Applications Operational Results of

Grid-Connected Photovoltaic
System With Different Inverters
Sizing Factors (ISF)
W. N. Macedo and R. Zilles
*
,y
Instituto de Eletrotecnica e Energia, Universidade de Sao Paulo, Av. Professor Luciano Gualberto, 1289, Cidade
Universitaria, CEP 05508-900, Sao Paulo SP, Brazil
This paper presents operational results of a 1107 kWp grid-connected photovoltaic
system. This system is made up by eight groups with different relationships between
the inverters rated power and the PV generators maximum power P
0
Inv
=P
0
PV
. The
obtained results led to the verication that the different studied relationships,
P
0
Inv
=P
0
PV
between 55 and 102%, do not affect signicantly the nal yields (Y
F
).
Copyright # 2006 John Wiley & Sons, Ltd.
key words: PV systems; grid connection; operation performance analysis
Received 8 September 2006; Revised 19 October 2006
INTRODUCTION
I
n agreement with some works,
110
it is observed that there is not a denitive orientation regarding what
relationship between the inverters rated power and the PV generators maximum power should be used,
since it depends on a series of factors intrinsic of each specic installation. On the other hand, the difference
between the PV generators maximum power under standard test conditions (STC) and the power that it really
supplies, coupled to the fact that the PV generator operates most of the time out of the test conditions, usually is
used as excuse for over-sizing the maximum power of the PV generator in relation to the inverters rated
power.
8,11
However, it should be adverted that the difference between the values of the maximum power and the
real power of the PV generator has been decreasing and, in some cases, becoming positive.
8
Unfortunately, there are few works based in experimental results of systems operating with different relative
capacities, P
0
Inv
=P
0
PV
. The most common is to nd operational results in kWh/kWp, without mentioning the
relationship between the inverters capacity and the capacity of the PVgenerator.
1214
In that context, the lack of
results of systems operating with different relationships P
0
Inv
=P
0
PV
is still veried.
This work presents operational results of an 1107 kWp grid-connected photovoltaic system, formed by eight
groups with different P
0
Inv
=P
0
PV
relationships. Each group has a PV generator connected to a 1 kW inverter. That
system, located in the geographical coordinates 238 32 36" S, 468 37 59" W, is incorporated to the north facade
of the administration building of the Instituto de Eletrotecnica e Energia da Universidade de Sao Paulo, Brazil.
PROGRESS IN PHOTOVOLTAICS: RESEARCH AND APPLICATIONS
Prog. Photovolt: Res. Appl. 2007; 15:337352
Published online 13 December 2006 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/pip.740
* Correspondence to: R. Zilles, Instituto de Eletrotecnica e Energia, Universidade de Sao Paulo, Av. Professor Luciano Gualberto, 1289,
Cidade Universitaria, CEP 05508-900 Sao Paulo, SP, Brazil.
y
E-mail: zilles@iee.usp.br
Copyright # 2006 John Wiley & Sons, Ltd.
Table I shows the maximum power of each PV generator obtained from the sum of the individual maximum
power under STC of each PV module and the real values of the generators maximum power obtained through
eld measures, accomplished in agreement with the procedure described in Caamano et al.
15
It is important to
stress that the relationship P
0
Inv
=P
0
PV
for each one of the groups is based in the measured power values (VM) of
each PV generator.
To help the reader to understand the results shown in this paper, some useful information, such as the PV
generators orientation and tilt angle, maximum power point voltage (Vmp) and open circuit voltage (Voc), both
at STC, number of series connected PVmodules, and temperature coefcients of maximumpower voltage of the
PV modules (l
mp
) can be seen complemented in Table I.
The presented results include a year of monitoring and they were collected using an acquisition system that
records the averages values in 5 min intervals. Among the main considered parameters in the analysis, as well as
the ones monitored by the data acquisition system, the following ones can be pointed out:
H
(t,b)
, is the incident irradiance in the plane of the PV generator;
T
c
, is the temperature in the rear part of the photovoltaic module;

P
PV
, is the DC power supplied by the PV generator;
P
mp
, is the DC power supplied by the PV generator when operating in the maxim power point;
P
out
, is the AC power delivered by the inverter;
P
0
PV
, is the maximum power of the PV generator;
P
0
Inv
, is the rated power of the inverter;
h
Inv
, is the energy efciency of the inverter;
h
MPPT
, is the efciency of the maximun power point tracking;
Y
F
, is the nal yield in kWh/kWp.
From the obtained results, it is veried that the relative capacity of the inverter regarding the PV generator
for the studied relationships, P
0
Inv
=P
0
PV
between 55 and 102%does not signicantly affect the nal yields (Y
F
).
On the other hand, the importance of power-limitation process for protection of the inverter as a necessary
prerequisite for PV generators over-sizing becomes evident.
This work presents a detailed analysis of the inuence of the relationship between the inverters rated power
and the PV generators maximum power in the behavior of a PV grid connected system. In this sense, 8 PV grid
connected congurations have been monitored during 1 year and the results are presented. The paper provides
empirical insight on the inuence of a widely used design practices of sizing the inverter and PV generator for
PV grid connected systems in sunny countries.
Table I. PV generators characteristics
Group N1 N2 N3 N4 N5 N6 N7 N8
Orientation North North North North North North North North
Tilt angle 238 238 238 238 238 238 238 238
V
mp
, STC (V) 220 221 205 204 187 160 272 274
V
oc
, STC (V) 273 273 254 252 231 205 336 339
Series connections 13 13 12 12 11 10 16 16
l
mp
, (%/8C) 05 05 05 05 05 05 05 05
P
0
PV
, STC (Wp) 1007 975 1940 1800 1650 1538 1200 1295
P
0
PV
, VM

(Wp) 978 1002 1802 1757 1663 1477 1188 1201

P
0
Inv
=P
0
PV
(%) 1020 998 555 570 600 680 842 830

VM is the measured value of each PV generator and extrapolated for the standard test conditions.

Assumed as the cells temperature.

Copyright # 2006 John Wiley & Sons, Ltd. Prog. Photovolt: Res. Appl. 2007; 15:337352
DOI: 10.1002/pip
338 W. N. MACE

DO AND R. ZILLES
PERFORMANCE ANALYSIS OF THE PV GENERATORS
Yields and capture losses of the PV generators
The reference yield (Y
R
), for 2004, corresponds to 1541 kWh/m
2
/year or 423 kWh/m
2
/day. That yield, together
with the array yield, Y
A
, allows to calculate the captures losses, L
C
, that are exclusive losses of the PV generator
and they can be calculated by the difference between Y
R
and Y
A
.
These losses are associates to several factors such as: cells operating out of the STCs; voltage drop in the DC
cables and protection diodes; dirt; partial shade; dispersion of parameters among the PV modules; operation
voltage out of the maximumpower point (MPP); spectrumand angle of incidence. It is observed that many of the
mentioned aspects cannot totally be represented in terms of simulations. Only through the aid of experimental
data it is possible to analyze the magnitude of the losses involved in those systems, for the subsequent
improvement of the forecasts accomplished in the project stage. Table II displays the values of Y
A
and L
C
obtained for each one of the generators here analyzed.
It is noticed that the largest array yields observed in annual terms were the ones regarding the generators of the
groups N2 and N4. However, those values are not very distant of the other ones, what is reected in daily array
yields somewhat close. Consequently, very similar capture losses can be observed, reaching 256% of the Y
R
in
the group N1, which corresponds to the generator with larger shadow inuence, and 20% of the Y
R
in the
generators belonging to the groups N2 and N4, resulting in differences between the losses presented by the
different PV generators always inferior to 6%, with average capture losses equal to 22% of Y
R
.
INVERTERS OPERATIONAL ANALYSIS
Maximum power point tracking (MPPT)
The parameter generally used to evaluate the functioning of the inverter close to the MPP is the MPPTefciency,
h
MPPT
. That parameter can be dened as the ratio between the energy obtained by the inverter of a given PV
generator, and the energy that could be obtained of that same generator, if the inverter was provided with a MPPT
ideal system.
The difculty to evaluate this parameter is associated both to its dependence of internal factors of the inverter,
for example, the MPPTs algorithm, and the external factors, such as photovoltaic generator, irradiance and
temperature.
1618
The method used in this paper consist in monitoring parameters of interest by a Sunny Data
Control Plus that have and accuracy of 05%. The irradiance (H
(t,b)
) is measured with calibrated reference cell
and cells temperature (T
C
) is obtained by temperature sensors located at modules back side. With these two
parameters is possible to calculated de power of each generators in the MPP using the Equation 1.
P
mp
P
0
FV

H
t;b
H
STC
1 l
mp
T
C
T
C;STC
; (1)
where H
STC
and T
(C, STC)
are, respectively, the irradiance and cell temperature on STC (1000 W/m
2
, 258C). Note
that the quality of the results obtained for h
MPPT
depends of quality of the measured data H
(t,b)
and T
C
. In this
Table II. Inverters relative capacity P
0
Inv
=P
0
PV
, annual array yield (Y
A
) and capture losses (L
C
) for each one of the
generators, regarding the eight analyzed groups
Group N1 N2 N3 N4 N5 N6 N7 N8
P
0
Inv
=P
0
PV
(%) 1020 998 555 570 600 680 842 830
Y
A
(kWh/kWp/year) 1146 1234 1197 1234 1182 1231 1211 1188
Y
A
(kWh/kWp/day) 313 337 328 337 323 336 331 325
L
C
(% Y
R
) 256 200 223 200 233 201 214 229
Copyright # 2006 John Wiley & Sons, Ltd. Prog. Photovolt: Res. Appl. 2007; 15:337352
DOI: 10.1002/pip
PHOTOVOLTAIC SYSTEM WITH DIFFERENT ISF 339
Figure 1. Variation of the DC power, calculated ( P
mp
) and measured ( P
PV
) delivered to the inverter along the day, for the
groups N1N8
Copyright # 2006 John Wiley & Sons, Ltd. Prog. Photovolt: Res. Appl. 2007; 15:337352
DOI: 10.1002/pip
340 W. N. MACE

DO AND R. ZILLES
case a good relationship between practicability and reliability of the data provide consistent results that have
coherence when compared with results existing in the literature.
1618
Figure 1 displays the power values calculated ( P
mp
) and measured ( P
PV
) along 1 day, while Figures 2 and 3
show the variation of the h
MPPT
with the DC power delivered to the inverter along the same day for each one of
the eight groups analyzed. It is possible to observe the effect of the power limitation, mainly in the groups N3 and
N4, and its inuence in the MPPT.
From experimental data, as can be seen in Figures 2 and 3, was possible to verify that the inverters works, in
almost the totality of time, with MPPTs efciency between 70 and 98%. The existing differences, performance
better in the morning than in the afternoon, can be associated to the fact that the DC power of the inverter
depends of its MPPT, which depends on the temperature of the inverter and of the conguration of the
photovoltaic generator in terms of operation voltage and current. Thus, the differences in the values obtained for
MPPTs efciency, can be associated to the inverters temperature differences between morning, when the
equipment is colder, and the afternoon when it is hotter.
In spite of the inverters used in this work are of the same model, it is important to mention that the inverters in
Figure 2 were manufactured in the year 2003 and the others, in Figure 3, were manufactured in the year 2000,
what can explain, in part, the observed performance differences. Note itself that the behavior of the group N7s
inverter is different from the others (about 10% smaller MPPT morning efciency), which can be associate to a
MPPTs system not sufciently optimized, what is observed by means of a larger differences between DC
measured and calculated voltages (Figure 3).
It should be emphasized that the daily average efciencies found, h
MPPT
, vary between 90 and 93% in days of
clear sky, depending on the conguration. On the other hand, in cloudy days those values can fall to 85% or less,
depending on the prole of the irradiance along the day.
Energy efciency
The power losses in the inverter can be interpreted by the sum of three components, which are: the self-
consumption losses, P
self
(W); the lineal losses with the electric current (voltage drops in the semiconductors,
etc.), K
1
P
out
; and the losses proportional to the squared electric current (resistive losses, etc.), K
2
( P
out
)
2
.
11
Thus, the losses in Watts are given by the Equation 2.
P
losses
P
self
K
1
P
out
K
2
P
out

2
(2)
Normalizing the Equation 1 by the rated power of the inverter P
0
Inv
, making the p
out
P
out
=P
0
Inv
and
considering k
0
P
self
=P
0
Inv
, k
1
K
1
and k
2
K
2
P
0
Inv
, the Equation 3 is obtained:
p
losses
k
0
k
1
p
out
k
2
p
out

2
(3)
The great advantage of this equation is that it makes possible to characterize the behavior of the losses in the
inverter and, consequently, its energy efciency, based in only three non-dimensional parameters, k
0
, k
1
, and k
2
,
which can be obtained experimentally, such as the ones displayed in Table III, where the characteristic
parameters and the average annual energy efciency of the inverter, h
Inv
, are shown associated to each one of the
groups experimentally analyzed. It is observed that those values also reect the features of each conguration in
terms of input voltage and of installed Wp.
Some experimental data obtained for an inverter with the same characteristics of the inverter studied here, for
three different values of input voltage, are presented in Table IV.
The main differences between the parameters of the two last tables above are due to the way how they were
originated. Table IVwas obtained in laboratory conditions, with a DCsource to maintain stable the input voltage
while the operation power was varied. On the other hand, Table III was obtained from eld operation data, 5 min
averages, of each one of the groups N1N8, considering a very larger number of situations regarding the voltages
and operation conditions. Figures 4 and 5 illustrate the working range of the input voltage (V
DC
), the
distributions of the power losses ( P
losses
) and the respective accumulated energy losses, for each one of the
analyzed groups.
Copyright # 2006 John Wiley & Sons, Ltd. Prog. Photovolt: Res. Appl. 2007; 15:337352
DOI: 10.1002/pip
PHOTOVOLTAIC SYSTEM WITH DIFFERENT ISF 341
Figure 2. Variation of the maximum power point tracking efciency (h
MPPT
) in function of the DC power ( P
PV
) extracted of
the PV generators and variation of the DC voltage (V
DC
) along the day, groups N1N4
Copyright # 2006 John Wiley & Sons, Ltd. Prog. Photovolt: Res. Appl. 2007; 15:337352
DOI: 10.1002/pip
342 W. N. MACE

DO AND R. ZILLES
Figure 3. Variation of the maximum power point tracking efciency (h
MPPT
) in function of the DC power ( P
PV
) extracted of
the PV generators, and variation of the DC voltage (V
DC
) along the day, groups N5N8
Copyright # 2006 John Wiley & Sons, Ltd. Prog. Photovolt: Res. Appl. 2007; 15:337352
DOI: 10.1002/pip
PHOTOVOLTAIC SYSTEM WITH DIFFERENT ISF 343
As can be seen in Figures 4 and 5, the groups N5 and N6 present a more concentrated distribution because they
have a small number of modules connected in series, with a larger prominence for the group N6, which presents
only 10 modules in series. Particularly, the operation voltages of those groups are closer to the necessary
minimum value for starting the inverter operation (134 V
DC
for 220 V
CA
). It is observed that approximately in
80%of the operation time the inverters input voltage of the group N6 is in 150 V
DC
, also implying in almost 90%
of the accumulated losses of energy at that level of operation voltage.
It is also important to observe that the shape of the power losses distribution ( P
losses
) is similar in all the
groups, with a small difference in the groups with large installed power, where power losses of 125 W are
observed. Those results show that besides the operation voltage, the relationship P
0
Inv
=P
0
PV
also interferes in the
characteristic parameters of the inverter.
In the right side of Figures 4 and 5, the power losses increase of the inverters N3, N4, and N5 are related with
power limitation process. On the other hand, power losses bar under the 15 W class has a considerable
percentage of occurrence (between 19 and 26%of operation in all inverters). This value of 15 Wis, depending on
the group, more than 90%higher than the inverters self consumption (as seen fromTable III) and is associated to
low level of AC power ( P
out
between 1 and 20% of the inverter rated power) that can occurs in the rst hours of
sunshine and sunset or in the moments with high level of shade; therefore, considerable operation in 15 W class
is related to considerable operation time in low power.
Verifying the relevance of using a parameter or another, the efciency h
Inv
was calculated using the
parameters shown in Table IV, regarding the levels of tension of 150 and 250 V
DC
, and that was compared with
the calculated values, taking the parameters in Table III, regarding the groups N6 and N7, which operate
respectively close to the mentioned tensions. These two groups were selected because they present in their
distributions a signicant number of occurrences of voltages close to 150 and 250 V
DC
, just as displayed in
Figure 5.
From that analysis, the relative errors with respect to the experimentally obtained values were extracted.
Figure 6 summarizes the results, in which it is observed, for both cases, quite high relative errors for
power-output values inferior to 10%of the rated power of the inverter, and that mistake lower to 5%are found for
power superior to 50% of the rated power of the inverter. When discussing the compared results between Tables
III and IV, must be considered that the Table III results from averaged measurements (5 min averages), whereas
Table IV (laboratory measurements) probably have been obtained from instantaneous measurements (DC
controlled source) than to Table IV and Figure 6 (Calculated-IES) h
Inv
represent the inverter instantaneous
Table III. Characteristic parameters, self-consumption and average annual energy efciency of the inverter
Group k
0
(%) k
1
(%) k
2
(%) P
self
(W) h
Inv
(%)
N1 081 423 400 81 914
N2 079 411 500 79 915
N3 095 232 700 95 913
N4 098 223 700 98 913
N5 097 235 700 97 913
N6 081 329 600 81 915
N7 082 397 500 82 913
N8 086 372 600 86 912
Table IV. Characteristic parameters and self-consumption
19
V
DC
(V) k
0
(%) k
1
(%) k
2
(%) P
self
(W)
150 11 04 64 11
250 13 24 48 13
330 15 43 36 15
Copyright # 2006 John Wiley & Sons, Ltd. Prog. Photovolt: Res. Appl. 2007; 15:337352
DOI: 10.1002/pip
344 W. N. MACE

DO AND R. ZILLES
Figure 4. Distribution of the operation voltage (V
DC
) and power losses ( P
losses
), as well as the accumulated energy losses in
2004 (Groups N1N4)
Copyright # 2006 John Wiley & Sons, Ltd. Prog. Photovolt: Res. Appl. 2007; 15:337352
DOI: 10.1002/pip
PHOTOVOLTAIC SYSTEM WITH DIFFERENT ISF 345
Figure 5. Distribution of the operation voltage (V
DC
) and power losses ( Plosses), as well as the accumulated energy losses in
2004 (Groups N5N8)
Copyright # 2006 John Wiley & Sons, Ltd. Prog. Photovolt: Res. Appl. 2007; 15:337352
DOI: 10.1002/pip
346 W. N. MACE

DO AND R. ZILLES
efciency. It is observed that both parameters are well adjusted to the inverters efciency curve, what implicates
to say that anyone of them characterizes its behavior in a satisfactory way.
Power limitation
Figure 7 shows a real example of how the power limitation works in the inverter for the groups N3, N4, and N5,
corresponding to the larger-installed-power ones and where that process is more critical.
Over-sizing the PV generator forces the manufacturers of inverters to build their equipments with protection
systems that impede its input DC power ( P
PV
) of overcoming the allowed maximum value, protecting them.
However, it is important to mention that such control implies in losses due to power limitation, which, in
agreement with the degree of the over-sizing and of the available solar resource in each place, can be quite
signicant.
Figure 6. Comparison between the efciency values calculated with the parameters obtained by the Instituto de Energ a
Solar (IES) of the UPM and by the Laboratorio de Sistemas Fotovoltaicos (LSF) of the IEE/USP and their respective
percentile errors regarding the obtained values of real efciencies
Copyright # 2006 John Wiley & Sons, Ltd. Prog. Photovolt: Res. Appl. 2007; 15:337352
DOI: 10.1002/pip
PHOTOVOLTAIC SYSTEM WITH DIFFERENT ISF 347
The limitation is characterized by the divergence of the PV generators MPP (voltage in the MPP) as a
function of the input power and/or temperature of the inverter (see Figure 7).
It is noticed that when the inverters input DC power reaches a certain value (approximately 1200 W), the
inverter starts limiting the power in that value. After some operation time in that condition, the inverters
Figure 7. Characterization of the of the inverters power limitation process (groups N3, N4, and N5)
Copyright # 2006 John Wiley & Sons, Ltd. Prog. Photovolt: Res. Appl. 2007; 15:337352
DOI: 10.1002/pip
348 W. N. MACE

DO AND R. ZILLES
temperature increases, and another process to maintain the temperature in a maximum constant allowed value
becomes necessary.
It is important to advice that problemas over-heating and software failures can impede the inverter of reaching
its rated power AC, just as veried by Keller.
5
In spite of not being this the case in the performed experiment, it is
observed in the groups N3, N4, and N5, where there is 1802, 1757, and 1663 Wp for each inverter of 1 kW, that in
days of clear sky, as can be see in Figure 7, the losses for limitation reaches 12, 10, and 4%, respectively. Figure 8
gives another example of the process of power limitation, for another day also with a good irradiance level,
where the variation of the solar resource is shown together with the variation of the power delivered by the
inverter for the eight inverter-generator congurations monitored in this experiment. It is noticed that in that
other situation there is only limitation on the groups N3 and N4, in other words, those with P
0
Inv
=P
0
PV
< 60%.
Figure 8. Inverters power output, power limitation in 1100 W AC, and the irradiance along the day: (a)generators of
mc-Si and (a)generators of c-Si
Table V. Production, performance ratio and nal yields, regarding 2004, for each group inverter-generator
Group N1 N2 N3 N4 N5 N6 N7 N8 GCPVS
VM (kWp) 0978 1002 1802 1757 1663 1477 1188 1201 11070
P
0
Inv
(kW) 10 10 10 10 10 10 10 10 80
P
0
Inv
=P
0
PV
(%) 1020 998 555 569 600 677 894 833 725
PA (kWh) 1026 1106 1971 1981 1794 1665 1255 1301 12 099
PR (%) 681 716 710 732 700 746 728 703 710
Y
F
(kWh/kWp) 1050 1104 1094 1127 1077 1150 1127 1083 1091
Copyright # 2006 John Wiley & Sons, Ltd. Prog. Photovolt: Res. Appl. 2007; 15:337352
DOI: 10.1002/pip
PHOTOVOLTAIC SYSTEM WITH DIFFERENT ISF 349
Figure 9. Seasonal variation of the performance ratio (PR) and nal yields (Y
F
) of each group and of the system as a whole,
along 2004
Copyright # 2006 John Wiley & Sons, Ltd. Prog. Photovolt: Res. Appl. 2007; 15:337352
DOI: 10.1002/pip
350 W. N. MACE

DO AND R. ZILLES
ANALYSIS OF THE INFLUENCE OF THE INVERTER-GENERATOR
RELATIVE POWER (P
0
Inv
/P
0
PV
)
The annual production (PA), for 2004, of each inverter-generator group with different P
0
Inv
=P
0
PV
relationships can
be visualized in Table V. This table also presents, among others, the measured values of the installed power of
each PV generator (VM) and the correspondent nal yields (Y
F
).
According to the Y
F
of each conguration, it is observed that the differences between one and another Y
F
,
regarding the different P
0
Inv
=P
0
PV
relationships, is not signicant and always being inferior or equal to 100 kWh/
kWp in 2004. That aspect is very important, because it provides a high degree of freedom in the project stage,
making it signicantly easy.
The parameters PR and Y
F
are typically reproduced, in an annual base, such as presented in Table V, or
monthly, how illustrates in Figure 9, where the seasonal variations of those parameters are displayed.
It is observed that seasonally the Y
F
parameter is the more inuenced due to its dependence of the solar
resource, and its monthly measurements usually are located between 65 and 117 kWh/kWp, regardless the
conguration. With respect the PR, it is noticed that the measurements tend to be larger in the winter than in the
summer, and they stay between 60 and 80%, with an annual average value, obtained for the GCPVS as a whole,
of approximately 71%. It is also veried that PR is less inuenced by the seasonal variations, since that
parameter is normalized regarding the solar resource. However, it is noticed that its values are inuenced by the
variations of the ambient temperature along the year.
It can be observed that, even in the months with grater solar resource availability, the productivity of the more
oversized groups, N3 and N4, do not differ signicantly when compared to the groups with smaller installed
power, N1 and N2, and the productivity differences among the several groups are always inferior to 10 kWh/
kWp per month.
FINAL COMMENTS
The tendency of over-sizing excessively the PV generator in relation to the inverter still exists, and this
procedure can affect the inverters operational lifetime. In agreement with the obtained experimental data, once
that takes the care which the inverter does not be excessively small in relation to the PV generator,
P
0
Inv
=P
0
PV
55%, the quality of the inverter is more important than the relationship between the inverters and
the PV generators powers. Regarding the above observations, the utilization of P
0
Inv
=P
0
PV
values between 60 and
100% should be recommended, depending on the location, of the type of installation and on the equipment kind.
REFERENCES
1. Macagnan MH, Lorenzo E. On the optimal size of inverters for grid connected PV systems. Proceedings of the 11th
European Photovoltaic Solar Energy Conference, Montreux, Switzeland, 1992; 11671170.
2. Louche A, Nortton G, Poggi P, Peri G. Global approach for an optimal grid connected PV system sizing. Proceedings of
the 12th European Photovoltaic Solar Energy Conference, Amsterdam, Netherlands, 1994; 16381641.
3. Peippo K, Lund PD. Optimal sizing of grid connected PV-systems for different climates and array orientations: a
simulation study. Solar Energy Materials and Solar Cells 1994; 35: 445451.
4. Peippo K, Lund PD. Optimal sizing of solar array and inverter in grid connected photovoltaic systems. Solar Energy
Materials and Solar Cells 1994; 32: 95114.
5. Keller L, Affolter P. Optimizing the panel area of a photovoltaic system in relation to the static inverterspractical
results. Solar Energy 1995; 55(1): 17.
6. Schalkwijk M, Kil AJ, Weiden TCJ, Paes PS. Undersizing of inverters: modeling and monitoring results of 15 PV/
Inverter units in Portugal and Netherlands. Proceedings of the 14th European Photovoltaic Solar Energy Conference,
Barcelona, Spain, 1997; 22292232.
7. Nofuentes G, Almonacid G. An approach to the selection of the inverter for architecturally integrated photovoltaic
grid-connected systems. Renewable Energy 1998; 15: 487490.
Copyright # 2006 John Wiley & Sons, Ltd. Prog. Photovolt: Res. Appl. 2007; 15:337352
DOI: 10.1002/pip
PHOTOVOLTAIC SYSTEM WITH DIFFERENT ISF 351
8. Fraunhofer institute for Solar Energy Systems (FISES). Alitle more wont hurt: in the past, inverters were often designed
too small. PHOTON International 2004; September: 6267.
9. Lorenzo E. Retratos de la conexion fotovoltaica a la red (V): De la AIE a los inversores. Era Solar 2005; 23(126): 5258.
10. Burger B, Ruther R. Inverter sizing of grid-connected photovoltaic systems in the light of local solar resource
distribution characteristics and temperature. Solar Energy 2006; 80: 3245.
11. Martin EC. Edicios Fotovoltaicos Conectados a la Red Eletrica: Caracterizacion e Analisis. Phd Thesis, Universidad
Politecnica de Madrid, 1998.
12. Jahn U, Nasse W. Operation performance of grid-connected PV systems on buildings in Germany. Progress in
Photovoltaics: Research and Applications 2004; 12: 441448.
13. Otani K, Kato K, Takashima T, Yamaguchi T, Sakuta K. Field experience with large-scale implementation of domestic
PV systems and with large PV systems on buildings in Japapn. Progress in Photovoltaics: Research and Applications
2004; 12: 449459.
14. Moore L, Post H, Hayden H, Canada S, Narang D. Photovoltaic power plant experience at Arizona Public Service: a
5-year assessment. Progress in Photovoltaics: Research and Applications 2005; 13: 353363.
15. Caamano E, Lorenzo E, Zilles R. Quality control of wide collections of PV modules: lessons learned from the IES
experience. Progress in Photovoltaics: Research and Applications 1999; 7: 137149.
16. Haeberlin H. A new apprach for semi-automated measurement of PV inverters, especically MPP Tracking Efciency.
Proceedings of the 19th European Photovoltaic Solar Energy Conference, Paris, Frace, 2004.
17. Haeberlin H, Borgna L, Kaempfer M. Total effciencya new quantity for better characterisation of grid-connected PV
inverters. Proceedings of the 20th European Photovoltaic Solar Energy Conference, Barcelona, Spain, 2005.
18. Abella MA, Chenlo F. Choosing the right inverter for grid-connected PV systems. Renewable Energy, v.March-
April2004.
19. IES-LS-I-0511. Informe de medida de inversores SUNNY BOY 1100/SMA Technologie. Abril de2005.
Copyright # 2006 John Wiley & Sons, Ltd. Prog. Photovolt: Res. Appl. 2007; 15:337352
DOI: 10.1002/pip
352 W. N. MACE

DO AND R. ZILLES