IET Renewable Power Generation

Research Article

Low-voltage ride-through control for ISSN 1752-1416

Received on 17th October 2019

photovoltaic generation in the low-voltage

Revised 8th July 2020
Accepted on 31st July 2020
E-First on 2nd October 2020
distribution network doi: 10.1049/iet-rpg.2019.1101
www.ietdl.org

Yufei He1, Minghao Wang1, Youwei Jia2, Jian Zhao3, Zhao Xu1
1Department of Electrical Engineering and Research Institute for Smart Energy, The Hong Kong Polytechnic University, Kowloon, Hong Kong
2Department of Electrical and Electronic Engineering, Southern University of Science and Technology, People's Republic of China
3College of Electrical Engineering, Shanghai University of Electric Power, Shanghai, People's Republic of China

E-mail: [email protected]

Abstract: The increasing penetration of photovoltaic (PV) energy in power grids will impose system instability issues, especially
in the occurrence of faults. However, very limited research has been conducted on the low-voltage ride-through (LVRT) control
of PV systems in the low-voltage distribution networks (LVDNs) with predominantly resistive line impedances. To fulfil this
remaining gap, the effects of active current injection (ACI) on the grid voltage support in LVDN is mathematically analysed in this
study. Subsequently, a novel LVRT control scheme for the PV grid-forming inverter is proposed, where the control distinguishes
itself from other existing methods due to its optimisation of ACI and PV energy harvesting with the premise of system safety and
specified reactive current injection as per grid codes. Multi-mode control modes are involved in the proposed method for dealing
with different environmental conditions and voltage dips. Meanwhile, the DC-link voltage is adaptively operated in a self-
adjustable control structure for improving grid resilience. The effectiveness of the proposed control method is verified by
simulations in MATALB Simulink and hardware experiments on a PV microinverter. Compared with the traditional LVRT control,
the post-fault power recovery and voltage support capability can be significantly improved.

1 Introduction capability of PV generators can be also enhanced via demand-side

management. Although the aforementioned methods could
The photovoltaic (PV) generation is a promising alternative of the maintain the MPPT and ensure generation system safety, additional
conventional fossil fuel-based power plants while great challenges circuit components are required. To keep system safe but avoid
of its large-scale grid integration are still pending to be addressed extra installation costs, de-loading control is receiving increasing
[1]. Traditionally, PV generators are operated in the maximum attention in recent research works [13–20], where the PV array is
power point tracking (MPPT) mode under normal grid conditions operated away from the maximum power point (MPP) for power
and tripped off as soon as voltage dip happens at the point of curtailment during faults.
common coupling (PCC). Such an operation will be detrimental for Considering the grid code progress in the past decade, the other
the power system with high PV-penetration since it may render the aspect is to provide reactive current injection (RCI) for supporting
collapse of system frequency [2]. To prevent this, the low-voltage the PCC voltage [21]. A typical pattern of reactive current support
ride-through (LVRT) capability is necessarily required by many required by the grid codes is shown in Fig. 1b. Generally, the grid-
international standards for PV generators. Typically, IEEE tied PV inverter is demanded to provide a 2% reactive current for
1547-2003 demands that PV generators must remain grid-tied with every 1% voltage drop. [13]. The RCI methods can be
unity power factor under specific voltage-dip circumstances. Since implemented on both the single-stage PV inverters [14] and two-
2010, reactive current in proportion with the severity of voltage stage PV inverters [15]. A decoupled current control on PV
dips has been required for supporting the voltage recovery at PCC systems is reported in [16] for improving LVRT capability, where
[3]. The latest LVRT criteria are summarised as shown in Fig. 1a, the output active and reactive currents of the grid-tied inverter are
where the PV generators must maintain to be grid connected when separately regulated in the synchronous rotating coordinate frame.
the PCC voltage is above the predefined values (borderlines in Direct power control [17] and predictive power control [18] with
Fig. 1a). the space vector pulse-width modulation are also adopted to
Much research has been conducted to pursue a safe and robust provide required reactive power for fulfilling the grid code
LVRT performance in [4–10, 11–20], of which the strategies focus requirements. In addition, PV inverters can be utilised as the static
on two main aspects. The first aspect is to ensure safe ride through. synchronous compensator (STATCOM) [19, 20] during the LVRT
Due to the operational limits of electrical components, the energy period.
injected into the grid must be curtailed to avoid over-current or In summary, the majority of previous studies on the LVRT
over-voltage under faults. Traditionally, there are methods based on capability for PV systems focus on simultaneously decreasing
crowbar circuits [5], dynamic resistors or DC choppers [6]. In these output active power for the system safety and increasing RCI for
methods, the excessive energy generated during the fault will be voltage support. This can only achieve satisfactory LVRT
dissipated for system safety. In [7, 8], a current amplitude performance in the transmission networks when the line impedance
limitation control is reported with the DC chopper, where the is significantly inductive. Due to the booming of distributed (PVs
inverter can consistently inject real power to the grid and the over- and enforcement of feed-in-tariff [22], large amounts of PV
current can be prevented. To avoid the unnecessary dissipation of generation systems are installed in the low-voltage distribution
renewable energy, energy storage systems (ESSs), such as network (LVDN) or micro-grids, such as building-integrated PVs
supercapacitors and batteries [9], are deployed for storing extra (BIPVs) [23] and residential PV generators [24] where the
energy during the faults and release it back to the grid after the generation will preferentially meet the local load demands. It has
fault clearance [10]. However, the ESS has a much shorter lifetime been concluded that LVRT capability should also be mandatory in
than the PV components do, and therefore additional care is needed LVDN [4, 25]. However, the active power curtailment control
in daily operation [12]. It is proposed in [11] that the LVRT involved in existing LVRT methods limits PV power production

IET Renew. Power Gener., 2020, Vol. 14 Iss. 14, pp. 2727-2737 2727
© The Institution of Engineering and Technology 2020
 17521424, 2020, 14, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-rpg.2019.1101 by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [26/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
• Adaptive DC-link voltage control is applied for buffering a
certain amount of PV energy with the self-adjusting control
structure to (i) accelerate post-fault recovery in the power grid,
(ii) provide more and accurate active current for grid voltage
support and (iii) compensate the power fluctuation caused by
random environmental changes.
• Different control modes are designed with respect to the various
voltage-drop depths at PCC and working conditions.

The paper is organised as follows: the configuration of the PV

generation system and traditional control for LVRT operation is
discussed in Section 2. The voltage support effects of ACI will be
analysed in Section 3. The proposed LVRT control scheme is
elaborated in Section 4. In Section 5, simulation results under
different scenarios are provided to verify the effectiveness of the
proposed control scheme. In Section 6, the experiment result of a
solar micro-inverter is provided to testify the functionality of the
proposed LVRT control.

2 PV generation system and traditional de-

loading and RCI-based control strategies
2.1 PV generation system in LVDN
PV generation tends to grid-connection from the LVDN in the
future [1, 22–24]. The power rating of these PV systems ranges
from a few hundred watts (e.g. solar microinverter) to several
hundred kilowatts (e.g. large BIPV) [23]. The output power of PV
arrays is influenced by the varying ambient temperature and solar
irradiance. As shown in Fig. 2, PV arrays are usually connected to
Fig. 1 Summary of grid code requirements on LVRT capabilities the PCC through an array-side voltage source converter (ASVSC),
(a) Various LVRT grid codes for distributed PV generators, (b) Inverter output current DC-link capacitors, a grid-side voltage source converter (GSVSC),
diagram with the typical RCI requirement an L-type filter and an isolation transformer. Practically, the boost
converter is usually applied as the ASVSC for the step-up function
and continuous input current feature. MPPT algorithms are
generally implemented on the PV generator for realising the
maximum solar energy harvesting in normal operation. Three/
single-phase inverters are applied as the GSVSC to deliver AC
power to the LVDN [22]. In case of short-circuit or transient power
imbalance, grid faults may occur, which can propagate throughout
the power system and pull down the PCC voltage. When the
voltage dip is detected, an effective LVRT control must be
Fig. 2 Traditional LVRT control diagram activated to protect and support the system.

that may cause local energy shortage and potential frequency 2.2 Traditional RCI-based LVRT control under the PCC
instability of the system under the faults. Besides, RCI will be less voltage fault
effective for voltage support in LVDN since the resistive
component dominates in the line impedance [25]. Therefore, it To fulfil the latest grid code, the decoupled current control in the
would be highly valuable for LVRT control to adopt the active synchronous coordinate frame (d–q control) is generally combined
current injection (ACI) control for injecting certain active current with the PV de-loading control for achieving the safe operation of
to the grid, which can produce pronounced effects on both local the system and reactive current compensation [16]. The control
power supply and voltage support for the LVDN. Furthermore, the block diagram can be plotted as shown in Fig. 2.
energy buffering potential of DC-link capacitors in the PV
generation system is, first, exploited in [26], where the capacitors 2.2.1 LVRT control on the GSVSC: To achieve decoupled
are controlled to reserve certain energy during fault periods and control on the active and reactive currents, Park (P) and Clarke (C)
release it back after fault clearance. However, only preliminary transformations are used, which can be expressed as
simulation works under standard testing conditions (STCs) are
conducted in this work. It still lacks full investigation in different xmcos(ωt)
working conditions and experimental implementation, which leads xd xa 2π
to the underutilisation of the existing resources in the PV system. x cos ωt −
xq = P ∗ C ∗ xb = P ∗ C ∗ m 3 (1)
In this paper, a novel LVRT control scheme is proposed for the
PV system in LVDN. The proposed control scheme can enhance x0 xc 2π
xmcos ωt +
LVRT and voltage support capabilities without any extra devices. 3
The main contributions of this work can be summarised as follows:
1 1
1 − −
• The effectiveness of ACI on voltage boosting in LVDN during 2 2
cos θ sin θ 0
voltage dips is verified through the mathematical analysis. 2 3 3
• The proposed LVRT control optimises ACI and solar energy P = −sin θ cos θ 0 ,C = 0 − (2)
3 2 2
harvesting under the premise of system safety and specified RCI 0 0 1
as per grid codes. This can maximise grid-supporting 2 2 2
capabilities and energy conversion efficiency of the PV 2 2 2
generators in the voltage transient.
are applied for mapping the three-phase signals onto the d–q
rotating coordinate system. In (1) and (2), the symbol x represents
2728 IET Renew. Power Gener., 2020, Vol. 14 Iss. 14, pp. 2727-2737
© The Institution of Engineering and Technology 2020
 17521424, 2020, 14, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-rpg.2019.1101 by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [26/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Z = R + jX (3)

where X and R are the equivalent inductance and resistance.

By referring to Fig. 3b, it can be expressed that

Ug − Ig ⋅ Z = Ue (4)

Fig. 3 Thevenin's equivalent circuit in LVDN Under the symmetrical voltage condition, only direct-sequence
(a) Equivalent circuit of a faulted system, (b) Phasor diagram component Ugd exists, which is in phase with the d-axis current Id
and 90° ahead of q-axis current Iq in the GSVSC. By substitute (3)
the output voltage ug or current ig of the GSVSC. The subscripts a, into (4), it can be derived that
b and c refer to the corresponding phases. The subscripts d, q, and
0 refer to the direct, quadrature and zero components, respectively. Ugd − (Id − jIq) ⋅ (R + jX) = Ue (5)
The θ is the phase angle of phase-a voltage, which is tracked by the
phase-locked loop (PLL). To quantify the voltage-supporting capability of the decoupled
As shown in Fig. 2, outer-voltage inner-current proportional- current control, it is assumed that the Ue and Z are fixed during the
integral (PI) controllers are applied for decoupled current control of fault period, so that
the GSVSC. The DC-link voltage is compared with a
predetermined reference to generate the output active current 2
(Ugd − Id R − Iq X)2 + (IqR − Id X)2 ≡ Ue = const . (6)
reference Idref through the DC-link controller, while the output
reactive current reference is set at zero in normal operation. During The derivative of (6) can be expressed as
the LVRT period, the active current control is still realised by the
double-loop control with unchanged DC-link voltage reference. dUgd = A ⋅ dId + B ⋅ dIq (7)
Meanwhile, a positive reactive current reference Iqref is calculated
based on the grid codes to feed more reactive power into the grid. ∂Ugd (IqR − Id X)X
A= = R +
∂Id A 1
U gd − Id R − Iq X (8)
2.2.2 LVRT control on the ASVSC: To realise a safe ride- A2
through process, the deloading proportion Kde is applied to curtail
the PV generation based on the voltage sag depth, as shown in ∂Ugd (Id X − IqR)R
B= = X +
Fig. 2. For the current-based control [7], the output current of the ∂Iq B
U gd − Id R − Iq X (9)
1
PV array is regulated to a fixed value. This assumes that PV output B2
power is linearly associated with its current when the voltage
exceeds the MPP [19]. The power-based control [16] can achieve Different from the high-/medium-voltage power grids, LVDN is
accurate control of the PV output power. The above two control usually of high R/X ratio, i.e.
strategies can be defined as the input-oriented control, of which the
active power demand is used as the reference in the ASVSC R
control. Thus, the real power injected to the grid will follow the >1 (10)
X
output power variation of the PV array. After the fault is cleared,
MPPT control will be resumed in the ASVSC. Based on the A1 and B1 in (8) and (9), it is concluded that in case of
Z with high R/X ratio, ACI can provide effective voltage support
2.2.3 Disadvantages of the traditional LVRT control: Three capability, as RCI does with high X/R ratio, which is illustrated in
disadvantages in traditional LVRT control can be listed as follows: Fig. 4. Therefore, after satisfying the RCI required by grid codes,
the remaining power capacity of the GSVSC should be prioritised
• Less effective for voltage support in LVDN, where the resistive for ACI in the LVDN.
component dominates in the line impedance. Based on the A2 and B2 in (8) and (9), it can be further
• Underutilisation of the existing resource in the generator. concluded that (i) under normal operation with MPPT, when
Specifically, DC-link capacitors can absorb or release a certain Id ≫ Iq, B2 is larger so that RCI can be adopted to mitigate the
amount of energy in the transient voltage event, which is hardly PCC voltage fluctuation and (ii) under LVRT operation, especially
been discussed in the previous research. with a severe PCC voltage drop when Iq ≃ 1 p . u . ≫ Id as per grid
• Unnecessary active power curtailment and sub-optimal energy codes, A2 is larger so that ACI will provide a significant effect on
harvesting as it lacks consideration for different working voltage support.
conditions and operational limitations in the LVRT operation. Considering the design margin of the power electronic
components, the inverter can be temporarily operated above the
3 Voltage support effect of ACI control rated point for achieving a larger power transferring capacity in the
LVRT operation. In this paper, the maximum allowable current
Previous LVRT research generally assumes the inductive line
output of GSVSC is temporarily set to be 1.1 p.u. during faults so
impedance [13] so that only RCI is required for voltage support. In
that more capacity for ACI can be generated. The feasible
most literature, PV generation is ceased when the PCC voltage is
operation region of GSVSC is illustrated with shadows in Fig. 5,
<0.5 p.u. [19]. Moreover, the PV inverter is entirely utilised as
where the same colour in the four subplots represents the same
STATCOM for reactive power production in [20].
operation point. The increase of ACI during faults can be
However, in the LVDN, the resistive characteristic of line
illustrated in Fig. 6. It is worth mentioning that in case that PCC
impedance is more significant [25]. To mathematically analyse the
voltage is <0.5 p.u., no ACI can be provided with GSVSC current
voltage support effect of decoupled current control of the GSVSC,
of 1 p.u. in traditional control. However, by adopting a slight
the Thevenin's equivalent model is adopted for simplification. As
increase of GSVSC current (from 1 to 1.1 p.u.), 0.46 p.u. ACI can
indicated in the dashed box of Fig. 3a, a low-voltage source Ue is
be retained in the fault period. This can potentially enhance the
used to represent the voltage vector at the faulted point. The voltage-supporting capabilities and power the local loads in LVDN.
impedance between the faulted point and PCC is denoted as Z.
In Fig. 3b, Ug, Uline and Ig are the vectors of the PCC voltage,
equivalent line voltage and output current of the GSVSC,
respectively. Z can be further expressed as

IET Renew. Power Gener., 2020, Vol. 14 Iss. 14, pp. 2727-2737 2729
© The Institution of Engineering and Technology 2020
 17521424, 2020, 14, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-rpg.2019.1101 by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [26/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Fig. 6 ACI capabilities with different GSVSC currents

T
ref
v = Ipv Idref Iqref
ref
Udc (11)

Obj: arg max Id + Epv (12)

v

ref mpp
subject to Ppv ≤ Ppv (13)

Ug − Urated
Iqref = min 2 ⋅ ⋅ Irated, Irated (14)
Urated

Idref ≤ 1.12 ⋅ Irated

2
− I ref
q
2
(15)

ref
min
Udc ≤ Udc max
≤ Udc (16)

where v is the vector of decision variables, including PV output

ref
current reference Ipv , d-axis GSVSC current reference Idref , q-axis
GSVSC current reference Iqref and DC-link voltage reference Udc ref
in
the LVRT operation. The objective function is to maximise the ACI
and PV energy harvesting Epv during faults. The environmental
limitation is described in (13), which indicates that operative PV
ref mpp
output power Ppv cannot exceed the MPP Ppv in the fault period.
Considering the environment condition during faults is usually not
the STC, it can be conceived that
mpp rated
Ppv ≤ Ppv ≃ Po = ph ⋅ Urated ⋅ Irated (17)

rated
where Ppv is the rated power output of PV arrays under STC. Po
is the rated power output of GSVSC. Urated and Irated are the rated
RMS value of GSVSC output phase voltage and current,
Fig. 4 Phasor diagrams of ACI and RCI control when respectively. ph denotes the total phase number of the generation
(a) R/X ratio is large, (b) R/X ratio is small system. During LVRT operation, Iqref is determined by using (14) as
per the grid code, where Ug is the retained RMS phase voltage of
PCC in the fault. Referring to the discussion in Section 3, 1.1 p.u.
Irated is adopted as the maximum allowable output current of
GSVSC to enhance the voltage boost capability by ACI, as
described in (15). Udc max
and Udc
min
denote the maximum and
minimum allowable DC-link voltage with respect to Ug. Based on
(14) and (15), the maximum permissive active power output of
GSVSC Pgmax can be obtained as

2Ug 2 Ug
Pgmax = 1.12 − 2 − ⋅ ⋅P (18)
Urated Urated o

According to the different PCC voltage and environmental

conditions, the proposed control can be categorised into three
different modes, which are as follows:

Fig. 5 Operational region of the GSVSC with GSVSC current of 1.1 p.u. in (i) Mode 1: If Ug > 0.9 p . u ., the PV generator will be operated in
the proposed LVRT control the normal mode with MPPT implementation in the ASVSC. In
order to provide a voltage response to the system, the output
4 Proposed LVRT control for PV systems reactive current of GSVSC is produced with respect to the PCC
voltage variation.
4.1 Supervisory control on mode selection
(ii) Mode 2: If Ug ≤ 0.9 p . u . & Pgmax ≤ Ppvmpp
, LVRT control will
Since (i) ACI is proved to be more effective on voltage support in be activated with a solution-oriented structure. After guaranteeing
the LVDN and (ii) local load demand still exists in the fault, the the specified RCI based on grid codes, the remaining power
proposed LVRT control will maximise the ACI and solar energy capacity of GSVSC will be deployed for the accurate ACI control.
harvesting during the fault, which should be under the premise of Meanwhile, the DC-link voltage is regulated to be higher than its
system safety and specified RCI as per grid codes. Essentially, it nominal value Udc norm
, storing certain extra PV energy as the
can be formulated into an optimisation problem as electrical potential energy of capacitors.

2730 IET Renew. Power Gener., 2020, Vol. 14 Iss. 14, pp. 2727-2737
© The Institution of Engineering and Technology 2020
 17521424, 2020, 14, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-rpg.2019.1101 by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [26/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
mpp
(iii) Mode 3: If Ug ≤ 0.9 p . u . andPgmax > Ppv , LVRT control will Table 1 Operations of different control modes
be activated with the input-oriented structure. In this mode, MPPT Mode Adaptive DC- Active Reactive Control
control is maintained. At the same time, the DC-link voltage is link voltage power power structure
regulated to be lower than its nominal value Udc norm
, releasing more 1 norm
constant at Udc MPPT volt-var voltage
energy to improve ACI to support PCC voltage. response
2 higher than accurate ACI grid code solution-
The characteristics of the above control modes are summarised in norm
Udc oriented
Table 1. control
3 norm
lower than Udc MPPT + grid code input-oriented
4.2 Mode 1: normal operation with voltage response enhanced ACI control

In case of a slight voltage drop, the Volt-Var voltage response is

programmed in the proposed control scheme. The block diagram of
this control mode is shown in Fig. 7a.
The GSVSC output active current id and reactive current iq are
regulated via the inner-current control loops. By linearising the
control inputs, the feedforward term can be expressed as

dd / q = (ud / q + ugd / gq + id / qRf ∓ ωLf iq / d)/udc (19)

where Lf and Rf denote the inductance of the filter and its

equivalent series resistor. ud / q, ugd / gq and dd / q denote the voltage
before feedforward, PCC voltage and duty ratio, respectively. udc is
the instantaneous voltage of DC-link capacitors.
By applying the loop-shaping method, the control law can be
designed as

did / q
ud / q = Lf
dt
= K1ei_d / q + K2 ∫e i_d / qdt (20)
Fig. 7 Operation mode 1
(a) Control diagram, (b) Energy flow
where

ei_d / q = Idref/ q − id / q (21)

Assuming the change of id / q is much faster than Idref/ q, (19) can be

reformulated as

dei_d / q
K1ei_d / q + K2 ∫e i_d / qdt + Lf
dt
=0 (22)

Thus, the GSVSC output current error eid / q will converge from its
initial value e(0) to zero according to (23)

sLf e(0)
ei_d / q(s) = (23)
Lf s2 + K1s + K2

To achieve a fast and robust dynamic response, the parameters of

the controller can be designed as
Fig. 8 Operation mode 2
K1 = 2 2πLf ⋅ 0.1 f sw K2 = 4π 2Lf (0.1 f sw)2 (24) (a) Control diagram, (b) Energy flow

where f sw is the switching frequency. As shown in Fig. 7b, by neglecting the power losses of the
As shown in Fig. 7a, MPPT is kept activated to maximise the converter, all the solar energy is delivered to the power grid in
energy harvesting in this mode. Unlike the conventional LVRT mode 1, where Eg and Epv denote the energy injected to the grid
control, a positive Iqref is regulated based on the PCC voltage and harvested by the solar arrays, respectively. The DC-link
variation through a PI controller, which can be expressed as capacitors are only used to filter high-frequency ripples.

4.3 Mode 2: LVRT operation with accurate ACI

Iqref = Kp_ac(Urated − Ug) + Ki_ac ∫ (U rated − Ug)dt (25)
During the LVRT period, if the MPP in the current environment
exceeds the maximum allowable output power of GSVSC, control
where Kp_ac and Ki_ac are the gains of the PI controller. mode 2 will be activated. As shown in Fig. 8a, a simultaneous
To ensure the smooth transition between the volt-var control control on the DC-link voltage regulation and PV power
(mode 1) with LVRT control (modes 2/3), the operational range of curtailment is performed in mode 2.
Iqref can be restricted as On the one hand, the DC-link voltage reference will be set
higher than the nominal value, so that the DC-link capacitors will
mpp 2 store certain real energy during the fault. Based on the statistical
Ppv
0≤ Iqref ≤ min (1.1Irated) − 2
, 0.2Irated (26) data record of voltage-dip circumstances [27], the time duration of
ph ⋅ Ug most grid faults is <100 ms. In view of this, a significant
proportion of energy can be stored in the DC-link capacitors during

IET Renew. Power Gener., 2020, Vol. 14 Iss. 14, pp. 2727-2737 2731
© The Institution of Engineering and Technology 2020
 17521424, 2020, 14, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-rpg.2019.1101 by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [26/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Po
Cdc = 2 in case of f = 50 Hz (28a)
norm
5π Udc

mpp
10πKetf Ppv − Pgmax
ref
Udc norm
= Udc 1+ (28b)
Po

mpp ref norm

In the case of Ppv = Po and tf = 0.1 s, Udc normalised by Udc
can be illustrated with respect to Ug in Fig. 9. Different Ke can be
selected to change the amount of energy stored in the DC-link
capacitors during fault periods. For example, when Ke equals to
0.3, at least 30% of PV energy that is wasted with constant DC-link
Fig. 9 Adaptive DC-link voltage reference with respect to the retained voltage control in the fault period can be reserved in the DC-link
PCC voltage capacitors.
As shown in Fig. 8b, a certain amount of extra solar energy is
stored as Ec in control mode 2, which will correspondingly reduce
the de-loaded energy Ede. The proposed adaptive DC-link voltage
control has the following advantages over the conventional LVRT
Fig. 10 OCC diagram of ASVSC in mode 2 strategies: (i) during the fault, more solar energy can be reserved
instead of being dissipated and (ii) after the fault clearance, the
stored energy in DC-link capacitors can help to potentially improve
the grid resilience by the rapid release of electrical potential
energy. In other words, the post-fault power recovery can be
accelerated, which is also specified by the grid code.
On the other hand, due to the unpredictable 5–10% power
Fig. 11 Active power flows with the proposed solution-oriented and
losses in the converters, the input-oriented control can lead to the
traditional input-oriented structures
unstable power output of GSVSC. To address this, the solution-
oriented control structure is proposed in control mode 2 for
achieving the accurate control of ACI. As shown in Fig. 8, the
GSVSC is directly controlled by the single-loop current controller.
The decoupled current references are calculated by using (14) and
(15). The compensator can be designed by the loop-shaping
method as expressed by (24).
The active power imbalance between the PV array and GSVSC
output is reflected by the variation of udc, i.e.

dudc
Cdcudc = UpvIpv − ph ⋅ UgIdref (29)
dt

where Upv and Ipv are the output voltage and current of the PV
array. In mode 2, the DC-link voltage feedback signal is switched
into the ASVSC control for deriving the current reference, which
can be expressed as

Fig. 12 Operation mode 3

ref
Ipv ref
= Kp_pv(Udc − udc) + Ki_pv ∫ (U ref
dc − udc)dt (30)
(a) Control diagram, (b) Energy flow
ref
fault periods. The calculation of the adaptive voltage reference of where Ipv is the average current reference flowing through the
DC-link Udcref
in mode 2 is based on the difference between Ppv mpp active switch during one switching period. Kp_pv and Ki_pv are the
parameters of the PI controller. In order to perform a fast PV output
and Pg , which can be derived from (27a) to (27b)
max
power de-loading in the fault, one-cycle control (OCC) is applied
in the ASVSC, which is illustrated in Fig. 10. IL and Spv are the
1
Ec = C (U ref 2 − Udc
norm2 mpp
) = Ketf (Ppv − Pgmax) (27a) current in the input inductor and gate signal of ASVSC,
2 dc dc
respectively.
Three advantages can be highlighted with the solution-oriented
ref 2 control: (i) accurate ACI is achieved since the current output of
K t (Pmpp − Pgmax) (27b)
2
Udc = U norm +
dc Cdc e f pv GSVSC is controlled by the fixed references, which can improve
the controllability of the system and the reliability of voltage
where Ec is the energy stored in the DC-link capacitors. Cdc is the support. (ii) Stable ACI is realised since the power variation caused
total DC-link capacitance. tf is the fault duration. Ke is a slack by environmental changes is buffered by the DC-link capacitors.
coefficient to regulate the percentage of extra energy stored in the (iii) The unnecessary power losses can be effectively reduced, as
DC-link capacitors. In this regard, the energy storage capability of shown in Fig. 11.
DC-link capacitors can be adjusted as per the actual needs and the
overvoltage circumstance can be avoided with a suitable selection 4.4 Mode 3: LVRT operation with enhanced ACI
of Ke.
To absorb the double-line frequency power in the DC-link During the LVRT period, if the MPP in the current environment
capacitors, Cdc can be calculated in (28a). By substituting (28a) condition is smaller than the maximum allowable output power of
GSVSC, control mode 3 will be activated. The control diagram and
into (27b), (28b) can be obtained
energy flow of this mode are illustrated in Fig. 12.
In this mode, all the solar energy harvested with MPPT control
is injected to the grid. After providing specified RCI as per the grid

2732 IET Renew. Power Gener., 2020, Vol. 14 Iss. 14, pp. 2727-2737
© The Institution of Engineering and Technology 2020
 17521424, 2020, 14, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-rpg.2019.1101 by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [26/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Table 2 Simulation parameters Ud+ Uq+ Ud− Uq−
P
Component Specifications
Q Uq+ −Ud+ Uq− −Ud− Id+
maximum input power 100 kW (at STC)
− − +
input capacitor Cin = 10 μF Pc2 3 Ud Uq Ud Uq+ Iq+
= (34)
output filter Lf = 510 μH, Rf = 660 μΩ, LR filter Ps2 2 Uq− −Ud− −Uq+ Ud+ Id−
DC-link capacitor Cdc = 15 mF Qc2 Uq− −Ud− Uq+ −Ud+ Iq−
equivalent line impedance R = 150 mΩ, L = 471 μH (normal) Qs2 −Ud− −Uq− Ud+ Uq+
MPPT control period 100 μs
PCC voltage 220 Vrms, 50 Hz (three-phase) where Ud+, Ud−, Uq+ and Uq− are the positive and negative
nominal DC-link voltage 750 V components of PCC voltage in the synchronous coordinate system.
switching frequency 20 kHz In order to (i) guarantee the accurate and stable ACI and (ii) avoid
nominal GSVSC current 151.5 Arms (phase current) DC-link voltage variation, which compromises the ACI capability
provided by the capacitors, pi should be regulated to be constant,
i.e.

Pc2 = Ps2 ≡ 0 (35)

Besides, the minimum DC-link voltage in (32) should be

reformulated as

2 3
min
Udc = max Ua − Ub , Ua − Uc , Uc − Ub (36)
3

where Ua, Ub, and Uc denote the vectors of three-phase peak

voltage at PCC.

5 Case studies
Fig. 13 P–V and I–V curves under different temperature and irradiance The model of a BIPV system is constructed in MATLAB-Simulink
conditions with the specified parameters shown in Table 2. The maximum
power of a PV module is 250 W under the STC (irradiance at 1000
code requirements, GSVSC still remains a certain power output W/m2, temperature at 25°C). 40 × 10 PV modules make up a PV
capacity. Since ACI is more effective on voltage support in the array with maximum power of 100 kW. The voltage and current at
fault of LVDN, especially for a severe PCC voltage drop, the DC- the MPP are 251.2 V and 398.8 A, respectively. P–V and I–V
norm
link voltage reference is regulated to be lower than Udc , which characteristics of this PV array can be plotted as shown in Fig. 13.
can generate more ACI from the electrical potential energy of DC- Different fault scenarios are simulated with respect to different
link capacitors. The enhanced active current output of GSVSC iden environmental conditions to verify the effectiveness of the
can be expressed as proposed LVRT control.
As the voltages of LVDN are relatively vulnerable, the
Cdc(U norm
2
− U ref
2 proposed LVRT control is programmed with voltage response
dc )
∫ idendt =
Ec
ph ⋅ Ug
=
dc
2ph ⋅ Ug
(31) capability in case of slight voltage drops at PCC. In this mode 1,
the MPPT control is always activated for delivering the maximum
PV power, as shown in Fig. 14c. In the meantime, the GSVSC is
As illustrated in Fig. 12b, the DC-link capacitors will release controlled to provide the desired iq for reducing the voltage
certain electrical potential energy to enhance the ACI during the fluctuations. As illustrated in Figs. 14b and d, 0.06 p.u. iq is
transient voltage fault. To avoid over-modulation with SPWM, the
ref injected to the grid at the time periods of 0.1–0.2 and 0.5–0.8 s
Udc should satisfy with the proposed control, which stabilises the PCC voltage at the
nominal reference. During the time periods of 0.2–0.4 and 0.4–0.5
ref
2 2Ug in three phase system s, the PCC voltage falls to 0.95 and 0.98 p.u., respectively. In these
(32)
min
Udc ≥ Udc = cases, 0.2 p.u. iq, which is the upper limit calculated, is generated
2Ug in single phase system
with the proposed control. This can boost the PCC voltage to 213
and 218 V, respectively. Compared with the control without voltage
4.5 Further discussion on LVRT control under the response, about 4 V of the PCC voltage variations can be reduced.
asymmetrical voltage drop A PCC voltage dip to 0.6 p.u. occurs at t = 0.3 s in Fig. 15,
In the case of asymmetrical voltage drop at PCC, the instantaneous when the maximum PV power output is larger than the remaining
real power pi and imaginary power qi can be expressed as active power capacity of GSVSC. Thus, LVRT operation with
accurate ACI control is activated (mode 2). After providing 0.8 p.u.
reactive current as per grid codes, the GSVSC with traditional
pi P + Pc2cos(2ωt) + Ps2sin(2ωt)
= (33) control only remains at 0.6 p.u. ACI, which corresponds to around
qi Q + Qc2cos(2ωt) + Qs2sin(2ωt) 0.36 p.u. Pg, as shown in Figs. 15c and d. However, GSVSC output
current of 1.1 p.u. is adopted in the proposed control for improving
where P¯ and Q¯ represent the average active and reactive power of ACI capabilities. By calculation, around 0.45 p.u. Pg is generated
the GSVSC output. Pc2, Ps2, Qc2 and Qs2 denote the double-line in the fault period, as illustrated in Figs. 15c and d. In addition,
frequency pulsating powers. Considering that the zero-sequence compared with the constant DC-link control in the traditional
current is zero in the three-phase three-wire system, only four LVRT method, the proposed one increases the udc from 750 to 800
control freedoms are available for the GSVSC, namely, positive V in the LVRT operation, as shown in Fig. 15b. In this operation,
and negative components of current in the synchronous coordinate more PV energy that is wasted with traditional control will be
system (Id+, Id−, Iq+ and Iq−). The steady-state model of the GSVSC stored in the DC-link capacitors during the fault period, which can
can be expressed as potentially increase the energy conversion efficiency. When PCC
voltage is restored at t = 0.7 s, Pg begins to recover, as illustrated in

IET Renew. Power Gener., 2020, Vol. 14 Iss. 14, pp. 2727-2737 2733
© The Institution of Engineering and Technology 2020
 17521424, 2020, 14, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-rpg.2019.1101 by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [26/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
references, which are calculated by the optimiser instead of the
DC-link voltage loop. In this way, the GSVSC output active power
is stable and irrespective to the environmental disturbances, as
shown in Fig. 15d, where the disturbances are absorbed by the DC-
link capacitors as shown in Fig. 15b. This will potentially improve
the stability and controllability of the system.
The performances of the traditional de-loading-based LVRT
control [16], DC-chopper-based LVRT control [8] and without
LVRT control are compared with the proposed one to further
demonstrate the advantages of proposed LVRT control in mode 2.
In this case, the PCC voltage falls to 0.7 p.u. at the time period of
0.5–1.7 s. The simulated waveforms are shown in Fig. 16. The
waveforms of the systems without LVRT control, with DC-
chopper, with traditional de-loading control and with proposed
LVRT control are labelled in green, yellow, blue and red,
respectively. By calculation, the maximum PV power output is
larger than the maximum permissive output active power of
GSVSC. As a result, the PV output power must be curtailed. As
illustrated in Fig. 16d, for the system without LVRT control, the
excessive solar energy has resulted in the over-current of GSVSC.
As shown in Fig. 16b, for the system with DC-chopper, the
excessive energy is dissipated on the power resistors, which causes
some unnecessary energy waste in the fault. As illustrated in
Fig. 16c, RCI is performed based on the grid code in the typical de-
loading and proposed LVRT control. By comparison, the proposed
control method injects more active current into the PCC and
increases the DC-link voltage to reserve energy in the fault. After
Fig. 14 Waveforms of the PV generation system with PCC voltage the fault clearance at the time instant of 1.7 s, the energy reserved
response activated (mode 1) in the capacitors will be rapidly released and injected into the grid,
(a) DC-link voltage, (b) GSVSC output current, (c) PV output power, (d) PCC phase which results in temporarily larger active current and power with
voltage the proposed control than the traditional de-loading control, as
shown in Figs. 16b and d.
As illustrated in Fig. 17c, the PCC voltage drops to 0.45 p.u.
during the time period of 0.5–1.06 s when the solar irradiance is
215 W/m2. Before the voltage dip, volt-var control is activated in
normal operation with the proposed control (mode 1). In this case,
reactive current of 21 A is produced to stabilise the PCC voltage at
1 p.u., while the generator with traditional control will not provide
any reactive current for PCC voltage regulation in normal
operation, as shown in Figs. 17b and d. During the fault period, the
line impedance is set as 100 mΩ and 70 μH. With traditional
control, the GSVSC generates 1 p.u. reactive current while the
active power is 0 p.u., as shown in Fig. 17b. However, since the
GSVSC current in the proposed control is temporarily set at 1.1
p.u, the 0.46 p.u. active current can be still injected. In this way, the
MPPT control can be maintained and the PCC voltage can be
effectively boosted. Compared with the effects of traditional
control in Fig. 17d, the PCC voltage is 8 V's higher with the
proposed control in mode 3. Besides, the DC-link voltage with the
proposed control is decreased to 350 V instead of being kept
constant, as shown in Fig. 17a. This operation releases the energy
of the capacitors and helps enhance the ACI and support PCC
voltage during the time period of 0.5–0.6 s, as shown in Figs. 17b
and d. After fault clearance, the proposed controller will restore
from LVRT control mode 3 to normal operation (mode 1).
Fig. 15 Waveforms of the LVRT operation with accurate ACI (mode 2 with
changing environments)
(a) Fluctuating irradiance and temperature, (b) DC-link voltage, (c) GSVSC output
6 Hardware validation
active power, (d) GSVSC output active power (zoom in from 0.3–0.7 s) To verify the feasibility of the proposed LVRT control in the
practical applications, a two-stage micro inverter is adopted for the
Fig. 15c. Simultaneously, Udc ref
is reset to be its nominal value of experiment. The ASVSC is implemented with MPPT control in
750 V. The post-fault recovery of Pg is accelerated with the normal operation, while the GSVSC is controlled for grid-side
current regulation. Second-order generalised integrator PLL is
proposed control due to the fast release of electrical potential
adopted to trace the phase of PCC voltage. The detailed
energy in the DC-link capacitors and the more rapid increase of Idref configuration is summarised in Table 3. The experimental
induced by larger error in the DC-link control loop. Furthermore, in environment is presented in Fig. 18.
this case, the temperature and irradiance are slightly fluctuating The behaviour of the PV panel is emulated by the Chroma
around 25°C and 800 W/m2 during the fault. With the traditional programmable DC source 62150H. The AC side of the inverter is
input-oriented control structure, the GSVSC output active power connected to the Chroma programmable AC source 61511, which
fluctuates following the variation of the environmental conditions, can generate time-based voltage dips to emulate the grid fault. The
as shown in the zoomed-in figure (see Fig. 15d) of Fig. 15c. control algorithm is embedded in the digital signal processor
However, with the proposed solution-oriented control structure, the TMS320F28035. When the grid is working under normal
GSVSC output currents are accurately controlled at the fixed

2734 IET Renew. Power Gener., 2020, Vol. 14 Iss. 14, pp. 2727-2737
© The Institution of Engineering and Technology 2020
 17521424, 2020, 14, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-rpg.2019.1101 by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [26/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Table 3 Hardware circuit parameters
Component Specifications
solar input CIGS panel: Pmpp = 95 W, 1200*600*19
mm or Chroma 62150H: Pmpp = 100 W
input capacitor Cin = 22 μF
DC-link capacitor Cdc = 630 μF
output filter Lf = 470 μH, Cf = 4.7 μF, LCL filter
output load 100 Ω
grid voltage (RMS) Chroma 61511: 110 V, 60 Hz
nominal DC-link voltage 275 V
switching frequency 50 kHz
state machine frequency 1 kHz
current sensor LEM 6-NP 2.5 V ∓0.625 V for ∓3 A

Fig. 18 Experiment environment in the laboratory

Fig. 16 Waveforms of the LVRT operation with different LVRT controllers

(mode 2)
(a) Voltage of DC-link, (b) GSVSC output active power, (c) GSVSC output reactive
power, (d) GSVSC output active current

Fig. 19 Alternative experimental inputs and PV output power curtailment

Fig. 20 PCC voltage and GSVSC output current

generated. In this case, the MPP of PV generation is at 100 W. By

calculation, LVRT control mode 2 is selected, where about 36% of
PV output power should be curtailed to avoid the overcurrent and
overvoltage conditions, as shown by the red point in Fig. 19. The
PCC voltage and GSVSC output current are shown in Fig. 20.
During the LVRT period, specified reactive current of 0.8 p.u. as
per grid code is supplied. Meanwhile, the apparent output current
Fig. 17 Waveforms of the LVRT operation with enhanced ACI (mode 3) of GSVSC is temporarily set at 1.1 p.u. of rated GSVSC current for
(a) DC-link voltage, (b) GSVSC output current, (c) Instantaneous PCC voltage, (d) enhancing the ACI capability, as shown in Fig. 20. As illustrated in
PCC phase voltage RMS Fig. 21, the PV array is operated at the MPP in the normal
operation. When the fault occurs, the PV output current is reduced
conditions, the maximum power point is constantly tracked, as to curtail the PV output power. Meanwhile, the DC-link voltage is
shown in Fig. 18. increased to 25 V higher for storing extra energy in the capacitors.
When the grid voltage dip occurs, the GSVSC output power is After the fault clearance, the MPPT control will be resumed and
restricted based on the current rating of power electronic the DC-link voltage will be regulated back to the nominal value.
components and grid codes. First, a PCC voltage dip to 0.6 p.u. is

IET Renew. Power Gener., 2020, Vol. 14 Iss. 14, pp. 2727-2737 2735
© The Institution of Engineering and Technology 2020
 17521424, 2020, 14, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-rpg.2019.1101 by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [26/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
This work was supported by Hong Kong Research Grants Council
(RGC) under Theme-based Research Project T23-701/14N and by
National Natural Science Foundation of China under grant no.
71971183.

9 References
[1] Lai, C.S., Jia, Y., Lai, L.L., et al. ‘A comprehensive review on large-scale
photovoltaic system with applications of electrical energy storage’, Renew.
Sust. Energy Rev., 2017, 78, pp. 439–451
[2] Marinopoulos, A., Papandrea, F., Reza, M., et al. ‘Grid integration aspects of
large solar Pv installations: Lvrt capability and reactive power/voltage support
requirements’. 2011 IEEE Trondheim PowerTech, Trondheim, Norway, 2011
[3] Hasanien, H.M.: ‘An adaptive control strategy for low voltage ride through
capability enhancement of grid-connected photovoltaic power plants’, IEEE
Trans. Power Syst., 2016, 31, (4), pp. 3230–3237
Fig. 21 DC-link capacitor voltage and PV output current [4] Dietmannsberger, M., Grumm, F., Schulz, D.: ‘Simultaneous implementation
of LVRT capability and anti-islanding detection in three-phase inverters
connected to low-voltage grids’, IEEE Trans. Energy Convers., 2017, 32, (2),
pp. 505–515
[5] Lee, C.T., Hsu, C.W., Cheng, P.T.: ‘A low-voltage ride-through technique for
grid-connected converters of distributed energy resources’, IEEE Trans. Ind.
Appl., 2011, 47, (4), pp. 1821–1832
[6] Saad, N.H., El-Sattar, A.A., Mansour, A.E.-A.M.: ‘Improved particle swarm
optimization for photovoltaic system connected to the grid with low voltage
ride through capability’, Renew. Energy, 2016, 85, pp. 181–194
[7] Yong-Sih, W., Chia-His, C., Yaow-Ming, C., et al.: ‘A current control strategy
for three-phase Pv power system with low-voltage ride-through’. Ninth IET
Int. Conf. on Advances in Power System Control, Operation and Management
(APSCOM 2012), Hong Kong, China, 2012
[8] Al-Shetwi, A.Q., Sujod, M.Z., Blaabjerg, F.: ‘Low voltage ride-through
capability control for single-stage inverter-based grid-connected photovoltaic
power plant’, Sol. Energy, 2018, 159, pp. 665–681
[9] Mohammadi, P., Eskandari, A., Milimonfared, J., et al.: ‘Lvrt capability
enhancement of single-phase grid connected Pv array with coupled
supercapacitor’. 2018 Ninth Annual Power Electronics, Drives Systems and
Technologies Conf. (PEDSTC), Tehran, Iran, 2018
[10] Saadat, N., Choi, S.S., Vilathgamuwa, D.M.: ‘A statistical evaluation of the
capability of distributed renewable generator-energy-storage system in
providing load low-voltage ride-through’, IEEE Trans. Power Deliv., 2015,
30, (3), pp. 1128–1136
[11] Wang, M., Mok, K., Tan, S., et al.: ‘Multifunctional Dc electric springs for
improving voltage quality of Dc grids’, IEEE Trans. Smart Grid, 2018, 9, (3),
pp. 2248–2258
Fig. 22 Experimental waveform results of the PV micro-inverter with [12] Lyu, X., Xu, Z., Zhao, J., et al.: ‘Advanced frequency support strategy of
proposed control in mode 3 photovoltaic system considering changing working conditions’, IET Gener.
Transm. Distrib., 2018, 12, (2), pp. 363–370
[13] Yang, Y., Blaabjerg, F., Zou, Z.: ‘Benchmarking of grid fault modes in single-
In the second case, the MPP of PV generation is 45 W. A PCC phase grid-connected photovoltaic systems’, IEEE Trans. Ind. Appl., 2013,
voltage dip to 0.7 p.u. is generated. By calculation, the GSVSC still 49, (5), pp. 2167–2176
retains sufficient capacity for active power output after 0.6 p.u. [14] Yang, Y., Blaabjerg, F., Wang, H.: ‘Low-voltage ride-through of single-phase
transformerless photovoltaic inverters’, IEEE Trans. Ind. Appl., 2014, 50, (3),
RCI is provided. Therefore, control mode 3 is selected, where the pp. 1942–1952
MPPT can be maintained, as shown in Fig. 22. In this case, the [15] Parvez, M., Elias, M.F.M., Rahim, N.A., et al.: ‘Current control techniques
DC-link voltage is adaptively reduced from 275 to 180 V, which for three-phase grid interconnection of renewable power generation systems:
will release certain energy from the capacitors to provide more ACI a review’, Sol. Energy, 2016, 135, pp. 29–42
[16] Yang, F., Yang, L., Ma, X.: ‘An advanced control strategy of PV system for
during the fault period. The transient responses of GSVSC output low-voltage ride-through capability enhancement’, Sol. Energy, 2014, 109,
current, DC-link voltage and PV output current in the transition pp. 24–35
between the normal operation (mode 1) and LVRT operation (mode [17] Bae, Y., Vu, T.K., Kim, R.Y.: ‘Implemental control strategy for grid
3) are illustrated in Fig. 22. stabilization of grid-connected Pv system based on German grid code in
symmetrical low-to-Medium voltage network’, IEEE Trans. Energy Convers.,
2013, 28, (3), pp. 619–631
7 Conclusion [18] Jain, S., Easley, M., Shadmand, M.B., et al.: ‘Decoupled active and reactive
power predictive control of impedance source microinverter with lvrt
The ACI-based LVRT control is necessarily required for the PV capability’. 2018 IEEE Power and Energy Conf. at Illinois (PECI),
generator in the LVDN. The effects of the ACI control on PCC Champaign, Illinois, US., 2018
[19] Tang, C.Y., Chen, Y.T., Chen, Y.M.: ‘Pv power system with multi-mode
voltage boosting is mathematically analysed in this paper. Based on operation and low-voltage ride-through capability’, IEEE Trans. Ind.
this, a novel LVRT control scheme is proposed. This control Electron., 2015, 62, (12), pp. 7524–7533
scheme utilises a supervisory optimiser to decide the different [20] Varma, R.K., Siavashi, E.M.: ‘PV-STATCOM - a new smart inverter for
control modes according to the various environmental and fault voltage control in distribution systems’, IEEE Trans. Sustain. Energy, 2018,
9, (4), pp. 1–1
conditions. Adaptive DC-link voltage regulation with the self- [21] Oon, K.H., Tan, C., Bakar, A.H.A., et al.: ‘Establishment of fault current
adjusting control topology is adopted. In the control mode 1, characteristics for solar photovoltaic generator considering low voltage ride
voltage-response capability is enabled to stabilise the PCC voltage through and reactive current injection requirement’, Renew. Sust. Energy Rev.,
in LVDN. In the control mode 2, more energy is reserved in the 2018, 92, pp. 478–488
[22] Chua, K.H., Wong, J., Lim, Y.S., et al. ‘Mitigation of voltage unbalance in
capacitor for accelerating post-fault power recovery. Besides, the low voltage distribution network with high level of photovoltaic system’,
solution-oriented control can achieve an accurate and stable Energy Procedia, 2011, 12, pp. 495–501
GSVSC output against the environmental changes. In the control [23] Biyik, E., Araz, M., Hepbasli, A., et al. ‘A key review of building integrated
mode 3, enhanced ACI is achieved for grid voltage support. As photovoltaic (BIPV) systems’, Eng. Sci. Technol., Int. J., 2017, 20, (3), pp.
833–858
verified by the simulation and experiment results, the proposed [24] Mundada, A.S., Nilsiam, Y., Pearce, J.M., ‘A review of technical
control can achieve more stable and effective operation in fault requirements for plug-and-play solar photovoltaic microinverter systems in
ride-through and PCC voltage boosting. the United States’, Sol. Energy, 2016, 135, pp. 455–470
[25] Choudhury, S.R., Gupta, A., Anand, S.: ‘Simulation of low voltage ride
through scheme for inverters connected to distribution system with high R/X
8 Acknowledgments ratio’. 2016 Tenth Int. Conf. on Compatibility, Power Electronics and Power
Engineering (CPE-POWERENG), Bydgoszcz, Poland, 2016

2736 IET Renew. Power Gener., 2020, Vol. 14 Iss. 14, pp. 2727-2737
© The Institution of Engineering and Technology 2020
 17521424, 2020, 14, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-rpg.2019.1101 by Egyptian National Sti. Network (Enstinet), Wiley Online Library on [26/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
[26] He, Y., Xu, Z., Li, Y.: ‘A novel control scheme for enhancing low voltage ride [27] Kinces, F.: ‘Voltage sag indices and statistics’ (Chalmers University of
through capability of solar generation’. 2017 China Int. Electrical and Energy Technology, Sweden, 2004)
Conf. (CIEEC), Beijing, China, 2017

IET Renew. Power Gener., 2020, Vol. 14 Iss. 14, pp. 2727-2737 2737
© The Institution of Engineering and Technology 2020