Content uploaded by Dominik Keiner
Author content
All content in this area was uploaded by Dominik Keiner on Oct 24, 2024
Content may be subject to copyright.
Assessing Yield Disparities: Anticipated versus Optimal
Rooftop Photovoltaic Systems and Implications for
Prosumer Viability
Dominik Keiner,* Dmitrii Bogdanov, Stefan Krauter, and Christian Breyer*
1. Introduction
Solar photovoltaics (PV) has been crowned as the “new king”by
the International Energy Agency (IEA).
[1]
Distributed energy
sources such as rooftop solar PV are expected to be a major pillar
in the ongoing energy transition.
[2]
The global potential for elec-
tricity generation from rooftop solar PV alone can be estimated to
about 27 000 TWh.
[3]
This importance is accentuated by ongoing
solar PV installations. In 2022, distributed solar PV installations
in the top 10 markets were 89.9 GW
p
,
which is 38.1% of the total installed solar
PV capacity that year.
[4]
Energy system tran-
sition research indicates the important role
of solar PV prosumers in the future. For
example, the global results of Bogdanov
et al.
[5]
estimate more than 15% of the
global installed solar PV capacity to be
assigned to solar PV prosumers by 2050.
Jacobson et al.
[6]
found a share of solar
PV prosumers in total installed PV capaci-
ties of about 39% in 2050, however, for a
lower total installed solar PV capacity as
in Bogdanov et al.
[5]
Higher installations
shares nowadays compared to partly lower
shares estimated for the future underline
the important role of rooftop solar PV for
the early phases of the global energy tran-
sition. As a fact, while rooftop solar PV
dominated the global solar PV market in
the 1990s and 2000s,
[4]
the share of rooftop
solar PV has stayed more or less constant
at around 50% since the renewable energy
installations see a strong uptake.
[7]
Rooftop
solar PV prosumerism,
[8]
that is, the use of
rooftop solar PV electricity within the household for covering
power demand, heat demand, charging the electric vehicle, as
well feeding electricity to the grid, has been shown to be benefi-
cial worldwide.
[9,10]
Prosumers are of high importance for the
overall energy transition.
[11]
Research suggests that distributed
renewable energy and prosumerism might help to support the
energy transition in societies.
[12,13]
There are numerous studies available assessing the perfor-
mance and expected yield of rooftop PV systems for different
sites, countries, or regions globally. Only very few studies cover
a comparative analysis of ground-mounted versus rooftop solar
PV power plants. Awan et al.
[14]
compared the performance of
ground-mounted versus rooftop solar PV for different row spac-
ing in a hot environment. In their approach, the energy saved for
cooling of the building is added to the electricity yield of the roof-
top PV system, leading to a better performance of the rooftop PV
system compared to the ground-mounted counterpart for the
overall energy balance. Bansal et al.
[15]
reviewed the different per-
formances of fixed and single-axis tracking solar PV power
plants, while except for degradation, the assessment is made
without differentiation of ground-mounted and rooftop solar
PV, leaving out major differences in these two system options.
Apart from these rare examples, no study could be identified
D. Keiner, D. Bogdanov, C. Breyer
School of Energy Systems
LUT University
Yliopistonkatu 34, 53850 Lappeenranta, Finland
E-mail: dominik.keiner@lut.fi; christian.breyer@lut.fi
S. Krauter
Faculty of Computer Science, Electrical Engineering and Mathematics,
Electrical Energy Technology-Sustainable Energy Concepts (EET-NEK)
Paderborn University
Warburger Str. 100, 33098 Paderborn, Germany
The ORCID identification number(s) for the author(s) of this article
can be found under https://doi.org/10.1002/solr.202400435.
DOI: 10.1002/solr.202400435
Solar photovoltaics, especially rooftop systems also called distributed solar
photovoltaics, are crucial in the ongoing energy transition. Modeling these
systems is vital to understanding their role in a decentralized energy system.
While ground-mounted photovoltaic power plants are easier to model, gener-
alizing yield profiles for rooftop systems is challenging. This study aims to
estimate yield loss effects for rooftop solar photovoltaic systems compared to
optimized ground-mounted systems. Anticipated yield losses are 18% for resi-
dential, 7% for commercial, and 4% for industrial rooftop systems. The impact on
residential prosumers’viability is assessed by comparing prosumer system
optimization results with and without yield losses. Results show a non-uniform
change in installed solar photovoltaic and battery capacities, with a tendency to
compensate for reduced yields by increasing photovoltaic capacity by up to 20%,
given favorable cost prospects by 2050. The annualized total cost of energy for
prosumer households could therefore increase by up to 20% by 2050. Despite
yield reductions, installing a solar photovoltaic prosumer system remains more
favorable than relying entirely on-grid electricity. This study highlights the
importance of considering yield losses in rooftop solar photovoltaics and the
significant role of prosumers despite identified yield losses.
RESEARCH ARTICLE
www.solar-rrl.com
Sol. RRL 2024, 2400435 2400435 (1 of 12) © 2024 Wiley-VCH GmbH
covering a dedicated overview on the yield differences between
the optimal and anticipated yield of rooftop solar PV plants ver-
sus ground-mounted utility-scale solar PV power plants.
In order to show pathways for the global energy transition,
energy system transition modeling is an important research area.
However, a dedicated modeling of solar PV prosumers is
included only in a few models, and if included, often insuffi-
ciently implemented. In the most used energy system model
for highly renewable energy systems
[16]
EnergyPLAN,
[17]
solar
PV resources are aggregated into one technology option, requir-
ing additional efforts for a differentiated modeling of various PV
options, such as utility-scale and rooftop solar PV. Some versions
of the TIMES model,
[18]
explicitly JRC-EU-TIMES
[19]
differentiate
utility-scale, high concentration, and rooftop solar PV. The
GENeSYS-MOD or OSeMOSYS model
[20,21]
also differentiate
the utility-scale, and rooftop solar PV.
[22]
However, TIMES and
GENeSYS-MOD/OSeMOSYS models use the time slices
approach instead of full hourly resolution, which may affect
the system optimization, especially short- and long-term energy
storage in the case of high shares of RE. Only the GATOR-
GCMOM/LOADMATCH
[23]
and LUT-ESTM
[5]
tools explicitly
differentiate prosumers PV and optimize the system in full
hourly resolution. Neumann et al.
[24]
include rooftop solar PV
in addition to utility-scale solar PV by assuming 16% of the total
PV capacity without individual rooftop solar PV optimization or
further detailing on the share as part of the PyPSA framework,
whereas an improved method with distributed PV modeling is
available for PyPSA.
[25]
Some studies use rooftop solar PV; how-
ever, they do not provide minimum information on what basis it
is considered in their energy system modeling,
[26,27]
or consider
rooftop PV potential while not disclosing whether it was consid-
ered in the modeling.
[28]
Others do not provide information
whether rooftop and utility-scale solar PV is differentiated at
all.
[29–32]
This overview on major energy system modeling teams
and models used indicate substantial gaps in proper inclusion of
rooftop solar PV in the analyses, as some even ignore rooftop
solar PV at all, and differentiated information on the rooftop
solar PV segments and their yield modeling are largely missing
across all major models and teams.
The photovoltaic power systems programmed of the IEA (IEA-
PVPS) Task 13 “Reliability and Performance of Photovoltaic
Systems”provides a set of reports assessing performance issues
of solar PV, including uncertainty in yield assessments
[33,34]
and
performance assessment of new solar PV systems.
[35]
A dedi-
cated study on different yield loss effects to assess the yield dis-
parity for rooftop PV systems in comparison to ground-mounted
systems on global scale is not yet available. Furthermore, an
assessment of the impact of the reduced yield on the techno-
economic performance of solar PV prosumer systems is not
yet available. This study aims to close these research gaps.
The novelties of this study comprise: 1) Global-local rooftop
PV yield disparity assessment of the anticipated rooftop PV yield
in comparison to the optimal yield of ground-mounted utility-
scale solar PV power plants. 2) Assessment of the system impact
of cost-optimized residential solar PV prosumer systems apply-
ing the optimal yield and anticipated yield to point out the impor-
tance of detailed consideration of solar PV prosumers in energy
system transition modeling.
This study aims to contribute new knowledge to the energy
system modeling research to correctly assess the role of solar
PV prosumers in detailed energy system transition research.
The identified values for yield losses of rooftop solar PV systems
versus easier to model ground-mounted solar PV systems aims
to improve the consideration of rooftop solar PV in future energy
system models with high spatial resolution and extent, which do
not yet consider dedicated rooftop solar PV resources.
2. Methods and Data
2.1. Loss Effects in Rooftop Solar Photovoltaic Systems
There are several loss effects identified for residential solar PV
rooftop systems. As structured in this sub-section, they can be
assigned to an environmental, cell and module, and system level.
The sub-section also provides information on the adaption of res-
idential yield losses to commercial and industrial system sizes.
The losses are compared on a normalized level to available data
from Killinger et al.
[36]
Therefore, the choice of values mentioned
in the following sub-sections is done having the target value in
mind. The comparison of total yield losses based on these values
and the yield results by Killinger et al.
[36]
however, is presented in
the Section 4.1. This study is carried out assuming that losses
from curtailment are not yet significantly important for rooftop
solar PV systems.
2.1.1. Environmental Level
This level consists of the orientation of the modules, so nonopti-
mal tilt or azimuth of the solar PV modules, and shading by
houses, structures, vegetation, urban obstacles, etc. Therefore,
the environmental level includes causes of yield reduction based
on the reduced irradiance on the modules.
The effect of nonoptimal orientation is estimated based on
irradiance data for several sites around the world. Table 1 lists
all studies included in the estimation of nonoptimal orientation
impact.
The polar contour plots for the tilt angle and azimuth provided
in the studies allow for the estimation of a worst-case effect of
5% yield reduction. The worst-case would be an orientation
toward East or West instead of facing the equator. A typical
Table 1. Studies included for estimating the nonoptimal module
orientation based on polar contour plots provided.
Study Year Location
Rowlands et al.
[74]
2011 Ontaria, Canada
Byrd
[75]
2012 Auckland, New Zealand
Khoo et al.
[76]
2014 Singapore
Hartner et al.
[77]
2015 Vienna, Austria
Matthiss et al.
[78]
2015 Widderstall, Germany
Kichou et al.
[79]
2019 Prague, Czechia
Oh and Park
[80]
2019 Seoul, South Korea
Yu et al.
[81]
2019 Tokyo, Japan
www.advancedsciencenews.com www.solar-rrl.com
Sol. RRL 2024, 2400435 2400435 (2 of 12) © 2024 Wiley-VCH GmbH
2367198x, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/solr.202400435 by Dominik Keiner - Lappeenranta University Of Technology , Wiley Online Library on [24/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
rooftop tilt angle of 25°–40° is presumed. As rooftop PV mod-
ules face on average different directions, however, with a ten-
dency toward the equator, an average effect of 2% for
nonoptimal module orientation is chosen.
Yield losses by obstacles, structures, vegetation, etc. are always
case-specific for each rooftop PV system, and are therefore chal-
lenging to estimate on a general basis. Another problem is that
such effects are rarely reported in literature. Killinger et al.
[36]
estimate 7.4% irradiation losses in Uppsala, Sweden.
Trzmiel et al.
[37]
report 7.5–11% irradiation losses for a residen-
tial house in Poland. Berardi and Graham
[38]
estimate up to 30%
losses in a worst-case scenario for the case of Toronto, Canada.
For this study, an effect of 7.5% is chosen for residential solar PV
irradiation loss due to obstacles. This loss only includes shading
that reduces the irradiation on the modules but still providing
enough irradiation for the modules to work. Shading that acti-
vates the bypass diodes of the module is assigned to the system
level in Section 2.1.3.
2.1.2. Cell and Module Level
The cell and module level refers to the physics of the solar PV
module. The only effect that is different from a utility-scale power
plant is the reduced ventilation of the modules, since they are
usually mounted directly to the rooftop. Again, such effects
are rarely studied for rooftop systems. For non-ventilated
façade systems in Stockholm, Sweden, London, Great Britain,
and Madrid, Spain, Yun et al.
[39]
estimated 2.5% yield reduction
due to higher cell temperatures. A nonventilated rooftop system
studied by Ritzen et al.
[40]
in the Netherlands showed 2.6%
yield reduction related to the module temperatures. The impacts
of a green roof versus concrete and the impact of wind versus no
wind for mounted modules on a rooftop in Bucaramanga,
Colombia was tested to be 2–3% 0.4% by Osma-Pinto and
Ordó˜nez-Plata.
[41]
Therefore, a global average of 2.5% yield losses
due to less sufficient module cooling on rooftops is chosen.
2.1.3. System Level
On the system level, the inverter performance plays a main role.
One loss factor might be the insufficient maximum power point
tracking of inverters if modules are partially shaded. Bründlinger
et al.
[42]
found this effect to reduce the yield by 2%. Another
issue is the inverter sizing. According to Kouro et al.
[43]
bigger
inverters show higher efficiencies than smaller ones, but this
effect is rather small with 1% yield loss. The biggest contribution
at this level, however, is the fact that smaller PV systems consist
of only a few strings. In the case that only one module is shaded
in such a way, that the full module is not operable, the whole
module string is not able to work accordingly. This effect can
only be estimated indirectly, that is, by equipping each module
with its own microinverter. No evidence could be found for a
large penetration of microinverters, also called module inverters,
for the foreseeable future. The studies done by Dong et al.
[44]
and
Mohd
[45]
allow for a yield reduction of 5–10%. In this study, an
average value of 7.5% is chosen.
2.1.4. Adaption for Commercial and Industrial Rooftop Systems
Solar PV prosumers cover residential houses as well as commer-
cial buildings (retailers, stores, supermarkets, etc.) and industrial
buildings (big manufacturing buildings, large logistic buildings).
The loss effects obtained for residential systems must therefore
be adapted to different applications. Unfortunately, due to the
lack of data allowing for a comparison of ground-mounted solar
PV simulation data to rooftop solar PV yield data as done for res-
idential systems, this must be done via an educated guess.
The nonoptimal orientation will in some cases be significantly
improved, for example, on flat roofs with mounting structures.
Therefore, this impact is assumed to be improved to 1% yield loss
for both commercial and industrial systems. Reduction in irradi-
ance from obstacles or vegetation, however, can be presumed to
have a much less impact for commercial and industrial systems
than for residential rooftops, since these systems are more situated
in industrial areas, or in areas with more spacing between build-
ings (e.g., for parking lots or similar). Furthermore, obstacles on
the roofs such as chimneys are not as much a problem than for
residential rooftop systems. This effect is chosen to contribute 2%
yield loss for commercial systems and 1% yield loss for industrial
systems.
Insignificant ventilation still plays a role, especially for commer-
cial systems, as many of these systems might not be installed on
flat roofs with a mounting structure, but on the rooftop itself.
Industrial systems can be presumed to be mostly flat roofs requir-
ing a mounting structure, which improves ventilation of the solar
PV modules. For commercial systems, this effect is only slightly
reduced to 2% compared to residential systems. For industrial
systems a contribution of 1% is assumed to account for higher
temperatures due to rooftop material other than bare ground.
One major issue of only a few strings in the system will be
significantly improved for both commercial and industrial sys-
tems due to the bigger capacities installed, allowing for, and
requiring more strings. Nevertheless, this effect will not
completely vanish. Therefore, a 2% yield reduction effect for
commercial systems and 1% for industrial systems is considered.
2.2. Other Loss Effects
Other causes for yield losses that impact rooftop solar PV sys-
tems equally as ground-mounted systems and, therefore, do
not cause an additional yield loss, are not included in this assess-
ment. Such loss affects are, for example, losses due to soiling of
the modules. The self-cleaning of the modules via precipita-
tion
[46,47]
is assumed to be equal for fixed-tilted, ground-mounted
solar PV power plants and rooftop systems. For other accumula-
tion of dirt on the modules, such as bird droppings,
[48]
no
evidence could be found in literature that rooftop solar PV sys-
tems would be more affected than ground-mounted systems. Air
pollution
[49]
and humidity or fog
[50]
could be factors for lower
rooftop solar PV systems, since rooftop systems are on average
located in metropolitan regions, which are commonly located
near the coast, in contrast to utility-scale solar PV power plants.
However, compared to the total rooftop PV installations of a
country or region, these effects might be negligible for
www.advancedsciencenews.com www.solar-rrl.com
Sol. RRL 2024, 2400435 2400435 (3 of 12) © 2024 Wiley-VCH GmbH
2367198x, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/solr.202400435 by Dominik Keiner - Lappeenranta University Of Technology , Wiley Online Library on [24/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
aggregated energy systems. Additionally, yield disparities due to
different degradation effects of rooftop solar PV systems com-
pared to ground-mounted systems are not covered in this study.
Degradation effects are investigated in ongoing research and
reports as of the IEA-PVPS Task 13 were able to present
insights
[51]
as well as academic research such as from Jordan
et al.
[52]
Herceg et al.
[53]
and Kaaya et al..
[54]
2.3. Solar Photovoltaic Yield and Prosumer Modeling
The calculation of the full load hours (FLH) per region is based
on a global 400 800 nodes gridded data (0.45° 0.45°) in full
hourly resolution and then aggregated with the method pre-
sented by Bogdanov and Breyer.
[55]
For the calculation of yield
profiles, an optimization of tilt angles for ground-mounted solar
PV power plants is done based on Breyer and Schmidt.
[56]
The
basis of the yield modeling is hourly global gridded data by NASA
for the year 2005,
[57,58]
reprocessed by the German Aerospace
Center.
[59]
Optimal ground-mounted solar PV power plants refer to opti-
mally tilted systems as used by Afanasyeva et al.
[60]
for compari-
son of single-axis tracking solar PV power plants to optimally
fixed-tilted solar PV power plants. The optimal tilt angles of
ground-mounted solar PV power plants have been calculated
based on an optimization of levelized cost of electricity (LCOE)
on high spatial resolution globally as done by Breyer and
Schmid.
[56]
Instead of using a rule-of-thumb for tilt angle optimi-
zation, the optimal tilt angle has been found in these studies by
calculating the irradiation on the titled surface of the PV modules
using the aforementioned resource data for different tilt angles
and subsequently calculating the LCOE for each tilt angle. The
optimal tilt angle in terms of lowest LCOE might differ from
the tilt angle with the highest global irradiation, as stated by
Breyer and Schmid.
[56]
Residential solar PV prosumers are modeled using the LUT-
PROSUME linear optimization model according to Keiner
et al.
[10]
The model structure with all its components is shown
in Figure 1.
The modeling details of the prosumer system are described by
Keiner et al.
[10]
including all techno-economic parameters and
demand numbers. Equally, the simulation is done for 145
regions as depicted in Figure 2. The regions are clustered in nine
major regions to simplify the classification of results and discus-
sion. Figure 2 also shows the FLH of optimal solar PV systems,
which means optimally tilted ground-mounted solar PV power
plants as used by Bogdanov et al.
[5]
To study the impact of reduced yield on solar PV prosumers,
the scenario including a grid connection and a standard car
(ON-STDC, cf. Keiner et al.
[10]
) is chosen for this study represent-
ing the most common system structure for residential solar PV
prosumers. The impact analysis is done by comparing the
installed rooftop PV capacity per average residential home, the
installed stationary battery capacity per average residential home,
and the annualized total cost of energy (ATCE). The ATCE
include the annualized cost for component investment (capital
expenditure), operational expenditure, and electricity drawn
from the grid (cf. Keiner et al.
[10]
).
3. Results
3.1. Optimal versus Anticipated Yield
In a recent study, Killinger et al.
[36]
provided detailed yield data
for rooftop PV systems from all over the globe. Table 2 shows the
Figure 1. Solar PV prosumer model structure including technology components and energy flows. STD-C, standard car (battery electric vehicle); V2H-C,
car with vehicle-to-home option (to be used as an electricity storage); WEL, water electrolyzer; HCP, hydrogen compressor; HES, hydrogen energy storage;
FC, fuel cell; PtH, power-to-heat (heat pump or heating cartridge); TES, thermal energy storage.
www.advancedsciencenews.com www.solar-rrl.com
Sol. RRL 2024, 2400435 2400435 (4 of 12) © 2024 Wiley-VCH GmbH
2367198x, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/solr.202400435 by Dominik Keiner - Lappeenranta University Of Technology , Wiley Online Library on [24/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
mean yield per installed capacity for the different countries
compared to the optimal ground-mounted yield according to
Bogdanov et al.
[5]
The comparison underlines the claim that yield disparities are
always location dependent. However, it can also be seen that most
of the regions show a yield deviationof at least 10%, except the UK
and Japan. Almost two thirds of the regions show a yield deviation
of at least 15%. Denmark shows the highest deviation of 31%. To
estimate a global average value, the average is taken weighted by
the number of systems, leading to 18% yield disparity of residen-
tial rooftop PV systems (rooftop PV systems ≤25 kW
p
).
Figure 3 shows the total anticipated rooftop PV yield versus
the optimal yield of ground-mounted solar PV power plants.
The yield loss effects chosen for the environmental, cell and
module, and system level based on available literature in
Section 3.2 are in line with the estimated global average value
of 18%. The reduction of the loss effects for commercial sys-
tems lead to a total discrepancy of 7% compared to optimal
ground-mounted systems. Industrial rooftop systems show
the smallest yield discrepancy with a total of 4%. The yield
differences of residential, commercial, and industrial rooftop
solar PV systems can be seen in Figure 4, if the aforementioned
Figure 2. Geographic overview on the 145 regions for simulations, clustered in nine major regions (top) and solar PV yield in optimal conditions for
ground-mounted power plants (bottom). MENA, Middle East and North Africa; SSA, sub-Saharan Africa; SAARC, South Asian Association for Regional
Cooperation.
www.advancedsciencenews.com www.solar-rrl.com
Sol. RRL 2024, 2400435 2400435 (5 of 12) © 2024 Wiley-VCH GmbH
2367198x, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/solr.202400435 by Dominik Keiner - Lappeenranta University Of Technology , Wiley Online Library on [24/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
yield discrepancy factors are applied to the optimal ground-
mounted yield as shown previously in Figure 3.
The relative yield difference applied to different regions globally
naturally leads to different total losses. Regions with high FLH
predominantly in the sunbelt may lose 250–360 FLH compared
to ground-mounted systems, while the absolute losses in the
Nordics may not exceed 150–200 FLH. The yield of rooftop PV
systems in the sunbelt with 18% yield reduction, however, still
exceeds the optimal yield of the Nordic regions (cf. Figure 2).
The sunbelt regions still show a yield of 1200–1800 FLH
whereas optimal ground-mounted systems in the Nordics have
yields of typically up to 1200 FLH. Numeric results for the optimal
and anticipated yield for 145 regions can be found in the
Supporting Information.
3.2. Impact on Prosumer Viability
The impact of the yield discrepancy on the yield of average sys-
tems in different regions presented in the previous subsection
leads to the question how prosumer viability is affected. This sub-
section presents differences in solar PV capacity, stationary bat-
tery capacity, and ATCE of residential PV prosumer households
in comparison.
3.2.1. Installed Solar PV Capacity
Presumable, less yield for rooftop PV systems will lead to higher
demand for solar PV capacity of prosumer households. However,
this question is more complex as the optimized system structure
of prosumers depends on several factors more. Figure 5 shows
the difference in solar PV capacity of residential PV prosumer
households if the yield losses are applied.
On the contrary to the presumption that less yield automati-
cally must lead to higher solar PV capacities, the results show that
in many regions less solar PV capacity is installed. Interestingly,
in more regions, less solar PV capacity is more beneficial than no
solar PV capacity at all (100% change) already in 2020, indicat-
ing that the cost advantages of solar PV compared to grid supply
of electricity is already favorable in many regions. However, a
higher solar PV capacity is only installed in a few regions, mainly
limited to South America, SSA, and Southeast Asia. This picture
already changes in 2030, where may regions are already close to
substitute 18% annual yield loss with up to 20% higher installed
solar PV capacities. Some regions would not yet have solar PV
capacities installed in 2020, but would have in 2030, either
reduce the capacity in 2030 or avoid installation at all due to
the yield losses. This effect is mostly evident in Eurasia. By
2040, negative changes get relatively rare. By 2050, almost all
sunbelt regions substitute the yield losses with higher installed
solar PV capacities, showing the very favorable economics of
solar PV by mid-century. Nordic countries, on the contrary,
Table 2. Comparison of Killinger et al.
[36]
yield numbers for rooftop solar PV systems ≤25 kWp to the optimally tilted, fixed ground-mounted yield number
as used in the LUT modeling (Bogdanov et al.
[5]
).
Region Killinger et al.
rooftop [kWh/kW
p
]
Number of systems LUT optimally tilted fixed,
ground-mounted [kWh/kW
p
]
Deviation [%]
Austria 1043 1000 1207 14
Belgium 922 15 000 1030 10
Denmark 784 2100 1131 31
France 1101 23 300 1317 16
Germany 862 5 885 000 1053 18
Italy 1143 5400 1455 21
Japan 1221 10 800 1340.5 9
Netherlands 853 6200 1030 17
UK 897 28 700 956 6
United States North 1014 11 900 1317 23
United States South 1432 73 100 1686.5 15
System-weighted average –– – 18
Figure 3. Total anticipated annual yield loss for residential, commercial,
and industrial rooftop solar PV systems. Orientation and shading refer to
the losses on environmental level, ventilation is the main reason for losses
on cell and module level, and the inverter and string configuration repre-
sents the system level.
www.advancedsciencenews.com www.solar-rrl.com
Sol. RRL 2024, 2400435 2400435 (6 of 12) © 2024 Wiley-VCH GmbH
2367198x, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/solr.202400435 by Dominik Keiner - Lappeenranta University Of Technology , Wiley Online Library on [24/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
are not able to substitute the yield losses with higher solar PV
capacities in an economically beneficial way, which can be seen
by small negative changes in northern Europe and central
Eurasia. In general, the cost reduction of solar PV in the future
by 2050 is promising enough that most regions can balance
lower yield compared to optimal systems with more installed
capacity. A prerequisite would be the available rooftop area.
Numeric differences in installed solar PV capacity for all regions
and years can be found in the Supporting Information.
3.2.2. Installed Stationary Battery Capacity
In the previous study by Keiner et al.
[10]
batteries have been iden-
tified to be the most flexible technology in terms of reacting to the
availability of the grid connection. As can be seen in Figure 6,
batteries also react in a very nonuniform way to the lower avail-
ability of rooftop solar PV yield.
In 2020, no change in stationary battery capacities is evident
due to the fact that in the chosen scenario, batteries are mainly
installed from 2025 onwards (cf. Keiner et al.
[10]
). In 2030, many
regions in the sunbelt increase their battery capacity together
with solar PV capacity. Only a few regions in central Eurasia
and Northeast Asia reduce the stationary battery capacity. For
most regions, however, it seems that the reduced yield and con-
sequently in general higher solar PV capacities (cf. Figure 5)
rather leads to slightly increased battery capacities as well. In
2040, the situation is not very different from 2030 with a mix
of increased stationary battery capacity mostly in South
America, SSA, Eurasia, and Southeast Asia, while in North
America and Northeast Asia battery capacities are slightly
reduced. By 2050, the change in stationary battery capacity is
more homogeneous with a clear tendency to reduce battery
capacities in North America, Europe, and Eurasia, whereas bat-
tery capacities stay more or less the same compared to no yield
reduction or are slightly increased in the rest of the world. The
relative change in stationary battery capacity is not as strong as
for solar PV capacity in 2020 and 2050; however, in the time steps
in between, namely 2030 and 2040, the optimized battery capac-
ity reacts more strongly to the reduced yield. Numeric differences
in installed stationary battery capacity for all regions and years
can be found in the Supporting Information.
3.2.3. Annualized Total Cost of Energy
The objective of the optimization of the PV prosumer system is
the ATCE. Figure 7 shows the changes in ATCE for the residen-
tial PV prosumer system if yield losses are applied compared to
the optimal system.
Contrary to the results presented up to now, the change in
ATCE is rather homogeneous over the globe. In 2020, a few
regions show a small change in ATCE. The reason is that, in
2020, the main supply via a rooftop PV system is not yet that pro-
nounced, and for many regions a supply of electricity via the elec-
tricity grid is more beneficial. Nevertheless, many regions already
Figure 4. Anticipated yield expressed as FLH for residential rooftop PV systems (top left), commercial rooftop PV systems (top right), and industrial
rooftop PV systems (bottom).
www.advancedsciencenews.com www.solar-rrl.com
Sol. RRL 2024, 2400435 2400435 (7 of 12) © 2024 Wiley-VCH GmbH
2367198x, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/solr.202400435 by Dominik Keiner - Lappeenranta University Of Technology , Wiley Online Library on [24/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
have the PV system as part of a least-cost solution. With the
increasing importance of rooftop PV systems, the share of PV-
supplied electricity grows as self-consumption becomes more
favorable. Consequently, the ATCE raises in most regions by
2030 compared to an optimal solar PV yield. Only in Eurasia
is this change not yet that evident as this is the region with
the least strong development of rooftop PV system with a major
increase post 2040. In 2040, the ATCEs of most of the regions are
up to 20% higher than if the rooftop PV system would have an
optimal yield. The same is the case in 2050. In the northern parts
of North America, northern Europe, and Eurasia, the ATCEs
increase by up to 10% compared to the optimal yield.
Conversely, in basically all other regions of the world, the
ATCE increases by up to 21% and a global average of 16%.
Though solar PV capacities and stationary battery capacities react
very inhomogeneously to the lower yield situation, the cost almost
linearly increases with the lower yield. If the optimized system
solution includes less solar PV capacity, more electricity has to
be drawn from the grid. If the solar PV system is already beyond
grid parity, the lower availability of solar PV electricity is counter-
balanced with an increased system design. Therefore, in techno-
economic modeling of energy systems, it is most important to
make a dedicated modeling of solar PV prosumers with accord-
ingly adapted yield profiles. Numeric differences in ATCE for all
regions and years can be found in the Supporting Information.
4. Discussion
This study estimates the yield loss for residential rooftop solar
PV systems to 18% annually, 7% annually for commercial sys-
tems, and 4% for industrial systems. A comparison of these val-
ueswithotherresultsinliteratureisnotpossibleduetothe
lacking availability of comparable studies. This absence of liter-
ature rather comes as a surprise since rooftop solar PV and solar
PV prosumers are usually assigned a significant role in the
energy transition with its own broad field of research.
[61]
Distributed solar PV power plants will support the energy tran-
sition on many levels, from using distributed inverters for fre-
quency support and system recovery,
[62]
via having a major
share in electricity supply,
[5]
to a decreased aging of transform-
ers if electric vehicles are charged at home via solar PV plants
instead of the grid.
[63]
As shown in the literature review in
Section 1, only very few energy system models include a differ-
entiated solar PV prosumer sub-model or rooftop PV which, in
the light of the “multi-talent”distributed solar PV, seems to be a
major shortcoming in energy system research. The two leading
reports on global energy transition, the World Energy Outlook
of the IEA currently available in the 2023 version
[64]
and the
assessment report of the Intergovernmental Panel on
Climate Change, latest available for the 6th assessment cycle,
[65]
do not differentially mention the important role of distributed
Figure 5. Difference in installed solar PV capacity for residential PV prosumer systems if yield losses are applied compared to an assumption of optimal
yield profiles. A positive change indicates that that with losses applied, the installed solar PV capacity by the model is higher than without yield losses.
Shown are the results for 2020 (top left), 2030 (top right), 2040 (bottom left), and 2050 (bottom right). No change (0%) might not only indicate no change
in least-cost solar PV capacity but also that regions without solar PV capacity as least-cost option might not change their status of not having solar PVas
part of their least-cost solution.
www.advancedsciencenews.com www.solar-rrl.com
Sol. RRL 2024, 2400435 2400435 (8 of 12) © 2024 Wiley-VCH GmbH
2367198x, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/solr.202400435 by Dominik Keiner - Lappeenranta University Of Technology , Wiley Online Library on [24/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
solar PV, rooftop solar PV, or solar PV prosumers, or simply
include it generally in solar PV.
Based on the chosen assumptions, however, results for roof-
top solar PV in energy systems might differ significantly when
looking at extreme cases. While Bogdanov et al.
[5]
find a high
share of rooftop solar PV electricity generation of about
14.3 PWh of a total 104.3 PWh from solar PV in 2050 for a global
transition toward 100% renewable energy in all energy sectors,
the results of Gernaat et al.
[66]
suggest 2.8 PWh in a total of
5.7 PWh electricity from solar PV by 2050 for the shared
socio-economic pathway 2 (SSP2), which represents a “mid-
dle-of-the-road”world, following historical developments in
social, economic, and technological trends.
[67]
The SSP2 marker
scenario, however, already suggests a total of 72 PWh electricity
from solar PV by 2050. As mentioned in the introduction, the
global rooftop solar PV market has been booming for several
years. Gernaat et al.
[66]
may serve as a negative example of mis-
interpreting the future role of solar PV in general and rooftop
solar PV in particular, studies such as Bogdanov et al.
[5]
will profit
from the results in this study to refine their assumptions about
rooftop solar PV. A showcase study in terms of consideration of
rooftop PV may be Rahdan et al.
[25]
who did a dedicated optimi-
zation on distribution grid level including rooftop PV, electric
vehicles, etc. Future comparable studies will show if the yield
reduction for rooftop solar PV as identified in the present article
will hinder the prospects of rooftop solar PV and prosumers, or
the low-cost electricity option powered by the sun is more resil-
ient to such changes due to a comfortable price buffer.
Yield profiles for rooftop solar PV are hard to obtain for global
energy system modeling since, the yield of rooftop solar PV is
different for each PV system. Therefore, the method presented
in this study to obtain a general loss factor helps to identify roof-
top solar PV profiles if ground-mounted yield profiles are avail-
able. Large-scale ground-mounted solar PV power plant profiles
are easier to simulate as the yield calculation can be more sim-
plified. Global simulations that rely on general data from satel-
lites rather than specific on-site measurements depend on a
simple, easy to use, and equation-based yield model for power
plants, such as applied by Breyer and Schmidt,
[56]
in contrast
to detailed electric models of solar PV modules as for example
presented by Cuce et al.
[68]
These two studies shall act as exam-
ples for the wide variety of different modeling approaches avail-
able in literature. Fixed tilted utility-scale power plants may show
yield losses due to mutual shading in the morning and evening
hours. These losses, however, may not come as a disadvantage
for the presented approach in this study since shading from
housing structures and other obstacles usually affects the yield
in the mornings and evenings on rooftops as well. Larger rooftop
solar PV systems on commercial and industrial roofs are also
often mounted and approximate with growing size the yield
Figure 6. Difference in installed stationary battery capacity for residential PV prosumer systems if yield losses are applied compared to an assumption of
optimal yield profiles. A positive change indicates that with losses applied, the installed stationary battery capacity by the model is higher than without
yield losses. Shown are the results for 2020 (top left), 2030 (top right), 2040 (bottom left), and 2050 (bottom right). No change (0%) might not only
indicate no change in least-cost stationary battery capacity but also that regions without battery capacity as least-cost option might not change their status
of not having a battery as part of their least-cost solution.
www.advancedsciencenews.com www.solar-rrl.com
Sol. RRL 2024, 2400435 2400435 (9 of 12) © 2024 Wiley-VCH GmbH
2367198x, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/solr.202400435 by Dominik Keiner - Lappeenranta University Of Technology , Wiley Online Library on [24/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
and loss characteristics of ground-mounted power plants. This
claim also has to be understood as an estimation, as actual shad-
ing from obstacles is again case- and roof-specific.
[37]
Depending on the definition, curtailment may be counted for
as yield loss, too. High solar PV concentration in low voltage dis-
tribution grids may cause issues
[69]
and regulatory measures may
be in order to overcome such challenges.
[70]
However, especially
in densely populated urban areas the electricity demand may
exceed the production from local solar PV.
[71]
Additional deploy-
ment of storage such as batteries and flexibility options, in par-
ticular smart charging of electric vehicles and vehicle-to-grid
[72]
may reduce the risk for curtailment. Seasonal storage for rooftop
solar PV prosumers may not be economically beneficial
[10]
and
better be solved on a community-level.
[73]
This study has been carried out with currently available data
on the loss effects of different system levels, which are rather
limited. However, in case the availability of respective data
improves, this study may have to be revised and adopted to
new findings. One example is the use of microinverters for roof-
top solar PV modules. Currently, no substantial market penetra-
tion can be identified from literature, nor for an outlook on the
foreseeable future. However, if the techno-economics of the
respective technology improves and becomes economically more
attractive compared to conventional inverters, then the loss
effects on system levels will have to be adapted. Long-term yield
disparities have to be checked including degradation effects of
rooftop solar PV systems versus ground-mounted systems.
5. Conclusions
Understanding the yield prospects of rooftop solar PV power
plants is necessary to allow for the rightful accounting of yield
losses in energy system transition modeling with a differentiated
consideration of solar PV prosumers. Usually, dedicated rooftop
solar PV profiles are not available. This article presented the lit-
erature and method to estimate the yield losses of residential,
commercial, and industrial rooftop solar PV systems in compar-
ison to optimal yield represented by utility-scale ground-mounted
solar PV power plants on three different levels: environmental
level, cell and module level, and system level. The assessment
indicated about 18% annual loss for average real-world residen-
tial systems compared to optimally simulated systems. This
impact was estimated to be 7% for commercial-sized systems
and to 4% for big-scale industrial rooftop PV systems.
The viability of residential solar PV prosumer systems with the
reduced, that is, anticipated, yield has been assessed by compar-
ing the total annualized cost of energy, including electricity
demand, heat demand, and one-battery electric vehicle, in an
on-grid case scenario for 145 regions globally. The results
Figure 7. Difference in ATCE for residential PV prosumer systems if yield losses are applied compared to an assumption of optimal yield profiles. A
positive change indicates that with losses applied, the ATCE of the system are higher than without yield losses. Shown are the results for 2020 (top left),
2030 (top right), 2040 (bottom left), and 2050 (bottom right). No change (0%) might not only indicate no change in ATCE with the PV system but also that
regions that did not install a PV system and relied fully on-grid supply do not change their 100% grid reliance.
www.advancedsciencenews.com www.solar-rrl.com
Sol. RRL 2024, 2400435 2400435 (10 of 12) © 2024 Wiley-VCH GmbH
2367198x, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/solr.202400435 by Dominik Keiner - Lappeenranta University Of Technology , Wiley Online Library on [24/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
indicate a nonuniform change in installed solar PV capacity and
stationary battery capacity, due to individual parameter combina-
tions and grid electricity cost for each region. For most regions,
the missing yield of the solar photovoltaic system is counter-bal-
anced with higher solar PV capacity by 2050 of up to 20%. The
change in installed stationary battery capacity is similar; however,
the change is not as significant as for the solar PV capacity. The
annualized total cost of energy of the prosumer systems
increases in all regions by 2050, as the cost-optimized average
households are all equipped with a solar PV system by mid-cen-
tury. On global average, an 18% yield reduction leads to about
16% higher system cost by 2050.
This study provided a novel view of the loss effect of rooftop
solar PV systems compared to ground-mounted solar PV sys-
tems. The results indicate the significance of the necessity to
include differentiated rooftop solar PV options in energy transi-
tion models, or solar PV prosumer sub-models in the best case,
with a respective consideration of the yield losses. The results
aim to improve the quality and detail of research of future energy
system transition studies with a better consideration of one of the
main pillars of the energy transition, which are distributed
renewable energy sources such as rooftop solar PV.
Supporting Information
Supporting Information is available from the Wiley Online Library or from
the author.
Acknowledgements
The authors gratefully acknowledge the public financing of European
Union’s Horizon 2020 research and innovation programmed under grant
agreement No. 953016 (SERENDI-PV), which partly funded this research.
D.K. would like to thank the Jenny and Antti Wihuri Foundation for the
valuable grant. The authors would like to thank Gabriel Lopez for
proofreading.
Conflict of Interest
The authors declare no conflict of interest.
Author Contributions
Dominik Keiner: Conceptualization (equal); Data curation (lead);
Methodology (equal); Resources (lead); Validation (equal); Visualization
(lead); Writing—original draft (lead). Dmitrii Bogdanov: Validation (equal);
Writing—review and editing (supporting). Stefan Krauter: Validation (sup-
porting). Christian Breyer: Conceptualization (equal); Methodology (equal);
Supervision (lead); Validation (equal); Writing—review and editing (lead).
Data Availability Statement
Data sharing is not applicable to this article as no new data were created or
analyzed in this study.
Keywords
battery, distributed, photovoltaics, prosumer, rooftop
Received: June 13, 2024
Revised: October 8, 2024
Published online:
[1] International Energy Agency (IEA), World Energy Outlook 2020, Paris
2020, https://www.iea.org/reports/world-energy-outlook-2020.
[2] International Renewable Energy Agency (IRENA), Smart Electrification
With Renewables: Driving the Transformation of Energy Services, Abu
Dhabi 2022, https://www.irena.org/-/media/Files/IRENA/Agency/
Publication/2022/Feb/IRENA_Smart-Electrification_Renewables_2022.
pdf?rev=f4add985f6284a7cb811977c65b281b1.
[3] S. Joshi, S. Mittal, P. Holloway, P. R. Shukla, B. Ó. Gallachóir, J. Glynn,
Nat. Commun. 2021,12, 5738.
[4] International Energy Agency Photovoltaic Systems Programme (IEA-
PVPS), Trends in Photovoltaic Applications 2023, Paris 2023, https://
iea-pvps.org/trends_reports/trends-2023/.
[5] D. Bogdanov, M. Ram, A. Aghahosseini, A. Gulagi, A. S. Oyewo,
M. Child, U. Caldera, K. Sadovskaia, J. Farfan, L. S. N. S. Barbosa,
M. Fasihi, S. Khalili, T. Traber, C. Breyer, Energy 2021,227, 120467.
[6] M. Z. Jacobson, M. A. Delucchi, M. A. Cameron, B. V. Mathiesen,
Renewable Energy 2018,123, 236.
[7] International Energy Agency Photovoltaic Systems Programme (IEA-
PVPS), Snapshot of Global PV Markets 2023, Paris 2023, https://iea-
pvps.org/snapshot-reports/snapshot-2023/.
[8] International Energy Agency - Renewable Energy Technology
Deployment (IEA-RETD), Residential Prosumers - Drivers And Policy
Options (RE-Prosumers), Paris 2014, https://www.researchgate.net/
publication/323663159_Residential_prosumers_-_Drivers_and_policy_
options_RE-PROSUMERS.
[9] D. Keiner, M. Ram, L. S. N. S. Barbosa, D. Bogdanov, C. Breyer, Sol.
Energy 2019,185, 406.
[10] D. Keiner, C. Thoma, D. Bogdanov, C. Breyer, Appl. Energy 2023,340,
121009.
[11] K. Kotilainen, in Affordable and Clean Energy (Eds: W. Leal Filho,
A. M. Azul, L. Brandli, P. G. Özuyar, T. Wall), Springer, Cham 2020,
pp. 1–14.
[12] J. M. Wittmayer, I. Campos, F. Avelino, D. Brown, B. Dora ˇci´c,
M. Fraaije, S. G¨
ahrs, A. Hinsch, S. Assalini, T. Becker, E. Marín-
González, L. Holstenkamp, R. , N. Dui´c, S. Oxenaar, T. Pukšec,
Energy Res. Soc. Sci. 2022,86, 102425.
[13] M. Child, D. Bogdanov, A. Aghahosseini, C. Breyer, Futures 2020,124,
102644.
[14] A. B. Awan, M. Alghassab, M. Zubair, A. R. Bhatti, M. Uzair, G. Abbas,
Energies 2020,13, 3639.
[15] N. Bansal, S. P. Jaiswal, G. Singh, Sustainable Energy Technol. Assess.
2021,47, 101526.
[16] S. Khalili, C. Breyer, IEEE Access 2022,10, 125792.
[17] H. Lund, J. Z. Thellufsen, P. A. Østergaard, P. Sorknæs, I. R. Skov,
B. V. Mathiesen, Smart Energy 2021,1, 100007.
[18] R. Loulou, G. Goldstein, A. Kanudia, A. Lettila, U. Remme,
Documentation for the TIMES Model,2016. https://iea-etsap.org/
docs/Documentation_for_the_TIMES_Model-Part-I_July-2016.pdf.
[19] EuropeanCommission - Joint ResearchCentre - Institute for Energy and
Transport,The JRC-EU-TIMES Model: Assessing the Long TermRole of the
SET Plan Energy Technologies, Publications Office, Luxembourg 2013.
https://data.europa.eu/doi/10.2790/97596 (accessed: May, 2024).
[20] F. Gardumi, A. Shivakumar, R. Morrison, C. Taliotis, O. Broad,
A. Beltramo, V. Sridharan, M. Howells, J. Hörsch, T. Niet, Y. Almulla,
E. Ramos, T. Burandt, G. P. Balderrama, G. N. Pinto De Moura,
E. Zepeda, T. Alfstad, Energy Strategy Rev. 2018,20,209.
[21] T. Niet, A. Shivakumar, F. Gardumi, W. Usher, E. Williams,
M. Howells, Energy Strategy Rev. 2021,35, 100650.
www.advancedsciencenews.com www.solar-rrl.com
Sol. RRL 2024, 2400435 2400435 (11 of 12) © 2024 Wiley-VCH GmbH
2367198x, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/solr.202400435 by Dominik Keiner - Lappeenranta University Of Technology , Wiley Online Library on [24/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
[22] K. Löffler, K. Hainsch, T. Burandt, P.-Y. Oei, C. Kemfert, C. von
Hirschhausen, Energies 2017,10, 1468.
[23] M. Z. Jacobson, M. A. Delucchi, M. A. Cameron, S. J. Coughlin,
C. A. Hay, I. P. Manogaran, Y. Shu, A.-K. Von Krauland, One Earth
2019,1, 449.
[24] F. Neumann, E. Zeyen, M. Victoria, T. Brown, Joule 2023,7, 1793.
[25] P. Rahdan, E. Zeyen, C. Gallego-Castillo, M. Victoria, Appl. Energy
2024,360, 122721.
[26] W. J. Cole, D. Greer, P. Denholm, A. W. Frazier, S. Machen, T. Mai,
N. Vincent, S. F. Baldwin, Joule 2021,5, 1732.
[27] K. Löffler, T. Burandt, K. Hainsch, P.-Y. Oei, F. Seehaus, F. Wejda,
Energy 2022,239, 121901.
[28] T. Weber, A. Blakers, D. Firnando Silalahi, K. Catchpole, A. Nadolny,
Energy Convers. Manage. 2024,308, 118354.
[29] J. Gea-Bermúdez, I. G. Jensen, M. Münster, M. Koivisto,
J. G. Kirkerud, Y. Chen, H. Ravn, Appl. Energy 2021,289, 116685.
[30] G. Luderer, S. Madeddu, L. Merfort, F. Ueckerdt, M. Pehl, R. Pietzcker,
M. Rottoli, F. Schreyer, N. Bauer, L. Baumstark, C. Bertram,
A. Dirnaichner, F. Humpenöder, A. Levesque, A. Popp, R. Rodrigues,
J. Strefler, E. Kriegler, Nat. Energy 2022,7,32.
[31] H. C. Gils, H. Gardian, J. Schmugge, Renewable Energy 2021,180, 140.
[32] S. Teske, editor, in Achieving the Paris Climate Agreement Goals,
Springer International Publishing, Cham 2019.
[33] International Energy Agency Photovoltaic Systems Programme (IEA-
PVPS), Uncertainties in PV System Yield Predictions and Assessments,
Paris 2018, https://iea-pvps.org/key-topics/uncertainties-in-pv-system-
yield-predictions-and-assessments/.
[34] International Energy Agency Photovoltaic Systems Programme (IEA-
PVPS), Uncertainty in Yield Assessments and PV LCOE, Paris 2020,
https://iea-pvps.org/key-topics/uncertainties-yield-assessments/.
[35] International Energy Agency Photovoltaic Systems Programme (IEA-
PVPS), Performance of New Photovoltaic System Designs, Paris 2021,
https://iea-pvps.org/key-topics/performance-of-new-photovoltaic-
system-designs/.
[36] S. Killinger, D. Lingfors, Y.-M. Saint-Drenan, P. Moraitis, W. van Sark,
J. Taylor, N. A. Engerer, J. M. Bright, Sol. Energy 2018,173, 1087.
[37] G. Trzmiel, D. Głuchy, D. Kurz, Renewable Energy 2020,153, 480.
[38] U. Berardi, J. Graham, Sustainable Cities Soc. 2020,62, 102402.
[39] G. Y. Yun, M. McEvoy, K. Steemers, Sol. Energy 2007,81, 383.
[40] M. J. Ritzen, Z. A. E. P. Vroon, R. Rovers, A. Lupíšek, C. P. W. Geurts,
Sol. Energy 2017,155, 304.
[41] G. Osma-Pinto, G. Ordó˜nez-Plata, Sol. Energy 2019,185, 112.
[42] R. Bründlinger, B. Bletterie, M. Milde, H. Oldenkamp, in 21st Eur.
Photovoltaic Sol. Energy Conf., Dresden 2006, pp. 2157–2160.
[43] S. Kouro, J. I. Leon, D. Vinnikov, L. G. Franquelo, IEEE Indus. Electron.
Mag. 2015,9, 47.
[44] D. Dong, M. S. Agamy, M. Harfman-Todorovic, X. Liu, L. Garces,
R. Zhou, P. Cioffi,IEEE Trans. Indus. Appl. 2018,54, 469.
[45] A. Mohd, in 26th Eur. Photovoltaic Sol. Energy Conf. Exhib., WIP,
Hamburg 2011, pp. 3739–3744.
[46] M. Gostein, B. Littmann, J. R. Caron, L. Dunn, in 2013 IEEE 39th
Photovoltaic Special. Conf. (PVSC), IEEE, Tampa 2013, pp. 3004–
3009.
[47] T. Townsend, P. Hutchinson, in Proc. Sol. Conf., American Solar
Energy Society, Madison 2000, pp. 417–422.
[48] R. J. Mustafa, M. R. Gomaa, M. Al-Dhaifallah, H. Rezk, Sustainability
2020,12, 608.
[49] A. M. Nobre, S. Karthik, H. Liu, D. Yang, F. R. Martins, E. B. Pereira,
R. Rüther, T. Reindl, I. M. Peters, Renewable Energy 2016,89, 389.
[50] K. Hasan, S. B. Yousuf, M. S. H. K. Tushar, B. K. Das, P. Das, Energy
Sci. Eng. 2022,10, 656.
[51] International Energy Agency Photovoltaic Systems Programme
(IEA-PVPS), Assessment of Performance Loss Rate of PV Power
Systems, Paris 2021, https://iea-pvps.org/key-topics/assessment-of-
performance-loss-rate-of-pv-power-systems/.
[52] D. C. Jordan, S. R. Kurtz, K. VanSant, J. Newmiller, Prog. Photovoltaics
2016,24, 978.
[53] S. Herceg, I. Kaaya, J. Ascencio-Vásquez, M. Fischer, K.-A. Weiß,
L. Schebek, Sustainability 2022,14, 5843.
[54] I. Kaaya, J. Ascencio-Vásquez, K.-A. Weiss, M. Topiˇc, Sol. Energy 2021,
218, 354.
[55] D. Bogdanov, C. Breyer, Energy Convers. Manage. 2016,112, 176.
[56] C. Breyer, J. Schmid, in 25th Eur. Photovoltaic Sol. Energy Conf. Exhib./
5th World Conf. Photovoltaic Energy Convers., WIP, Valencia 2010, pp.
4715–4721.
[57] P. Stackhouse, C. Whitlock, Surface Meteorology and Solar Energy
(SSE) Release 6.0, NASA SSE 6.0, https://searchworks.stanford.
edu/catalog?q=%2522NASA%2bLangley%2bAtmospheric%2bScie
nces%2bData%2bCenter%2522%26search_field=search_author,
Langley 2008.
[58] P. Stackhouse, C. Whitlock, Surface Meteorology and Solar Energy
(SSE) Release 6.0 Methodology, NASA SSE 6.0, https://power.larc.
nasa.gov/docs/methodology/, Langley 2009.
[59] D. Stetter, Dissertation, Faculty of Energy-, Process- and Bio-engineer-
ing, University of Stuttgart 2014.
[60] S. Afanasyeva, D. Bogdanov, C. Breyer, Sol. Energy 2018,173, 173.
[61] Y. Shen, L. Ji, Y. Xie, G. Huang, X. Li, L. Huang, Energy Build. 2021,
251, 111333.
[62] H. Dehghanitafti, G. Konstantinou, J. Fletcher, L. Callegaro,
G. Farivar, J. Pou, IEEE Trans. Power Electron. 2022,37, 4742.
[63] M. K. Gray, W. G. Morsi, Electric Power Syst. Res. 2017,143, 563.
[64] International Energy Agency (IEA), World Energy Outlook 2023, Paris
2023, https://origin.iea.org/reports/world-energy-outlook-2023.
[65] Intergovernmental Panel on Climate Change (IPCC), Climate Change
2023: Synthesis Report. Contribution of Working Groups I, II and III to
the Sixth Assessment Report of the Intergovernmental Panel on Climate
Change, Geneva 2023.
[66] D. E. H. J. Gernaat, H.-S. de Boer, L. C. Dammeier, D. P. van Vuuren,
Appl. Energy 2020,279, 115705.
[67] O. Fricko, P. Havlik, J. Rogelj, Z. Klimont, M. Gusti, N. Johnson,
P. Kolp, M. Strubegger, H. Valin, M. Amann, T. Ermolieva,
N. Forsell, M. Herrero, C. Heyes, G. Kindermann, V. Krey,
D. L. McCollum, M. Obersteiner, S. Pachauri, S. Rao, E. Schmid,
W. Schoepp, K. Riahi, Glob. Environ. Change 2017,42, 251.
[68] E. Cuce, P. M. Cuce, I. H. Karakas, T. Bali, Energy Convers. Manage.
2017,146, 205.
[69] T. Aziz, N. Ketjoy, IEEE Access 2017,5, 16784.
[70] G. Tévar, A. Gómez-Expósito, A. Arcos-Vargas, M. Rodríguez-
Monta˜nés, Int. J. Electr. Power Energy Syst. 2019,106, 68.
[71] M. Ram, A. Gulagi, A. Aghahosseini, D. Bogdanov, C. Breyer,
Renewable Energy 2022,195, 578.
[72] D. Bogdanov, C. Breyer, Energy 2024,301, 131635.
[73] P. Grunow, Energies 2022,15, 2820.
[74] I. H. Rowlands, B. P. Kemery, I. Beausoleil-Morrison, Energy Policy
2011,39, 1397.
[75] H. Byrd, The Solar Potential of Auckland, The University of Auckland,
Auckland 2012.
[76] Y. S. Khoo, A. Nobre, R. Malhotra, D. Yang, R. Ruther, T. Reindl,
A. G. Aberle, IEEE J. Photovoltaics 2014,4, 647.
[77] M. Hartner, A. Ortner, A. Hiesl, R. Haas, Appl. Energy 2015,160, 94.
[78] B. Matthiss, D. Stellbogen, M. Ebersp¨
acher, J. Binder, in 31st
Eur. Photovoltaic Sol. Energy Conf. Exhib., WIP, Hamburg 2015,
pp. 2311–2314.
[79] S. Kichou, N. Skandalos, P. Wolf, J. Phys.: Conf. Ser. 2019,1343,012093.
[80] M. Oh, H.-D. Park, Energies 2019,12, 3262.
[81] C. Yu, Y. Khoo, J. Chai, S. Han, J. Yao, Sustainability 2019,11, 2016.
www.advancedsciencenews.com www.solar-rrl.com
Sol. RRL 2024, 2400435 2400435 (12 of 12) © 2024 Wiley-VCH GmbH
2367198x, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/solr.202400435 by Dominik Keiner - Lappeenranta University Of Technology , Wiley Online Library on [24/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
