), (f), (i) the proposed DS-PSGFLI in a strong grid with SCR = 5.4.

power. However, if the inverter reactive power is increased TABLE 3. A Comparison Summary of the Proposed PLL-LESS DS-PSGFLI
with other existing PLL-Based and PLL-LESS Control Strategies
while the active power is small or zero, as depicted in Fig. 6,
the grid current magnitude will be noticeably altered.

B. TEST 2: PERFORMANCE IN STRONG GRID

The performance of the proposed controller is tested in a
strong grid with SCR = 5.4, and the results are displayed in
Fig. 11(c), (f), and (i). Similarly to the previous two tests,
the permanent asymmetrical voltage of 50% in phase B is
applied. At t1 = 0.3 s, the initial active power reference com-
mand Pref = 400 W is changed to Pref = 700 W, followed
by a change in the reactive power reference command from
Qref = 200 W to Qref = 350 W at t2 = 0.6 s.
As seen in Fig. 11(c), the proposed controller injects bal- is in line with commonly used control strategies in conven-
anced current into the grid under the simulated asymmetrical tional GFLI and PNSC GFLI. Finally, among all the PLL-less
grid voltage. In Fig. 11(f), it can be observed that the con- control strategies, only the proposed controller remains stable
troller tracks the power set-points and injects active and and handles abnormal grid frequency deviations and unbal-
reactive power without any oscillations. Additionally, the anced grid voltages. Overall, Table 3 highlights the superior
estimated frequency by the inverter is steady and free of oscil- performance and advantages of the proposed DS-PSGFLI
lations, as seen in Fig. 11(i) control strategy. It is important to note that the inverter system
Table 3 compares the proposed PLL-less DS-PSGFLI considered in this article is a 3-phase 3-wire system. However,
with two conventional PLL-based control strategies (GFLI future work could extend this concept to 3-phase 4-wire sys-
and PNSC), and two PLL-less control strategies (PSGFLI tems, where the AC neutral is connected to the midpoint of
and LPV-PSGFLI). The results show that the proposed DS- the DC bus. To apply this method to a 3-phase 4-wire system,
PSGFLI outperforms existing control techniques and has an additional control loop for the zero sequence should be
several notable advantages. Firstly, the proposed control strat- included in the control structure of the inverter.
egy is capable of functioning effectively in both strong and
weak grid connections and does not require a PLL for grid VI. CONCLUSION
synchronization. Secondly, unlike PSGFLI and LPV-PSGFLI, This article proposes a double-synchronous-reference-frame-
the proposed control strategy regulates the PCC power, which based control strategy for PLL-less power-synchronized
484 VOLUME 4, 2023
grid-following inverters. This control strategy outperforms [13] M. Cespedes and J. Sun, “Adaptive control of grid-connected inverters
other existing structures for grid-connected inverters, includ- based on online grid impedance measurements,” IEEE Trans. Sustain.
Energy, vol. 5, no. 2, pp. 516–523, Apr. 2014.
ing PLL-based GFLIs, PLL-less GFMIs, and conventional [14] X. Chen, Y. Zhang, S. Wang, J. Chen, and C. Gong, “Impedance-
PLL-less power-synchronized GFLIs. The proposed control phased dynamic control method for grid-connected inverters in a weak
strategy employs two inner loops to independently control grid,” IEEE Trans. Power Electron., vol. 32, no. 1, pp. 274–283,
Jan. 2017.
positive and negative sequences, preventing high oscillations [15] C. Guo, W. Liu, C. Zhao, and R. Iravani, “A frequency-based syn-
of twice the grid frequency during asymmetrical faults. The chronization approach for the VSC-HVDC station connected to a weak
proposed approach leverage the benefits offered by existing AC grid,” IEEE Trans. Power Del., vol. 32, no. 3, pp. 1460–1470,
Jun. 2017.
power-synchronized control strategies as the inverter does [16] M. Z. Mansour, M. H. Ravanji, A. Karimi, and B. Bahrani, “Small-
not depend on the grid strength due to the removal of the signal synchronization stability enhancement of grid-following invert-
PLL from the control structure. However, unlike the existing ers via a feedback linearization controller,” IEEE Trans. Power Del.,
vol. 37, no. 5, pp. 4335–4344, Oct. 2022.
power-synchronized control strategies, the proposed method [17] X. Zhang, D. Xia, Z. Fu, G. Wang, and D. Xu, “An improved feedfor-
regulates the inverter power at the PCC instead of the ter- ward control method considering PLL dynamics to improve weak grid
minal, avoiding unnecessary voltage drops and high voltage stability of grid-connected inverters,” IEEE Trans. Ind. Appl., vol. 54,
no. 5, pp. 5143–5151, Sep./Oct. 2018.
harmonics. The robust performance of the proposed control [18] T. Noguchi, H. Tomiki, S. Kondo, and I. Takahashi, “Direct power
strategy is verified in simulation using MATLAB/Simulink control of PWM converter without power-source voltage sensors,” IEEE
and experimentally using a scaled-down inverter test-bed. Trans. Ind. Appl., vol. 34, no. 3, pp. 473–479, May/Jun. 1998.
[19] S. Vazquez, A. Marquez, R. Aguilera, D. Quevedo, J. I. Leon, and L. G.
The proposed control strategy enables reliable integration of Franquelo, “Predictive optimal switching sequence direct power con-
large-scale renewable energy resources into established power trol for grid-connected power converters,” IEEE Trans. Ind. Electron.,
systems by mitigating instabilities caused by several factors vol. 62, no. 4, pp. 2010–2020, Apr. 2015.
[20] Y. Gui, C. Kim, C. C. Chung, J. M. Guerrero, Y. Guan, and J.
such as grid strength variations, grid frequency deviations, and C. Vasquez, “Improved direct power control for grid-connected volt-
unbalanced/fault events. age source converters,” IEEE Trans. Ind. Electron., vol. 65, no. 10,
pp. 8041–8051, Oct. 2018.
[21] Y. Gui, X. Wang, H. Wu, and F. Blaabjerg, “Voltage-modulated
direct power control for a weak grid-connected voltage source invert-
REFERENCES ers,” IEEE Trans. Power Electron., vol. 34, no. 11, pp. 11383–11395,
[1] F. Blaabjerg, R. Teodorescu, M. Liserre, and A. V. Timbus, “Overview Nov. 2019.
of control and grid synchronization for distributed power generation [22] B. Bahrani, “Power-synchronized grid-following inverter without
systems,” IEEE Trans. Ind. Electron., vol. 53, no. 5, pp. 1398–1409, a phase-locked loop,” IEEE Access, vol. 9, pp. 112163–112176,
Oct. 2006. 2021.
[2] S. Xu, Y. Xue, and L. Chang, “Review of power system support [23] M. Z. Mansour, M. H. Ravanji, A. Karimi, and B. Bahrani, “Linear
functions for inverter-based distributed energy resources-standards, parameter-varying control of a power-synchronized grid-following in-
control algorithms, and trends,” IEEE Open J. Power Electron., vol. 2, verter,” IEEE J. Emerg. Sel. Topics Power Electron., vol. 10, no. 2,
pp. 88–105, 2021. pp. 2547–2558, Apr. 2022.
[3] H. Gong, X. Wang, and L. Harnefors, “Rethinking current controller [24] O. P. Mahela, N. Gupta, M. Khosravy, and N. Patel, “Comprehensive
design for PLL-synchronized VSCs in weak grids,” IEEE Trans. Power overview of low voltage ride through methods of grid integrated wind
Electron., vol. 37, no. 2, pp. 1369–1381, Feb. 2022. generator,” IEEE Access, vol. 7, pp. 99299–99326, 2019.
[4] J. Rocabert, A. Luna, F. Blaabjerg, and P. Rodriguez, “Control of power [25] M. A. G. López, J. L. G. de Vicuña, J. Miret, M. Castilla, and
converters in AC microgrids,” IEEE Trans. Power Electron., vol. 27, R. Guzmán, “Control strategy for grid-connected three-phase invert-
no. 11, pp. 4734–4749, Nov. 2012. ers during voltage sags to meet grid codes and to maximize power
[5] R. Teodorescu, M. Liserre, and P. Rodriguez, Grid Converters for Pho- delivery capability,” IEEE Trans. Power Electron., vol. 33, no. 11,
tovoltaic and Wind Power Systems. Hoboken, NJ, USA: Wiley, 2011. pp. 9360–9374, Nov. 2018.
[6] P. Rodríguez, J. Pou, J. Bergas, J. I. Candela, R. P. Burgos, and D. [26] R. Kabiri, D. G. Holmes, and B. P. McGrath, “Control of active and re-
Boroyevich, “Decoupled double synchronous reference frame PLL for active power ripple to mitigate unbalanced grid voltages,” IEEE Trans.
power converters control,” IEEE Trans. Power Electron., vol. 22, no. 2, Ind. Appl., vol. 52, no. 2, pp. 1660–1668, Mar./Apr. 2016.
pp. 584–592, Mar. 2007. [27] R. Errouissi and A. Al-Durra, “Disturbance-observer-based control for
[7] Q. Huang and K. Rajashekara, “An improved delayed signal cancella- dual-stage grid-tied photovoltaic system under unbalanced grid volt-
tion PLL for fast grid synchronization under distorted and unbalanced ages,” IEEE Trans. Ind. Electron., vol. 66, no. 11, pp. 8925–8936,
grid condition,” IEEE Trans. Ind. Appl., vol. 53, no. 5, pp. 4985–4997, Nov. 2019.
Sep./Oct. 2017. [28] M. Reyes, P. Rodriguez, S. Vazquez, A. Luna, R. Teodorescu, and J. M.
[8] Y. Cheng et al., “Real-world subsynchronous oscillation events in power Carrasco, “Enhanced decoupled double synchronous reference frame
grids with high penetrations of inverter-based resources,” IEEE Trans. current controller for unbalanced grid-voltage conditions,” IEEE Trans.
Power Syst., vol. 38, no. 1, pp. 316–330, Jan. 2023. Power Electron., vol. 27, no. 9, pp. 3934–3943, Sep. 2012.
[9] F. Arraño-Vargas, Z. Shen, S. Jiang, J. Fletcher, and G. Konstantinou, [29] N. Mohammed, M. Ciobotaru, and G. Town, “Online parametric esti-
“Challenges and mitigation measures in power systems with high share mation of grid impedance under unbalanced grid conditions,” Energies,
of renewables–the australian experience,” Energies, vol. 15, no. 2, 2022, vol. 12, no. 24, 2019, Art. no. 4752.
Art. no. 429. [30] N. Mohammed, M. H. Ravanji, W. Zhou, and B. Bahrani, “Online grid
[10] B. Wen, D. Boroyevich, R. Burgos, P. Mattavelli, and Z. Shen, “Anal- impedance estimation-based adaptive control of virtual synchronous
ysis of DQ small-signal impedance of grid-tied inverters,” IEEE Trans. generators considering strong and weak grid conditions,” IEEE Trans.
Power Electron., vol. 31, no. 1, pp. 675–687, Jan. 2016. Sustain. Energy, vol. 14, no. 1, pp. 673–687, Jan. 2023.
[11] W. Zhou, N. Mohammed, and B. Bahrani, “Comprehensive model- [31] J. Sun, “Impedance-based stability criterion for grid-connected invert-
ing, analysis, and comparison of state-space and admittance models of ers,” IEEE Trans. Power Electron., vol. 26, no. 11, pp. 3075–3078,
PLL-Based grid-following inverters considering different outer control Nov. 2011.
modes,” IEEE Access, vol. 10, pp. 30109–30146, 2022. [32] Q. Peng, G. Buticchi, N. M. L. Tan, S. Guenter, J. Yang, and P. Wheeler,
[12] A. A. A. Radwan and Y. A.-R. I. Mohamed, “Improved vector control “Modeling techniques and stability analysis tools for grid-connected
strategy for current-source converters connected to very weak grids,” converters,” IEEE Open J. Power Electron., vol. 3, pp. 450–467,
IEEE Trans. Power Syst., vol. 31, no. 4, pp. 3238–3248, Jul. 2016. 2022.

VOLUME 4, 2023 485

MOHAMMED ET AL.: DOUBLE-SYNCHRONOUS-REFERENCE-FRAME-BASED POWER-SYNCHRONIZED PLL-LESS GRID-FOLLOWING INVERTERS

NABIL MOHAMMED (Member, IEEE) received BEHROOZ BAHRANI (Senior Member, IEEE) re-
the bachelor’s degree (Hons.) in electrical power ceived the [Link]. degree in electrical engineering
engineering from Tishreen University, Latakia, from the Sharif University of Technology, Tehran,
Syria, in 2013, the [Link]. degree in electrical Iran, in 2006, the [Link]. degree in electrical engi-
engineering from Universiti Teknologi Malaysia, neering from the University of Toronto, Toronto,
Johor Bahru, Malaysia, in 2017, and the Ph.D. ON, Canada, in 2008, and the Ph.D. degree in elec-
degree in power electronics from Macquarie Uni- trical engineering from the Ecole Polytechnique
versity, Sydney, NSW, Australia, in 2022. During Fédérale de Lausanne (EPFL), Lausanne, Switzer-
the summer of 2019, he was a Visiting Researcher land, in 2012. From September 2012 to September
with the Department of Energy Technology, Aal- 2015, he was a Postdoctoral Fellow at several in-
borg University, Aalborg, Denmark. He is currently stitutions including EPFL, Purdue University, West
a Postdoctoral Research Fellow with Monash University, Clayton, VIC, Lafayette, IN, USA, Georgia Institute of Technology, Atlanta, GA, USA, and
Australia. His research interests include power electronic converters, grid the Technical University of Munich, Munich, Germany. Since 2015, he has
integration of renewable energy resources, microgrids, energy storage and been with Monash University, Clayton, Australia, where he is currently a
management systems, and modeling and control of electric power systems. Senior Lecturer and the Director of the Grid Innovation Hub. His research
interests include control of power electronics systems, applications of power
electronics in power and traction systems, and grid integration of renewable
WEIHUA ZHOU (Member, IEEE) was born in energy resources.
Anhui, China, in 1993. He received the [Link].
degree in electrical engineering from the Honors
College, Northwestern Polytechnical University,
Xi’an, China, in 2014, the [Link]. degree in elec-
trical engineering from the School of Automation,
Northwestern Polytechnical University, in 2017,
and the Ph.D. degree in power electronics from the
Department of Energy Technology, Aalborg Uni-
versity, Aalborg, Denmark, in 2020. From 2015 to
2016, he was a Junior Research Assistant with the
Department of Mechanical and Automation Engineering, The Chinese Uni-
versity of Hong Kong, Hong Kong, Hong Kong. From August to November
2020, he was a Visiting Scholar with the Department of Marine Technol-
ogy, Norwegian University of Science and Technology, Trondheim, Norway.
In 2021, he was a Postdoctoral Fellow with the Department of Electrical
Engineering, KU Leuven, Leuven, Belgium. He is currently a Postdoctoral
Research Fellow with the Department of Electrical and Computer Systems
Engineering, Monash University, Melbourne, VIC, Australia. His research
interests include modeling and control of power electronic converters, energy
storage systems, and stability analysis and enhancement of power electronic-
and underground cable-based modern power systems.

486 VOLUME 4, 2023