Aalborg Universitet
Reactive Power Injection Strategies for Single-Phase Photovoltaic Systems
Considering Grid Requirements
Yang, Yongheng; Wang, Huai; Blaabjerg, Frede
Published in:
IEEE Transactions on Industry Applications
DOI (link to publication from Publisher):
10.1109/TIA.2014.2346692
Publication date:
2014
Document Version
Early version, also known as pre-print
Link to publication from Aalborg University
Citation for published version (APA):
Yang, Y., Wang, H., & Blaabjerg, F. (2014). Reactive Power Injection Strategies for Single-Phase Photovoltaic
Systems Considering Grid Requirements. IEEE Transactions on Industry Applications, 50(6), 4065-4076.
https://doi.org/10.1109/TIA.2014.2346692
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners
and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.
- Users may download and print one copy of any publication from the public portal for the purpose of private study or research.
- You may not further distribute the material or use it for any profit-making activity or commercial gain
- You may freely distribute the URL identifying the publication in the public portal -
Take down policy
If you believe that this document breaches copyright please contact us at vbn@aub.aau.dk providing details, and we will remove access to
the work immediately and investigate your claim.
MANUSCRIPT ACCEPTED IN THE IEEE INDUSTRY APPLICATIONS MAGAZINE 1
Reactive Power Injection Strategies for
Single-Phase Photovoltaic Systems Considering
Grid Requirements
Yongheng Yang, Student Member, IEEE , Huai Wang, Member, IEEE, and Frede Blaabjerg, Fellow, IEEE
Abstract—As the development and installation of photovoltaic
(PV) systems are still growing at an exceptionally rapid pace,
relevant grid integration policies are going to change conse-
quently in order to accept more PV systems in the grid. The
next generation PV systems will play an even more active role
like what the conventional power plants do today in the grid regu-
lation participation. Requirements of ancillary services like Low-
Voltage Ride-Through (LVRT) associated with reactive current
injection and voltage support through reactive power control,
have been in effectiveness in some countries, e.g. Germany
and Italy. Those advanced features can be provided by next-
generation PV systems, and will be enhanced in the future to
ensure an even efficient and reliable utilization of PV systems. In
light of this, Reactive Power Injection (RPI) strategies for single-
phase PV systems are explored in this paper. The RPI possibilities
are: a) constant average active power control, b) constant active
current control, c) constant peak current control and d) thermal
optimized control strategy. All those strategies comply with the
currently active grid codes, but are with different objectives. The
proposed RPI strategies are demonstrated firstly by simulations
and also tested experimentally ona1kWsinge-phase grid-
connected system in LVRT operation mode. Those results show
the effectiveness and feasibilities of the proposed strategies with
reactive power control during LVRT operation. The design
and implementation considerations for the characterized RPI
strategies are also discussed.
Index Terms—Reactive power injection, single-phase systems,
photovoltaic (PV) systems, grid requirements, low-voltage ride-
through (LVRT), power (PQ) control, junction temperature,
reliability
I. INTRODUCTION
T
HE ADVANCEMENTS of power electronics technolo-
gies have shown great potential for renewable energy
integration into the grid, as proved by the continuously
booming penetration level of PhotoVoltaic (PV) systems [1]–
[5], which leads to increased grid decentralization and vul-
nerability. Hence, catering for further more PV installations
calls for advanced control strategies in compliance with grid
Manuscript received December 2, 2013; revised March 12, 2014 and June
14, 2014; accepted June 16, 2014. Paper 2013-IPCC-0972.R2, presented at
the 2014 IEEE Applied Power Electronics Conference and Exposition, Fort
Worth, TX, USA, March 16-20, and approved for publication in the IEEE
I
NDUSTRY APPLICATIONS MAGAZINE by the Industrial Power Converter
Committee of the IEEE Industry Applications Society.
The authors are with the Department of Energy Technology, Aalborg
University, 9220 Aalborg, Denmark (e-mail: yoy@et.aau.dk; hwa@et.aau.dk;
fbl@et.aau.dk).
This is the preprint version of the manuscript. When it is published, color
versions of one or more of the figures in this paper will be available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TIA.xxxx.xxxxxxx
requirements or standards. Currently, it is required that the
PV systems cease to energize local loads in the presence
of grid abnormal conditions, e.g. voltage sags and frequency
variations [6]–[11]. Meanwhile, it is required in those grid
regulations for most systems to operate at unity power factor
(or a minimum power factor, e.g. power factor ≥ 0.9) with
Maximum Power Point Tracking (MPPT) control in order
to extract as much energy as possible from the PV panels
[11]–[14]. Those grid requirements are valid in case of a
low penetration degree of PV systems, including the most
commonly used single-phase systems.
However, a still increasing adoption of PV systems will
violate the grid integration. For example, potential overload-
ing or voltage rises may appear at distributed grid feeders,
especially when a very high penetration level of PV systems
is reached, due to the intermittent nature of solar PV source
and the unbalance between PV supply and load demands [2],
[4], [10], [15]–[21]. Possibilities to solve those issues include
limiting feed-in maximum power from PV systems [22] and
reducing installations, which are against the goal of carbon
reduction in most countries, e.g. Germany, by enabling an
even more wide-scale adoption of renewable energies. Thus,
those countries have put forward specific grid requirements for
large-scale PV systems, which should be able to participate in
voltage regulation through reactive power control (injecting or
absorbing reactive power), as static grid support [23]–[26].
Meanwhile, the trip-off of an aggregated PV system owing
to anti-islanding protection may induce grid variations, leading
to more serious events, e.g. power outage [4], [10], [21], [27]–
[29]. Hence, in response to grid disturbances, it is better for
next-generation PV systems to provide dynamic grid support
in terms of Low-Voltage Ride-Through (LVRT) with Reactive
Power Injection (RPI), in order to: a) stabilize the grid in
case of failures and b) to avoid loss of massive PV generation
systems. For instance, in Italy, any generation system with
the total power exceeding 6 kW should have LVRT capability
[24]. In Germany, it has been in effectiveness for medium-
or high-voltage systems, including grid-connected PV systems
[4], [28]–[32]. Other countries also keep the pace with grid
code revisions [10], [33]–[35]. Those requirements were firstly
introduced to wind turbine systems, but today tend to be
extended to all PV systems, even for PV modules [34].
Obviously, the implementation of LVRT function violates the
anti-islanding requirement. Hence, as it is shown in Fig. 1,
compatibility of those two functions should be taken into
considerations when upgrading grid requirements.
MANUSCRIPT ACCEPTED IN THE IEEE INDUSTRY APPLICATIONS MAGAZINE 2
7LPH V
9ROWDJH /HYHO SX
$QWL,VODQGLQJ
3URWHFWLRQ
W
W
W
W
W
/957
5HPDLQ FRQQHFWHG
Y
Y
Y
0D\ GLVFRQQHFW
VKRUWWHUP LVODQGLQJ
0XVW GLVFRQQHFW
1RUPDO RSHUDWLRQ
/
9
5
9
9
7
P
P
D
L
Q
L
L
F
R
Q
Q
Q
Fig. 1. Suggestion on a compatible implementation of low-voltage (and
zero-voltage) ride-through and anti-islanding requirements for single-phase
PV systems connected to low-voltage networks.
Nonetheless, as the penetration level is continuously grow-
ing, much controllability of active power and reactive power
should be ensured in future PV systems, including reactive
power control to support the grid voltage statically (volt-
age rise mitigation) and also ride-through faults dynamically,
which is associated with RPI control during the transients. In
light of the above issues, this paper explores single-phase RPI
strategies, including: a) constant average active power control,
b) constant active current control, c) constant peak current con-
trol and d) thermal optimized reactive power control strategy.
Firstly, a brief introduction of the power control for single-
phase PV systems is given in § II, followed by the proposed
RPI methods. Simulations and experiments were carried out on
a 1 kW singe-phase system in the LVRT operation mode and
presented in § IV. Both results have verified the effectiveness
of the proposed RPI strategies.
II. P
OWER CONTROL OF SINGLE-PHASE SYSTEMS
Since the PV systems are still dominantly for residential ap-
plications at present, single-phase topologies are more widely-
used solutions for PV systems. Fig. 2 represents a typical
single-phase grid-connected PV system, where, in some cases,
a DC-DC converter is adopted to boost up the PV panel voltage
within an acceptable range of the PV inverter [6], [9], [11]. It
also offers the flexibility of MPPT control, which is a basic
requirement for such systems operating at unity power factor.
Meanwhile, the injected current should be synchronized with
the grid voltage, and as mentioned previously, the system
should disconnect from the grid when it presents disturbances
(e.g. frequency or voltage variation) at the Point of Common
Coupling (PCC) as shown in Fig. 2.
As for the control of single-phase systems with the RPI
function, the droop control concept [8], [36] for single-phase
PV systems is not suitable, since it requires that the line
is mainly inductive (i.e. XR). The utilization of adaptive
filtering technique leads to an instantaneous power control
solution [37]. This power control method is a good candidate
for single-phase systems when a satisfactory synthesis of the
power references is achieved. Besides the above possibilities,
the power control can also be developed in the dq-orαβ-
frame, based on the single-phase PQ theory [11], [37]–
[42]. The implementation of this control solution is intuitive
396WULQJV
3
39
%RRVW
RSWLRQDO
L
39
Y
39
Y
J
L
J
4
'&
'&
'&
$&
&
GF
,QYHUWHU
)LOWHU
Y
GF
3:0
LQY
3:0
E
&RPPXQLFDWLRQ
0RQLWRULQJDQG&RQWURO
6XSHUYLVRU\
FRPPDQG
0LVVLRQ3URILOHV
*ULG
3
R
=
J
3&&
&
/RDG
0337
,QYHUWHU
&RQWURO
4
Y
GF
3
Fig. 2. Typical power and control configuration of a single-phase grid-
connected PV system.
L
J
L
J
L
J
Y
JĮ
Y
Jȕ
Y
JĮ
Y
Jȕ
3
4
3
4
3
4
L
SY
Y
SY
Y
SY
0337
4
/957
Y
J
3ODQW
+DUPRQLF
FRPSHQVDWRUV
3,
FRQWUROOHUV
*
S
V
*
T
V
5HIHUHQFH
3URILOHV
*
S
V
Y
LQY
&XUUHQW
FRQWUROOHU
*
F
V
*
KF
V
Fig. 3. Control structure in the αβ-frame for single-phase single-stage PV
systems based on single-phase PQ theory [11], [40].
with less complexity, but it requires an Orthogonal Signal
Generation (OSG) system to create quadrature components
(v
gα
, v
gβ
and i
gα
, i
gβ
) corresponding to the real grid voltage
v
g
and current i
g
, as shown in Fig. 3. Moreover, a power
calculation method in terms of fast computation and high
accuracy can contribute to the control performance.
Thus, the RPI control can be implemented in this control
solution as the “Reference Profiles” by setting the references
for active power P
∗
and reactive power Q
∗
, and then the grid
current reference i
∗
g
is generated. In normal operation mode,
the active power reference P
∗
is the tracked maximum power,
P
MPP
, of the PV panels (P
∗
= P
MPP
) and Q
∗
=0Var.
When the RPI control is enabled by a detected grid condition
(voltage and frequency range), the reactive power is injected
according to the grid requirements, e.g. the German grid code
shown in Fig. 4(a) [4], [29]. It is noted that, during fault ride-
through operation, the system should inject sufficient reactive
current according to the grid voltage level [4], [29], [31]. This
relationship can be defined as,
k =
(I
q
− I
q0
)/I
N
(1 − v
g
)
, when I
q
<I
N
(1)
where I
q0
is the initial reactive current before grid failure, and
v
g
is the instantaneous voltage in p.u. during voltage fault.
Since the PV systems are required to operate at unity power
factor in MPPT mode, there is no reactive power injection
before voltage sags, i.e. I
q0
= 0 A. Moreover, it is required
by this grid code that k should be larger than 2 p.u., i.e.
k ≥ 2 p.u., for a minimum reactive current injection [29]. For
example, when the grid voltage sags to 0.8 p.u., a minimum
reactive current I
q
(40 % of the rated current I
N
) should be
MANUSCRIPT ACCEPTED IN THE IEEE INDUSTRY APPLICATIONS MAGAZINE 3
Y
J
SX
'HDG %DQG
,
T
,
1
/957 6XSSRUW YROWDJH
N SX
3
4
9
JQ
,
PD[
3
033
6
Q
4
PD[
6
PD[
6
ij
4
/957 6XSSRUW YROWDJH
D
E
N SX
)XOO UHDFWLYH
FXUUHQW LQMHFWLRQ
Fig. 4. Reactive power profiles for single-phase systems: (a) during LVRT for
medium- and/or high-voltage systems [29], [31], [43] and (b) reactive power
capability of a PV inverter.
injected into the grid. It is also shown in Fig. 4(a) that under
a severe voltage fault (e.g. v
g
= 0.3 p.u.) full reactive power
injection should be enabled, where the active power injection
could be deactivated. However, the amount of reactive power is
limited by the inverter apparent power, S
max
, as it is illustrated
in Fig. 4(b). This constraint should be taken into account when
designing the RPI strategies, i.e. the avoidance of inverter trip-
off due to over-current protection.
III. R
EACTIVE POWE R INJECTION STRATEGIES
Some grid specifications have been imposed on the next-
generation PV systems, especially for medium-/high-voltage
applications [26], [27]. In the future, PV systems, covering
a wide range of applications, have to provide reactive power
under grid faults. As discussed in the last paragraph, when
designing the RPI control strategies, both the grid requirements
(e.g. Fig. 4(a)) and the inverter current limitation shown in
Fig. 4(b) have to be complied. Those constraints are given as,
I
q
= I
N
, 0 ≤ v
g
< (1 −
1
k
) p.u.
I
q
= k(1 − v
g
)I
N
, (1 −
1
k
) p.u. ≤ v
g
< 0.9 p.u.
(2)
where k ≥ 2 p.u. is defined in (1), and
I
gmax
=
I
2
d
+ I
2
q
≤ I
max
(3)
in which I
d
is the active current, I
gmax
is the amplitude of the
injected current, and I
max
is the inverter allowable maximum
current level. In accordance with (2) and (3), the following
RPI strategies are proposed:
A. Constant Average Active Power Control (Const.-P)
The objective of this RPI control strategy is to maximize
the output energy with MPPT control during LVRT operation.
Therefore, the average active power is maintained constant in
the short-term period. Based on the single-phase PQ theory,
the average active power can be given as,
P =
1
2
v
gm
I
d
(4)
where v
gm
is the amplitude of the grid voltage during MPPT
operation and I
d
is the active current of the injected grid
current. In the normal operation mode, I
d
= I
N
, and hence,
under LVRT situation with Const.-P control, the average active
power P = k
d
P
N
=
k
d
2
v
gmn
I
N
, with v
gmn
, I
N
being the
nominal values of the grid voltage and current, respectively,
and k
d
being the power derating factor.
According to (2) and (4), when the instantaneous grid
voltage level v
g
:
1 −
1
k
p.u ≤ v
g
< 0.9 p.u., the current
in the dq-frame can be expressed as,
⎧
⎨
⎩
I
d
=
k
d
v
g
I
N
I
q
= k(1 − v
g
)I
N
(5)
in which k is defined in (1), and k
d
has been given previously.
When the grid voltage level sags to lower than
1 −
1
k
p.u.
(i.e. a severe voltage sag occurs), the system is required to
fully inject reactive power while the active power output may
be disabled (i.e. I
q
= I
N
). In that case, the system might still
operate at Const.-P mode, depending on the inverter current
limitation, and the current in the dq-frame is given by,
⎧
⎨
⎩
I
d
=
k
d
v
g
I
N
I
q
= I
N
(6)
However, when the required injection of reactive power
is fulfilled, according to (3), it might pose the inverter at a
risk of over-current and thus over-heating with this control
strategy to maintain a constant output power (i.e. maximum
power with MPPT control). Thus, based on (5) and (6), the
following constraints should be satisfied in order to avoid
inverter shutdown during LVRT:
1
v
g
k
2
d
+ k
2
(v
g
− v
2
g
)
2
≤
I
max
I
N
, (7)
when
1 −
1
k
p.u ≤ v
g
< 0.9 p.u., and
1
v
g
k
2
d
+ v
2
g
≤
I
max
I
N
(8)
when v
g
<
1 −
1
k
p.u.. Those could be the design criterions
for component selection, and can be illustrated in Fig. 5.
It is observed in Fig. 5 that the minimum value of the
inverter current limitation (I
max
) should be 2.24I
N
when
k =2p.u. so that the RPI strategy can be adopted in case
of a wide range of voltage drop (i.e. the grid voltage is within
0.5 p.u ≤ v
g
< 0.9 p.u.) without power derating (k
d
=1
p.u.). As for a predesigned PV inverter with a robustness
margin, the system has to derate the output power in order
to inject enough reactive power. For example, if the allowable
maximum current of a PV inverter, I
max
=1.5I
N
and k =2
p.u., the PV systems should reduce the active power output
(e.g. k
d
= 0.5 p.u.), when the voltage drops below 0.72 p.u.,
as it is shown in Fig. 5. It is also noted in Fig. 5 that, by
derating operation, the Const.-P strategy can be adopted for
a wider range of voltage sags.
MANUSCRIPT ACCEPTED IN THE IEEE INDUSTRY APPLICATIONS MAGAZINE 4
9ROWDJH/HYHOSX
,
JPD[
,
1
SX
N
G
SX
6ROLGOLQHN SX
'DVKHGOLQHN SX
,
PD[
,
1
N
G
SX
:LGHUUDQJH
Fig. 5. Design constraint of the Const.-P considering the inverter over-current
protection, where k
d
= P/P
N
and k is defined in (1).
B. Constant Active Current Control (Const.-I
d
)
Another RPI control possibility under LVRT operation is to
keep the active current constant (i.e. I
d
= const.). According
to (4), the active current I
d
can be obtained as,
I
d
=
2P
v
gm
= mI
N
= const. (9)
in which m is defined as the scaling factor for the design
consideration in case of derating operations, and 0 ≤m ≤ 1
p.u.. According to (9), the active power will automatically
be reduced when this RPI control strategy is adopted in
the response to voltage sags, i.e. P ∝ v
gm
. Meanwhile,
the reactive current I
q
can be calculated according to the
requirement shown in Fig. 4(a) and (2). Subsequently, the
current in the dq-frame can be given as,
I
d
= mI
N
I
q
= k(1 − v
g
)I
N
(10)
where (1 −
1
k
) p.u. ≤ v
g
< 0.9 p.u. and k are defined
previously. Notably, when a severe voltage fault happens (very
low voltage), the PV system should inject full reactive power.
In that case, the current in dq-frame can be expressed as,
I
d
= mI
N
I
q
= I
N
(11)
when v
g
< (1 −
1
k
) p.u..
With the Const.-I
d
control strategy, the amplitude of the
injected current may also exceed the inverter limitation ac-
cording to (3), and then trip the inverter protection. In order
to avoid this, the following conditions should be fulfilled,
m
2
+ k
2
(1 − v
g
)
2
≤
I
max
I
N
, (12)
when
1 −
1
k
p.u ≤ v
g
< 0.9 p.u., and
m
2
+1≤
I
max
I
N
, (13)
when v
g
<
1 −
1
k
p.u.. For simplicity, the level of active
current can be controlled to be that of the rated current (i.e.
m =1p.u., I
d
= I
N
).
Similarly, a design guide for this RPI control strategy can
be given in Fig. 6. It is seen from Fig. 5 and Fig. 6 that the PV
9ROWDJH/HYHOSX
,
JPD[
,
1
SX
P
6ROLGOLQHN
'DVKHGOLQHN
P SX
P SX
,
PD[
,
1
Fig. 6. Design constraint of the Const.-I
d
considering the inverter over-
current protection, where m = I
d
/I
N
and k is defined in (1).
inverter with Const.-I
d
control can be designed with a lower
I
max
/I
N
when it is compared to the one with Const.-P control
strategy. Therefore, it offers the possibility to select power
devices with lower current ratings and thus lower cost. It is
also worth to point out that derating operation of a PV system
can be achieved by changing m because of the proportional
relationship between the active power and the grid voltage
amplitude, P ∝ v
gm
. A smaller m leads to the possibility to
select power devices of further lower ratings.
C. Constant Peak Current Control (Const.-I
gmax
)
A PV inverter with the previous discussed RPI strategies has
a risk of over-current loading when it is operating in LVRT
mode. Thus, the Const.-I
gmax
is proposed. With this control
strategy, there is no unintentional inverter shutdown due to
over-current protection, since the peak of the injected grid cur-
rent is kept constant and lower than the inverter current limita-
tion during LVRT, i.e. I
gmax
= nI
N
= const., and I
gmax
≤
I
max
, where n is defined as the peak current scaling factor.
According to (2), when the grid voltage is within the range:
(1 −
1
k
) p.u. ≤ v
g
< 0.9 p.u., the current in dq-frame can be
given by,
I
d
=
n
2
− k
2
(1 − v
g
)
2
I
N
I
q
= k(1 − v
g
)I
N
(14)
while, if the grid voltage goes further lower than
1 −
1
k
p.u.,
according to (3) the current in dq-frame should be,
I
d
=
√
n
2
− 1I
N
I
q
= I
N
(15)
where v
g
and k are defined previously.
It should be noted that n has a maximum value of
I
max
I
N
p.u. considering inverter current protection shown in (3). For
example, when a inverter is designed with a margin of 2 p.u.
(i.e. I
max
=2I
N
), the maximum n should be 2 p.u. to ensure a
stable RPI without tripping the inverter during LVRT. Thus, if
n ≤
I
max
I
N
, riding-through operation of the PV inverter will not
give an amplitude rise to the injected grid current. Meanwhile,
according to (4) and (14), the active power will be reduced in
order to inject sufficient reactive power during LVRT.
![Fig. 8. (a) Electrical and thermal models of a power module and (b) Foster model of the thermal impedance from junction to case [47], [51], [52].](/figures/fig-8-a-electrical-and-thermal-models-of-a-power-module-and-vgc7jkvj.webp)
![Fig. 19. Experimental results of a single-phase grid-connected system with different RPI control strategies (voltage level: vg = 0.55 p.u., k = 2 p.u., grid voltage vg [100 V/div], grid current ig [10 A/div], active power P [500 W/div], reactive power Q [500 Var/div], and time [40 ms/div]): (a) Const.-Id with m = 1 p.u. and (b) Const.-Igmax with n = 1 p.u..](/figures/fig-19-experimental-results-of-a-single-phase-grid-connected-2qckmh7h.webp)
![Fig. 20. Experimental results of a 1 kW single-phase system with TOptimized RPI control strategy (voltage level: vg = 0.55 p.u. and k = 2 p.u.): (a) grid voltage vg [100 V/div] and grid current ig [10 A/div]; (b) active power P [500 W/div] and reactive power Q [500 Var/div]; time [40 ms/div].](/figures/fig-20-experimental-results-of-a-1-kw-single-phase-system-2nywemps.webp)
