Voltage control of grid - connected PV system facing voltage sags

s, most of the worldwide production of 
energy is ensured by fossil sources. The consumption 
of energy from the fossil sources faces to the 
exhaustion of these sources, the climate changes, and 
the emission of CO2. This result to the exponential 
growth of implementation of renewable energy 
generation systems. Among the renewable energy 
sources, photovoltaic solar energy (PV) is a promising 
source. In the current economic context (tariffs of 
purchase, tax credit, national or regional aids, etc.), the 
number of requests for connection of PV is increasing 
in an exponential way. There are interactions between 
PV system and power network. The behavior of PV 
systems connected to the distribution grid has been 
reported by many groups [1], [2]. PV system can have 
a significant impact on the operation of the electrical 
system or may cause malfunctions [3]. Besides, a 
disturbance in the network could have an important 
impact on PV system operation. Voltage sags are 
acknowledged to be one of the major power quality 
disturbances. A low voltage caused by voltage sag can 
provoke the PV system disconnection due to the 
decoupling protection [4], [5]. 
Currently, almost inverters integrated with a 
classic regulation (P/Q regulation) cannot control 
voltages. In the worst case, if the voltage sags are very 
deep (i.e. the voltage can drop practically to a few V 
for a few hundred milliseconds), it is impossible to 
maintain the PV systems in such conditions. But in less 
* Corresponding author: Tel.: (+84) 943842803 
Email: tung.leduc@hust.edu.vn 
severe voltage sags, for example in case of a 
momentary faults or short circuits on the adjacent 
feeder, the maintenance of PV systems connected to 
the grid is achievable. The connection is maintained by 
a control system integrated into the inverters which 
allows the voltage at the connection point greater than 
an accepted threshold. This threshold voltage depends 
on the rules of each country and each type of the grid. 
In this paper, the proposed method relates to the 
integration of an "intelligent" control/command 
system in the PV inverters. This control system allow 
participation in maintaining the voltage at the 
connection point during a voltage dip or voltage 
disturbances on the grid and participate in the optimal 
regulation of the grid voltage by using a “auto-adaptive 
voltage control”. It permits to increase the rate of 
insertion, the performances, and the flexibility of 
operation of PVs in an intelligent and adaptive way. 
Section 2 of the paper presents firstly building an 
auto-adaptative voltage control for PV system. 
Secondly, section 3 will evaluate the effectiveness of 
the proposed control through Matlab/Simulink 
software. Finally, the conclusions and perspectives 
will be presented in section 4. 
2. Development of auto-adaptative voltage control 
for PV system 
PV inverters can be operated with different 
control schemes according to their operation mode 
[6][7]. Three types of reactive power compensation 
schemes can be applied for grid-tied inverters: an 
Journal of Science & Technology 144 (2020) 001-005 
2 
active and reactive power control scheme (P/Q-
control); a control of active power and power factor 
(P/PF-control); a control of active power and voltage 
(P/V-control). The voltage/frequency (V/f) control 
scheme is generally used for grid-forming inverters 
[8]. 
For the P/Q control scheme, the active and 
reactive power outputs of PV are fixed to set-point 
values Psetpoint and Qsetpoint. Similarly, for the P/PF 
control scheme, the active power and the power factor 
are fixed to set-point values by changing the reactive 
power in order to maintain a constant power factor. 
For the V/f control scheme, the voltage and the 
frequency are fixed to set-point values Vsetpoint and 
fsetpoint. The active and reactive powers are controlled 
in order to maintain a constant voltage and frequency. 
A “frequency-active power” and “voltage-reactive 
power” droop is used. 
The energy source is represented by a PV-power 
source. The authors suppose that the dynamic of the 
entire system of the up-stream PV system (primary 
source) could be represented by a first-order response 
which enables to change the time constant according 
to the characteristics of the primary source. In addition 
to this dynamic part of the PV’s characteristics, the 
operation limits of active and reactive power are 
included. Only the described parameters, enhanced by 
primary energy availability, e.g. with variations of 
solar irradiation, define the characteristics of PV. 
Fig 1. P/Q control scheme. 
The operation principle of the P/Q control 
scheme is described as Fig. 1. From the current and 
voltage measured at the connection point, the power 
(Pmes and Qmes) and the corresponding voltage are 
determined. These powers will be adjusted by two 
proportional-integral (PI) controllers. The difference 
between the setpoint power Psetpoint and Qsetpoint and the 
measured power Pmes and Qmes will be handled by the 
ratio (Kp) and the integral (Ki/p). From the output 
power through the PI, the desired current is calculated 
by the Park transformation: 
⎩
⎪
⎨
⎪
⎧𝐼𝐼𝑑𝑑 = 2(𝑃𝑃.𝑉𝑉𝑑𝑑 + 𝑄𝑄.𝑉𝑉𝑞𝑞)3(𝑉𝑉𝑑𝑑2 + 𝑉𝑉𝑞𝑞2)
𝐼𝐼𝑞𝑞 = 2(𝑃𝑃.𝑉𝑉𝑞𝑞 − 𝑄𝑄.𝑉𝑉𝑑𝑑)3(𝑉𝑉𝑑𝑑2 + 𝑉𝑉𝑞𝑞2) 
(1) 
where Id, Iq and Vd, Vq are Park transformation of 
currents and voltages at the output of the inverter; P 
and Q are the reference power (normally Q=0). 
Fig 2. Auto-adaptive voltage controller of PV 
Fig. 2 presents the scheme of auto-adaptive 
voltage controller. The model of PV with this regulator 
is developed in three phases. It composes a P/Q control 
and a P/V control. In P/V control mode, the voltage 
setpoint is changed in an auto-adaptive way by using a 
fuzzy logic module or droop control. The change of 
setpoint voltage values is carried out, correlatively 
with the operation and location of PV, by respecting 
reactive power limits of each PV. 
Three operating modes of the control are 
possible. They correspond to three possible states: 
- Normal state: where the voltage is located 
inside a window of “desired” voltage (Vmin_desired ≤ V 
≤Vmax_desired). In this state, PV is in P/Q control (or 
PF/VAR control). 
- Disturbed state: where voltage leaves the 
desired limits (V> Vmax_desired or V< Vmin_desired). The 
goal of the adaptive control is to maintain, within the 
limits of the system, the voltage between these fixed 
values. Thus, under disturbed conditions, PV 
commutates in voltage regulation mode (P/V control). 
Here, only reactive power is used to control voltage at 
the PV connection point. The voltage set point is set at 
Vmin_desired or Vmax_desired according to whether the 
network voltage profile is too low or too high. If PV is 
in reactive power limitation (Q=Qmin or Q=Qmax), it 
cannot ensure any more the control in the desired 
voltage. The voltage moves and reaches critical state 
when voltage admissible limits are crossed. 
Journal of Science & Technology 144 (2020) 001-005 
3 
- Critical state: where the voltage is out of the 
admissible limits (V>Vmax_admissible or V< Vmin_admissible, 
in France Vmax_admissible=1.1 pu, Vmin_admissible=0.9 pu) 
and, as previously explained, PV cannot act any more 
by compensation of reactive power. In the critical state 
regulation of active power becomes necessary. So, PV 
commutates in active power regulation mode (Mode 
P). It means that PV changes active power generation 
in order to bring back the voltage in the admissible 
values. 
The change of control operating mode is 
automatic and auto adaptive. Moreover, the proposed 
method only uses voltage or current measurements at 
the connection point and does not need any 
communication link with DNO or other PVs. 
The control changes in an adaptive way the 
desired voltage value. The desired voltage depends on 
the voltage at the PV connection point, and the level of 
reactive power used compared to the Qlimit of each PV. 
The calculation of the desired limit is based on fuzzy 
logic as shown in equation (2), where 
Vmesure(pu)=Vmesure/Vnominale and 
Qmesure(pu)= Qmesure/Qlimit: 
𝑉𝑉max = 𝑉𝑉max_𝑎𝑎𝑑𝑑𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 − 𝑉𝑉𝑛𝑛𝑛𝑛𝑎𝑎𝑎𝑎𝑛𝑛𝑎𝑎𝑎𝑎 
𝑉𝑉min = 𝑉𝑉𝑛𝑛𝑛𝑛𝑎𝑎𝑎𝑎𝑛𝑛𝑎𝑎𝑎𝑎 − 𝑉𝑉max_𝑎𝑎𝑑𝑑𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 
𝑉𝑉max_desired = 𝑉𝑉𝑛𝑛𝑛𝑛𝑎𝑎𝑎𝑎𝑛𝑛𝑎𝑎𝑎𝑎 + 𝐶𝐶.𝑉𝑉max 
𝑉𝑉min_desired = 𝑉𝑉𝑛𝑛𝑛𝑛𝑎𝑎𝑎𝑎𝑛𝑛𝑎𝑎𝑎𝑎 − 𝐶𝐶.𝑉𝑉min 
(2) 
After identifying the desired voltage in equation 
(2), the authors can calculate the reactive power 
required for voltage regulation. The coefficient C is 
identified by fuzzy logic as in the Fig. 3. 
Fig. 3. Calculation of the coefficient C by fuzzy logic 
Adaptive limits allow all PVs to contribute to 
voltage profile without communication system, even 
PVs located on not critical voltage feeders. In fact, the 
more the voltage measured is closed to 1pu the more 
voltage desired window will be narrow. This window 
moves according to the quantity of reactive power 
provided or absorbed compared with physical limits of 
the PV considered. More the contribution of reactive 
power is important more the window of voltage will 
increase by respecting the limits 
(Vmin_admissible≤Vmin_desired≤Vmax_desired≤Vmax_admissible). 
3. Simulation Results 
This section evaluates the effectiveness of the 
proposed control through Matlab/Simulink software. 
In order to show the capacity of the proposed local 
voltage control of PVs, a medium voltage (MV) grid 
(Fig. 4) is used for the study. This MV network is 
supplied by a 110/22kV, 40 MVA transformer. It 
composes 53 and a 1000kW PV system. 
Fig. 4. Medium voltage grid studied. 
It assumes that the PV systems are disconnected 
facing a voltage dip U ≤ 85% of the nominal voltage 
for PV systems connected in MV grid (such as French 
requirements [8]), this threshold voltage depends on 
the rules of each country and each type of the grid. A 
voltage sag caused by a momentary fault or a short 
circuit on the adjacent feeder appears at time t = 0.5s 
for 500m. For PV system, two types of control are 
used: classical control (P/Q control) and Auto-
adaptive voltage control. 
With a short-circuit on the adjacent feeder L_05, 
in case of operating in P/Q control mode, Fig. 5 
illustrates the power of the 1000kW PV system and the 
voltage at the connection point 
The simulation results show that the voltage at 
the connection point of PV systems exceeds the limit 
voltage normalized by the assumed recommendation 
(0.85pu), then this PV system can be disconnected by 
their associated protections and the reactive power 
remains zero at the time of voltage sags. The reactive 
power of PV systems should therefore be modified to 
keep the voltage within the admissible limits. 
In case of using the auto-adaptative voltage 
control, the grid structure, parameters and scenario are 
identical to those of the above study. The Fig. 6 shows 
that the PV inverters participate in the voltage 
regulation. For the P/Q control regulator, the reactive 
power always remains zero. Facing voltage sags, the 
PV system produces reactive power to restore the 
voltage in the admissible threshold by the decoupling 
protections. Therefore, the voltage at the connection 
node (with auto adaptative voltage control) is greater 
Journal of Science & Technology 144 (2020) 001-005 
4 
Fig 5. PV system powers connected to the MV grid and voltage variation with P/Q control. 
Fig 6. PV system powers and voltage variation with Auto-adaptative voltage control. 
than 0.85pu. In this case, this PV system remains 
connected to the grid. 
The proposed voltage control is so capable to 
maintain the PV system connected facing the voltage 
sags and voltage disturbances on the grid. The degree 
of reactive production or absorption depends 
ondifferent factors such as the connection location, the 
reactive supply capacity of PVs, grid voltage profile, 
and grid parameters. 
4. Conclusion 
This paper presents a local voltage control based 
on auto-adaptive voltage control integrated into PV 
inverter, this control uses local information. Base on 
absorption/production of reactive power, the voltage of 
PV systems at the connection is so improved and 
restored in the admissible threshold during a voltage 
sags (in case of momentary faults or short circuits on 
the adjacent feeder). A lot of advantages are brought 
by using such inverter control such as reducing 
connection costs, increase the rate of insertion, the 
performances of operation of PV systems, the power 
quality of grid and without reducing the efficiency of 
the decoupling device of the inverters. 
References 
[1] Y. Xue, M. Manjrekar, C. Lin, M. Tamayo and J. N. 
Jiang, Voltage stability and sensitivity analysis of grid-
connected photovoltaic systems, IEEE Power and 
Energy Society General Meeting, Detroit, MI, USA, 
(2011), 1-7. 
[2] Naomi Stringer, Navid Haghdadi, Anna Bruce, Jenny 
Riesz and Iain MacGill, Observed behavior of distributed 
photovoltaic systems during major voltage disturbances 
and implications for power system security, Applied 
Energy, 260-114283(2020), 1-13. 
[3] M.Q. Duong, N.T.N. Tran, G.N. Sava, S. Leva, M. 
Mussetta, The Impact of 150MWp PhoAn Solar 
Photovoltaic Project into Vietnamese QuangNgai - Grid, 
International Conference and Exposition on Electrical 
And Power Engineering (EPE), Romania, (2018), 498-
502. 
[4] C. Le Thi Minh, T. Tran-Quoc, S. Bacha, C. Kieny, P. 
Cabanac, D. Goulielmakis, C. Duvauchelle, Behaviours 
of photovoltaic systems connected to MV network 
Journal of Science & Technology 144 (2020) 001-005 
5 
during faults, 26th EUPVSEC, Humburg, Germany, 
(2011), 4221 – 4226. 
[5] A Mahmud. M.A, Hossain. M.J, Pota. H.R, Voltage 
Variation on Distribution Networks With Distributed 
Generation: Worst Case Scenario, IEEE Systems 
Journal, 8(2014), 1096 – 1103. 
[6] Björn Lindgren, Topology for Decentralised Solar 
Energy Inverters with a Low Voltage AC-Bus, European 
conference on power electronics and applications, 
Lausanne (Switzerland), (1999), 1-10. 
[7] Sarina Adhikari; Fangxing Li; Huijuan Li, P-Q and P-V 
Control of Photovoltaic Generators in Distribution 
Systems, IEEE Transactions on Smart Grid, 6(2015), 
2929 – 2941. 
[8] T.Tran-Quoc, G.Rami, A.Almeida, N.Hadjsaid, 
J.C.Kieny, J.C.Sabonadiere, Méthode et dispositif de 
régulation pour un dispositif de production décentralisée 
d’énergie, et installation comportant au moins deux 
dispositifs de production dotes dudit dispositif de 
régulation, Brevet d’invention international, (2005), 1-
50.