Evaluation of Photovoltaic Systems for Reactive Power

Compensation in Low Voltage Power Systems

Mesut Uğur, Erencan Duymaz, Murat Göl, Ozan Keysan
Department of Electrical and Electronics Engineering
Middle East Technical University
Ankara, Turkey

Abstract— The four-quadrant operation ability of photovoltaic reactive power compensation, voltage regulation and
(PV) inverters makes them promising candidates for reactive harmonic distortion elimination. In [3], reactive power
power compensation in low voltage systems. In this paper, injection strategies in single-phase inverters by considering
utilization of PV inverters instead of conventional reactive grid requirements are investigated in order to support grid
power compensation units is evaluated. The use of PV inverters voltage and to comply with the LVRT dynamics. Moreover, in
for reactive power compensation as well as active power the study [4], two techniques for reactive power compensation
supplying is investigated considering a real life system. The are given for the low voltage grid with high PV penetration.
considered system suffers from low capacitive power factor due The methods are mainly investigated in terms of
to the connected online UPS system. The paper firstly analyzes
communication infrastructure and power quality factors.
utilization of low voltage PV systems for reactive power
Studies [5], [6] emphasize the operation of PV inverters as
compensation purpose technically, and then presents a detailed
economic study in terms of short and long-term costs. The costs Static Synchronous Compensator (STATCOM) during night
are evaluated considering reactive demand charge to the operation. In [7], the potential of distributed renewable
customer. sources is emphasized due to the fact that conventional
reactive power sources are in transmission level and system
Index Terms-- Solar power generation, reactive power operators ask for reactive power generation to maintain a
compensation, PV inverters, reactive demand charge stable grid operation. The paper mainly focuses on the
incentives that system operators pay to the costumers.
I. INTRODUCTION
In this paper, utilization of PV inverters instead of
Apart from the environmental concerns, low prices due to conventional capacitor banks or namely reactive power
technological advances and government incentives are factors compensation units (RPCUs) will be investigated from an
that affect the choice of investment in power plants. As a economic point of view. The study will be conducted using
result, the share of renewable sources in total installed real system of the Ayaslı Research Center located at Middle
capacity reached significant levels in recent years. Aside from East Technical University, Ankara, Turkey. In Section II, the
its benefits, renewable energy sources bring challenges to problem will be defined in detail. Section III will explain the
system operators in voltage regulation and frequency stability. static solution and in Section IV, the static solution is analyzed
Photovoltaic (PV) systems are renewable sources that are and evaluated in terms of short-term and long-term costs. In
commercially available even for small sizes. Regulations that Section V, use of RPCUs and PV inverters as reactive power
allow installation of PV systems at residential buildings compensators is discussed.
caused also increased the demand for those systems. However, II. PROBLEM DEFINITION
the radial structure of the low voltage network creates several
power quality problems especially when the generation In the power system network, active power is the main
exceeds consumption. Reverse power flow phenomenon component of the apparent power that does useful work.
creates voltage rises on the load side, which may result in However, reactive power, especially inductive reactive power
unintentional load or generator tripping. Hence, PV inverters is also required for the successful operation of some electrical
are expected to cooperate with system operators. loads such as industrial motors. However, demanding reactive
power from the grid causes voltage drops as well as additional
losses in the transmission and distribution systems. Hence,
PV inverters are capable of supporting additional services system operators restrict such demands on the point of
to the grid thanks to their four-quadrant operation ability. In common coupling (PCC), such that customers are responsible
[1], PV inverters are used as Active Power Filter (APF) to for complying defined power factors at the PCC. Considering
suppress the current harmonics and reactive currents. Study in the fact that load is assumed to be inductive most of the time,
[2] focuses on smart utilization of the PV inverters such as common practice is utilizing RPCUs to compensate the

978-1-5386-4505-5/18/$31.00 ©2018 IEEE

Authorized licensed use limited to: Universidad Nacional de Colombia (UNAL). Downloaded on September 22,2022 at 01:11:19 UTC from IEEE Xplore. Restrictions apply.
 inductive reactive power at the PCC by using capacitor banks. loss of profit. This is not experienced in the studied case;
Nevertheless, there has been a paradigm shift in the sense of hence it is kept out of the scope of this paper.
power system network over the last decade with the increasing
trend towards renewable energy. Due to the fact that rooftop III. STATIC SOLUTION
PV systems spread everywhere in the low voltage network, the The PV systems can be utilized for reactive power
system experiences bidirectional power flow, voltage rise at compensation instead of the RPCUs. In the studied case, the
the PCC and disturbance in the power factor. RPCUs actually do not operate at all due to the load being
This paper discusses a case study based on a real life case capacitive because of the UPS. This situation results in an
experienced at Ayaslı Research Center located in METU, extra installation cost to the customer despite it cannot solve
Turkey. The building is equipped with a 50 kWp PV system, the power factor problem. The first advantage of the proposed
two distinct loads with 20 kVA and 120 kVA, each of which static solution is the four-quadrant operation of the PV
has its own uninterruptible power supply (UPS), and an RPCU inverters. In this way, reactive power compensation can be
with 50 kVAr rating. The single line diagram of the system is achieved not only when the load in inductive, which is the
shown in Fig. 1. general case, but also when the load is capacitive as in this
case. Moreover, PV inverters can operate as reactive power
PV System RPCU
compensators even if the active power input from the panels is
zero, e.g. at night. In such a case, they operate just like a
STATCOM, and can be used whenever compensation is
34.5kV
Dist. Tr. Grid needed. Finally, the amount of reactive power flow can be
PCC Line adjusted in a continuous manner with PV inverters, although
RPCUs can only achieve this with discrete steps.
0.4kV/34.5kV
Reactive power compensation with a PV system can be
achieved in two ways. Either the size of the PV system is
Loads On-line UPS increased according to the maximum amount of reactive
power, or the amount of active power injected to the grid is
Figure 1. Ayaslı Research Center single line diagram
reduced whenever the apparent power rating is exceeded. In
The loads in the system have electronic components such this research, both cases will be considered and evaluated in
as computers and air conditioners, with assumed power factor terms of their long-term cost to the system. In the first case,
of 0.98 inductive. On the other hand, on-line UPS systems the static solution will have impact on the installation cost as
equipped in the building are adjusted to generate capacitive well as long-term cost due to the increased losses. The latter
reactive power when there is no interruption in the system. case will decrease the efficiency of the system since the
First, it is aimed to extend the lifetime of the batteries by system is required to operate outside of the maximum power
continuous charge-discharge operation. Second, the loads are point (MPP), and the profitability because of the reduced
assumed to operate in inductive mode so that UPS systems are active power injection. The static solution may also have
adjusted to operate in capacitive mode by default. Therefore, impact on the DC link capacitor size in terms of capacitance
buildings equipped with such loads and UPSs may encounter requirement and RMS ripple current requirement, which is
capacitive operations although loads are assumed to be highly dependent on each specific design and type of the DC
inductive normally. In Ayaslı Research Center, it has been link capacitor used. In both of the cases, additional power
noticed that the RPCU has not worked for five years due to the factor measurement will be required and control complexity
capacitive nature of the load. This results in impair will increase.
investment which brings unnecessary installation cost.
Im
Moreover, the power factor limitation constitutes another
problem associated with this approach. In Turkey, for Trajectory of Is
buildings with power rating more than 9 kW, it is obligatory to
use electricity meter with the capability to reactive power 4 P>0
P<0
measurement, and impose sanctions to limit the power factor. Id Q>0
Q>0 2
PWM Rectifier

In the normal operation, power factor might be within these

Operation as a

Is
acceptable limits. However, if the PV power generation is Iq
around the active power demand, then very low or even zero 1 Vs
Re
active power might be observed at the PCC. Therefore, even if
the reactive power drawn or injected to grid is negligible in the P>0
normal operation, poor power factor will be observed at the P<0
3 Q<0
PCC. In such a case, the customer will be penalized by the Q<0
distribution system operator. Hence, the reactive power should 5
be compensated to maintain the power factor within the
Figure 2. Phasor diagram of a PV inverter
acceptable limits. Another problem in systems employing PV
panels is voltage regulation issues due to the reverse power A grid-connected inverter can operate at four quadrants of
flow when loads are interrupted. Depending on the design, PV the V-I plane as mentioned before. PV inverters use only two
inverters may trip resulting in loss of power generation and quadrants where active power is always positive according to

Authorized licensed use limited to: Universidad Nacional de Colombia (UNAL). Downloaded on September 22,2022 at 01:11:19 UTC from IEEE Xplore. Restrictions apply.
 the sign convention. Low power PV systems connected to low capacitance, higher current rating and better lifetime [8]. In
voltage grid are aimed to operate at unity power factor as this study, both types will be considered for evaluation.
marked in the phasor diagram in Fig. 2 as Point 1. By keeping
the magnitude of the current (and hence the apparent power) Capacitance selection of a DC Link capacitor ( ) is
constant, a PV system can be operated as inductive or determined by voltage ripple constraint as well as hold-up
capacitive as in Point 2 and Point 3, respectively. They can energy requirement. In (1) and (2), general expressions for
also be operated as STATCOMs with purely inductive and ripple requirement and hold-up energy are provided where M
purely capacitive as in Point 4 and Point 5, respectively. is the modulation index, fsw is the switching frequency, Vdc,r is
the DC-link voltage.
This wide range of operation can be achieved by using the
×
reactive component of the reference current as an additional > (1)
√ × , ×
control variable to the control system for reactive power
compensation. The modification on the control system is ×
shown in Fig. 3 on a conventional control block diagram of a > (2)
PV inverter in which active and reactive components (d and q
components) of the injected current are controlled separately. Considering a generic design for a 50 kWp central PV
These components are also marked in Fig. 2 (Point 2) for an inverter with a switching frequency of 5 kHz (fsw), DC Link
inductive case. average voltage of 750 V (Vdc), peak to peak ripple constraint
Vdc* + PI Id* +
Σ
PI +
Σ
Vd*
Va*
of 1%, and hold-up requirement of 500 V for a full
Σ
-
Regulator
-
Regulator
+
dq0
Vb*
Pulse
Width
fundamental cycle of the grid, the capacitance requirements
Vdc
Iq* ≠ 0 +
Σ
PI
Regulator
+
Σ
V q* abc Vc* Modulation come out to be 5.3 mF and 0.58 mF, respectively. As a result,
-
capacitance requirement is determined by hold-up energy,
measurement

+
S and pf

Iq Id
Reference
calculation
Vq Vd which is dependent on the kVA rating of the inverter. For a
Ia Va
PV inverter, which can operate down to 0.8 pf at the same
dq0

dq0

Ib Vb
Ic Vc active power, the capacitance is increased by 25%, which
abc

abc

Va
Vb
Phase θ should be taken into account in capacitor size evaluation.
Vc
Locked
Loop When the total power rating is kept constant, there will be no
Figure 3. The conventional control block diagram of a PV inverter and change.
modification for RPC
RMS ripple current requirement of the DC Link capacitor
Conventionally, the reference to the reactive component
has also been derived as in Equation (3) [8], where pf is the
(Iq*) is set to zero to operate the PV system at unity power
factor (Point 1 on Fig. 2). With the proposed modification, this power factor.
reference can be used to adjust the reactive power by either an
external reference signal or additional measurements. √3 √3 9
, = , 2 + − (3)
IV. COST ANALYSIS AND EVALUATION 4 16
In this section, the static solution will be evaluated in terms
of short-term and long-term cost and will be compared to the In our case, the ripple current of the DC Link capacitor
conventional system specific to the case study, which is under increased by 18% if the size of the inverter is increased for
investigation. 0.8 pf operation and by 5% if the total kVA rating is kept
constant.
A. DC Link Capacitor Size
B. Power Semiconductor Losses
In grid connected PV inverters, DC Link capacitors have
the duties of supplying the inverter current at switching Making a PV inverter to operate at non-unity power factor
frequency, smoothing the DC Link voltage waveform, will affect the long-term cost of the system due to additional
supplying the transient power deviations and hold-up energy semiconductor losses even if the total kVA rating is kept
when power failure occurs in the system. The ability of a PV constant. In this study, power semiconductor losses will be
inverter to transfer reactive power is also achieved by DC evaluated with selected commercial IGBTs, including
Link capacitor; i.e., it handles the power fluctuation on the DC transistor conduction (Ptc) and switching loss (Pts) and anti-
side due to reactive power. There are two major parameters of parallel diode conduction (Pdc) and reverse recovery loss (Pdr).
a DC Link capacitor, which will be affected directly by General formulation of semiconductor losses is shown in (4) –
reactive power compensation capability of a PV inverter: (7). In this formulation, Eon and Eoff are turn-on and turn-off
capacitance and RMS current rating. switching energies, Icp and Iep are the peak current, Vce,sat is the
saturation voltage drop, Vec is the forward voltage drop of the
Aluminum electrolytic capacitors are the most commonly diode, Irr and trr are recovery current and time, and Vce,p is the
used types in DC Link applications due to their low cost and peak reverse recovery voltage.
high capacitance per volume. However, their RMS current
ratings per volume are very low and lifetime is short. On the 1
contrary, metal film capacitors have higher cost, lower = , + (4)
8 3

Authorized licensed use limited to: Universidad Nacional de Colombia (UNAL). Downloaded on September 22,2022 at 01:11:19 UTC from IEEE Xplore. Restrictions apply.
 C. Long Term Cost Evaluation
= + (5)
The interpretation of the loss results to cost is required for
1 the comparison of the two proposed methods. For this
= − (6)
8 3 purpose, the actual case study in Ayaslı Research Center is
investigated for long term. The daily power generation of the
= , (7) PV panels in winter and summer are shown in Fig. 4. The
8 load active and reactive power demand and the UPS reactive
For loss calculation, a 150A, 1200V IGBT is selected for power injection, which are assumed to be constant throughout
50kVA case and a 200A, 1200V IGBT is selected for 70kVA the year are also shown in Fig. 4.
case from the same HV-IGBT family of Powerex. There are
three cases under investigation for comparison where solar 100
PV Gen (summer) - kW
panel power is considered as 50 kW for all cases, as shown in PV Gen (winter) - kW

PV Generation, Load and UPS

Table 1. Case-1 is the conventional system with no reactive 80
Load P - kW
Load Q - kVAr
power injection; Case-2 is the same system operating at 0.8 pf UPS Q - kVAr

and Case-3 is the system where inverter size is increased for

reactive power compensation. 60

TABLE 1: CASES CONSIDERED FOR LOSS CALCULATION

40
Pout Qout pf Inverter size

Case - 1 50 kW 0 1.0 Same

20
Case - 2 40 kW 30 kVAr 0.8 Same
Case - 3 50 kW 50 kVAr 0.7 Increased 0
0 5 10 15 20 25
Time (hours)
For these cases, all the losses are found and listed as in
Table 2. First, each loss component is found and total loss of Figure 4. Daily profiles of PV generation, load and UPS
the inverter (Ploss) and active power injected to the grid (Pgrid)
are determined. In addition, total effective loss (Ploss_eff) is The apparent power of the PV inverter and daily reactive
found considering the capacity of the solar panels, since the power demand from the PV inverter is shown in Fig. 5,
actual effect of reactive power compensation to the customer supposing that reactive power compensation is performed by
is the effective power loss. the PV inverters to make the PCC power factor unity.

TABLE 2: POWER SEMICONDUCTOR LOSS RESULTS

Reactive Power Demand and Apparent Powers

Reactive Power Demand

60 PV Apparent Power (summer) - kVA
Case - 1 Case - 2 Case - 3 PV Apparent Power (winter) - kVA

50
Ptc (W) 44.39 25.51 59.07
Pts (W) 26.6 26.66 40.78 40

Pdc (W) 6.81 26.15 18.9

30
Pdr (W) 15.92 15.92 20.29
Ploss (W) 562.68 565.44 834.24 20

Pgrid (kW) 49.43 39.43 49.166

10
Ploss_eff (W) 562.68 10565.44 834.24
0
0 5 10 15 20 25
It has been shown that, both of the methods have their own Time (hours)
benefits and drawbacks. Increasing the size of the PV inverter Figure 5. PV inverter apparent power and daily reactive power demand in
(Case-3) yields size increase on power semiconductors, case it is used for reactive power compensation
capacitor and possibly other elements which will result in
installation cost increase, while the increase on the effective Moreover, the active power, reactive power and the power
loss is 50% (nearly 270W) for rated conditions. On the other factor at the PCC are also obtained and shown in Fig. 6 and
hand, using the same inverter for reactive power compensation Fig. 7, for the cost and penalty calculation due to reactive
(Case-2) does not require much size increase for components power. The electricity prices are 6.5 ¢/kWh and 3.1 ¢/kVArh,
while the effective loss is too much (18 times the original however, the reactive power charge is only evaluated when Q
case) which will result in operational cost to the customer. > 0.2P which has been taken into account in the cost
evaluation.

Authorized licensed use limited to: Universidad Nacional de Colombia (UNAL). Downloaded on September 22,2022 at 01:11:19 UTC from IEEE Xplore. Restrictions apply.
 12
Energy lost (kWh) - summer
PCC Active Power (Summer) - kW
100 P loss X5 (kW) - summer
PCC Active Power (winter) - kW 10
Energy lost (kWh) - winter
PCC Reactive Power - kVAr
PCC Active and Reactive Powers

P loss X5 (kW) - winter

80 8

60 6

40 4

2
20
0
0 0 5 10 15 20 25
Time (hours)

-20 Figure 9. Daily power loss (Case-2)

-40 10
0 5 10 15 20 25 Energy lost (kWh) - summer
Time (hours) Power loss X5 (kW) - summer
8 Energy lost (kWh) - winter
Figure 6. PCC reactive power and active power Power loss X5 (kW) - winter

6
1.5
PCC power factor (Summer)
PCC power factor (winter) 4
1
2

0.5
0
0 5 10 15 20 25
Time (hours)
0
Figure 10. Daily power loss (Case-3)

-0.5
TABLE 3: LONG TERM COST EVALUATION
Q CostQ E CostP
-1 (MVArh) ($) (MWh) ($)
0 5 10 15 20 25
Time (hours) Case - 1 156 4848 1.500 97
Figure 7. PCC power factor Case - 2 0 0 3.154 205

6 Case - 3 0 0 2.751 179

Energy lost (kWh) - summer
P loss X5 (kW) - summer
5 Energy lost (kWh) - winter
P loss X5 (kW) - winter
Q injection x0.01 (kVArh)
V. CONCLUSION
4 Q charged x0.01 (kVArh) - summer
Q charged x0.01 (kVArh) - winter It has been shown that the customer is faced with a high
3
reactive power injection charge when no additional reactive
2 power compensation is applied, even though RPCU is
installed in the building. The reason is the capacitive nature of
1 the UPS systems and the active power injection by the PV
0
system. When the same inverters are used for reactive power
0 5 10 15 20 25 compensation, the cost due to active power losses increased
Time (hours)
over 100%, which is still very small compared to no reactive
Figure 8. Daily power loss and reactive power charge for the power compensation case.
conventional system (Case-1)
In the case-3, increased inverter size yields a little less long
In Fig. 8, daily power loss, cumulative energy lost and term cost than using original inverters thanks to the utilization
reactive power injection and charge are shown for case-1. of solar panels at full power. However, size increase of the
Fig. 9 and 10 show the daily power loss and cumulative IGBTs, DC Link capacitors and other major equipment on the
energy lost when reactive power is compensated with PV PV system will bring higher installation cost. On the inverter
inverters for case-2 and case-3, respectively. power stage, the cost increase of these fundamental
components has been estimated as 25%, however, exact
Long term cost evaluation results are shown in Table 3 in installation cost increase cannot be told. The reason is that
per year base, where reactive power charged to the customer these inverters are subject to mass production and cost of each
by the distribution system operator (Q), total cost of reactive item will not be the same as commercial off-the-shelf
power penalty (CostQ), total lost energy (E) and total cost of products. In addition, there are R&D investments and service
active power loss (CostP) are shown for the three cases.

Authorized licensed use limited to: Universidad Nacional de Colombia (UNAL). Downloaded on September 22,2022 at 01:11:19 UTC from IEEE Xplore. Restrictions apply.
 expenditures. Even so, it is very clear that the installation cost [2] F. Barrero-Gonzalez, V. Minambres-Marcos, M. Guerrero-Martinez, E.
will be much larger for case-3. Romero-Cadaval, M. Milanes-Montero, and E. Gonzalez-Romera,
“Photovoltaic inverter with smart grid functions,” EEEIC 2016 - Int.
One of the major reasons that such case exists in the low Conf. Environ. Electr. Eng., no. Dcdc, 2016.
voltage network is the lack of monitoring mechanisms. To [3] Y. Yang, H. Wang, and F. Blaabjerg, “Reactive power injection
strategies for single-phase photovoltaic systems considering grid
avoid such problems, low voltage should be monitored requirements,” IEEE Trans. Ind. Appl., vol. 50, no. 6, pp. 4065–4076,
extensively. In this way, more secure low voltage can be 2014.
achieved as well as such problematic consumers can be [4] M. Castilla, M. Velasco, J. Miret, P. Martí, and A. Momeneh,
detected. Finally, it should also be noted that the penalization “Comparative study of reactive power control methods for photovoltaic
by power factor should be replaced by the amount of reactive inverters in low-voltage grids,” IET Renew. Power Gener., vol. 10, no. 3,
pp. 310–318, 2016.
power demand in the regulations to avoid these problems with [5] M. S. E. Rao and K. Obulreddy, “Utilization Of Pv Solar Farm As
the increasing number of photovoltaic systems in low voltage Statcom During Night Hours In A Distribution Network,” Int. J. Sci.
network. Technol. Res., vol. 3, no. 8, pp. 33–39, 2014.
[6] N. Solanki and J. Patel, “Utilization of PV solar farm for Grid Voltage
ACKNOWLEDGMENT regulation during night; analysis &amp; control,” in 2016 IEEE 1st
International Conference on Power Electronics, Intelligent Control and
This work is partially supported by Scientific and Energy Systems (ICPEICES), 2016, pp. 1–5.
Technological Research Council of Turkey (TUBITAK) under [7] a. J. Von Appen, B. C. Marnay, C. M. Stadler, D. I. Momber, E. D.
the TUBITAK project number 115E146. Klapp, and F. A. Von Scheven, “Assessment of the economic potential
of microgrids for reactive power supply,” 8th Int. Conf. Power Electron.
REFERENCES - ECCE Asia, no. June, pp. 809–816, 2011.
[8] J. W. Kolar and S. D. Round, “Analytical calculation of the RMS current
[1] S. Biricik, H. Komurcugil, and M. Basu, “Photovoltaic Supplied Grid- stress on the DC-link capacitor of voltage-PWM converter systems,”
Tie Three-Phase Inverter with Active Power Injection and Reactive IEE Proc. - Electr. Power Appl., vol. 153, no. 4, p. 535, 2006.
Harmonic Current Compensation Capability,” pp. 3–8, 2016.

Authorized licensed use limited to: Universidad Nacional de Colombia (UNAL). Downloaded on September 22,2022 at 01:11:19 UTC from IEEE Xplore. Restrictions apply.