Performance evaluation of Phase Change Materials for cooling Photovoltaics and enhancing

efficiency in the Global South

(leave this line blank)
A1, B2, C2
1
Institution 1, City 1, Country 1
2
Institution 2, City 2, Country 2
(The names and affiliations SHOULD NOT be included in the draft submitted for review)
(leave blank up to line 10 – remove line numbering from final version)

Abstract thus preventing poor performance in hot

climatic locations.
Despite the carbon reduction potential of Photovoltaics
(PV) for buildings energy use, its adoption in the • PCM integration resulted in a lower average
residential sector is limited. For effective building annual efficiency for the remaining majority
integration, it is crucial that PV panels operate close to the locations.
standard conditions (~25°C). Overheating of PV in hot Practical Implications
climates leads to low PV efficiencies, which can Besides identification of the locations and weather
adversely affect the technology adoption. In this study, conditions where integrating PCM into PV may prove
the effectiveness of Phase Change Materials (PCM) in beneficial, the automated co-simulation approach
reducing panel surface temperatures of Concentrated presented in this study can be utilised to study other novel
Photovoltaic (CPV) in the Global South (GS) is explored. PV systems for which experimental data may be available
Utilising lab-based experimental data from a CPV-PCM but EnergyPlus does not have the capability to model.
system, and a co-simulation approach using the software While 256 co-simulations with different IDFs
BCVTB, MATLAB (v2018b) and EnergyPlus (v8.8), the (EnergyPlus Input Data Files), weather files and
panel surface temperature and CPV efficiency were experimental data were launched programmatically, a
simulated. A total of 128 locations across 17 countries in delay of less than 60s between successive simulations
the GS were identified for simulation that satisfied the resulted in failed BCVTB runs. Thus, the method used in
following conditions: (i) EnergyPlus weather files (.epw) this study is time intensive.
availability, (ii) yearly averaged diurnal swing greater
than 10°C, and (iii) peak summer temperature in excess Introduction
of 25°C. From the results, we identified 36 locations in Integration of Photovoltaics (PVs) into buildings can
the GS where PCM integrated PV resulted in a 2% - 7% contribute to large reductions in carbon emissions, but
relatively improved annual average efficiency over the with a total installed capacity of ~133 GW worldwide
non-PCM baseline, while in the remaining 92 locations (2020 IEA), and a potential of 707.5 GW, solar PV
PCMs either did not make a notable improvement or adoption in the residential sector currently only stands at
reduced the efficiency. 18%. It is a matter of great concern as countries across the
Key Innovations globe work together to contribute to the United Nation
Sustainable Development Goals 7 (Affordable and Clean
• 885 Global South locations were analysed, out Energy), Goal 9 (Industry, Innovation, and
of which 128 locations had at least a 10°C Infrastructure), and Goal 13 (Climate Action).
monthly average diurnal swing, and sufficiently
For effective building integration, it is crucial that the PV
hot summers to warrant PCM-PV integration.
panels operate close to the standard temperature
• PCM-CPV performance was simulated for the conditions (~25°C), Overheating in hot climates leads to
selected 128 Global South locations. A co- reduction in efficiency and panel degradation, leading to
simulation approach using EnergyPlus was used shortened lifetime of the PV/CPV panels and impeding its
to facilitate implementation of buildings energy uptake.
modelling.
In this paper, we aim to clearly distinguish the geographic
• PCM integration resulted in improved locations across the whole of Global South where addition
performance in 36 locations, because of lower of Phase Change Materials (PCM) would prove
maximum and minimum daily efficiencies – advantageous from the ones where it would be
detrimental to the PV system efficiency. For the purpose
of this communication, the term Global South refers
broadly to the regions of Latin America, Asia, Africa, and
Oceania; “Third World” and “Periphery,” that denote
regions outside Europe and North America, mostly
(though not all) low-income and often politically or
culturally marginalized (Dados & Connell, 2012).
In this study, the effectiveness of PCMs as a means to effective for the free-standing PVs with a daily average
reduce the operating temperature of Concentrated efficiency increase by 7.1% (Nada & El-Nagar, 2018).
Photovoltaic (CPV) panels in the Global South has been The efficiency of a vertical building-added PV-PCM
explored. Utilising lab-based experimental results from a system showed an average increase of 1.5% in the power
CPV-PCM system, and a co-simulation approach using output, while the peak increased by a maximum of 3%
the software BCVTB, MATLAB and EnergyPlus, the experimentally with the annual weather conditions taken
panel surface temperature and CPV efficiency were into account (Park et al., 2014).
simulated.
In a separate application, it was revealed that adding flat
The results and conclusions from this study on one hand plate solar collectors, absorption chiller with fan coil, and
will help increase the understanding about the PCM to the walls of a building could reduce the energy
effectiveness of CPV-PCM systems as a whole and on the consumption by 450 kWh (Amirahmad et al., 2021).
other hand assist decision makers with quantitative data
The enviro-economic analysis on the impacts of various
for wide geographies to make appropriate investment
PV-cooling techniques on CO2 emissions revealed that
decisions.
passive technologies such as nano-PCM, hybrid PCM,
and hybrid PCM-water, respectively proved more
Literature review effective than the use of individual techniques
(Ghadikolaei, 2021).
PV performance in hot climates
PV-PCM performance modelling and simulation
The electrical efficiency of a PV panel has a linear
correlation with the operating temperature; the PV A building integrated PV-PCM system installed on the
material being a primary influence for the temperature building façade of an office was theoretically simulated
sensitivity. For instance, the most commonly used and field-tested for the winter season in Lisbon, Portugal
crystalline silicon-based PV panels have a permissible (Aelenei et al., 2014). The test prototype (0.73 m × 1.75
degradation rate of 0.8 % per year with warranties for up m) consisted of a polycrystalline PV-panel and a gypsum
to 25 years. The rates, however, demonstrate an annual wallboard incorporating PCM Alba® balance type. A 1-
median value of 0.5% as per an international review of the D dynamic simulation program and a control-volume
available field test results (Gaglia et al., 2017). based finite-difference scheme was used through
MATLAB/SIMULINK® with SIMSCAPE® library on a
The key methods for PV cooling could be broadly
staggered grid. It was found that the thermal efficiency of
classified into active and passive means depending on
the PV-PCM reached as high as 10% and an overall
whether electrical power is used or not (through a pump,
(thermal plus electrical) efficiency of about 20% with
blower, or a fan). While forced air and water cooling fall
PCM (Aelenei et al., 2014). Unfortunately, a comparison
under the category of active cooling, passive cooling
with the baseline (no PCM) case was not provided in the
includes natural/free convection of air and water, heat
study.
pipes, and use of PCM.
Integration of PCM with PV in glazed buildings was
PV-PCM performance experimentation
numerically studied using a 1-dimensional transient
Several authors from literature advocate PCM as an model using MATLAB/TRNSYS and validated with the
effective passive method for PV cooling for temperature experimental results (Elarga et al., 2017). The
reduction, increase in electrical output thereby positive temperature profile of PCM on a buildings window, and
environmental implications through reducing the use of with a PV layer was analysed. The overall numerical
energy-intensive silicon material. The effectiveness is model was implemented to a double skin façade building
determined by thermo-physical properties of the material with a PV-PCM layer in a ventilated cavity in Venice and
as well as befitting external conditions such as irradiance, Helsinki. It was revealed that a PV-PCM layer was
ambient temperature affecting the convection effect sufficient to reduce the demand for cooling by 60% in
within the melted PCM, the wind velocity, and the angle Venice, and the demand for heating in Helsinki by 36%
of inclination of the PV panel as tilt angles and wind (Elarga et al., 2017). The authors noted that substantial
velocities are associated with lowering the operating amounts of heating/cooling energy demands could be met
temperatures (Kant et al., 2016). through building-integration of PV-PCM in Nordic
In a study (Nada & El-Nagar, 2018), four different PV countries, depending on the thermo-physical material
modules, namely, (a) a free-standing, (b) a building- properties.
integrated, (c) a PCM-integrated, and (d) a nano-enhanced A simulation study for PV-PCM system with the climatic
PCM-integrated system were compared experimentally conditions of Incheon, South Korea, demonstrated that the
for temperatures distributions and output electrical best outputs in terms of generated power were observed
parameters at a site in Egypt. The results showed that for October with high irradiance and moderate ambient
using PCM and nano-PCM increased the average daily temperature; later being an important criterion for PCM
power in free-standing PVs by 4.1% and 5.6% effectiveness (Park et al., 2014). Irrespective of the
respectively. The peak power increase was 6.7% and direction of installation, the PCM melting temperature of
8.5% respectively. The building integrated PV showed 298 K (~25°C) was found optimum.
high temperature rise (from 50 °C to 75 °C) compared to
the free-standing and that the PCM addition was only
In another simulation, the dynamic thermal performance necessitate PV surface temperature management (ii)
model for a ventilated PV-PCM façade showed a peak experiment details for with and without PCM integrated
shift time of over five hours, with a temperature reduction into PVs (iii) co-simulation workflow. These are
of over 20 °C (Curpek & Hraska, 2017). Design Builder described as follows.
software was used to study the climatic conditions of Geographical locations and weather data selection
Bratislava (Slovakia) in this case.
The simulation methodology in this study employs
Through the above literature review, it was identified that EnergyPlus as the core simulation engine, for which
so far only a few studies have targeted the simulation of weather data is in the form of hourly typical reference
PV systems with PCM and within these, almost all of years, in EPW(Energy Plus Weather) format. In this
them focussed on one, two or a few global locations to test study, all freely available epw files (3034 in total) were
the effectiveness of PV-PCM systems. Therefore, we downloaded in bulk from (EnergyPlus, 2022). From the
attempt to cover a majority of regions from the Global
global dataset, there were 885 Global South location
South to address this knowledge gap.
weather files which were then further analysed for PV-
While the studies described in this section show that PCM suitability.
PCM-PV modelling has been carried out, the following
research gaps were observed For PV-PCM systems to work effectively, it is imperative
that the PCM undergo congealing or solidification during
• Generally, studies consider limited geographical the non-sunny hours (Lazaro et al., 2009). For varying
locations, with very few studies done for Global solar irradiance, ambient temperatures and other climatic
South locations.
parameters, the selection of an appropriate PCM material
• Most studies employ simulation approaches that is indeed of paramount importance. Together with the fact
may not be easily integrated into Buildings
that PV cooling is only required for locations where the
Energy Modelling (BEM). Few studies use
elevated PV surface temperture may lead to lower
widely accepted BEM tools, such as
efficiency, the 885 GS locations were further filtered into
DesignBuilder, but none analysed PCM
integration into PV panels. 128 locations across 17 countries using the following two
criteria:
To address these gaps, this paper used an approach well
suited for BEM and considers 885 locations in the Global • A minimum of 10°C monthly average diurnal
South for assessing the PV-PCM potential. Moreover, the swing.
experimental data used for this study is for a CPV-PCM • A minimum peak summer temperature of 25°C.
system fully tested in the laboratory with the It should be noted here that with such a climate selection,
concentrating element added to the study, which adds to it is assumed that the PCMs completely charge and
the novelty of this research. discharge over the passage of a day. This assumes careful
Methods selection of the PCM melting point with reference to the
This study employs a co-simulation approach using ambient temperature (Lazaro et al., 2009).
BCVTB (Wetter et al., 2008) to couple MATLAB 2018b Experimental data
(Mathworks, 2021) and EnergyPlus v8.8 (EnergyPlus, The indoor experimental work was carried out under the
2012). Such an approch allows for modelling that is not AAA Wacom Super Solar Simulator® (WXS-300S-50
readily available in EnergyPlus (Khattak et al., 2020). As AM1.5G) in the UK (Sharma et al., 2016). A CPV-PCM
a PCM integrated PV model was not available in system was designed and fabricated in the laboratory. The
EnergyPlus, this co-simulation approach was considered key components for the experimental set up were:
appropriate for this study. Laboratory experimental data
i. A PV module with crystalline silicon solar cells
is utilized within the simulation timestep while making
(dimensions 116×6 mm2) soldered in series using
use of the full simulation capability of EnergyPlus. While
thin tin-plated copper strips, and assembled on a 0.6
building performance is not analysed in this study, a
mm thick Aluminium back plate.
typical mid-rise appartment reference building (DOE,
2020) was used where the PV were positioned ii. Optical elements or concentrators: linear asymmetric
horizontally on the roof. The co-simulation approach compound parabolic concentrator (LACPC) with
allowed us to understand how the efficiency of typical acceptance half-angles as 0° and 55° providing a low
solar PV panels may be affected if PCM are integrated geometrical concentration of ~2.7.
into them, while taking into account weather conditions iii. PCM containment made using 13 mm thick Perspex
for typical reference years across selected Global South sheet walls with the inner dimensions of 144×134×38
locations. Moreover, the approach is suitable for mm3.
Buildings Energy Modleling and can be utlised for iv. PCM Rubitherm RT-42® was used; latent heat
another other experimental data set. The methodology capacity of 174 kJ/kg, and melting temperature range
employed is comprised of (i) selecting appropriate between 38-43°C.
weather location where PCM can be fully charged and
v. The electrical connections and peripheral devices for
discharged and where the summers may be hot enough to
measurements.
The CPV-PCM system was characterised performed Simulation workflow:
using a highly collimated illumination from the solar In this study, 128 locations had to be considered for two
simulator at 0° from horizontal. cases, with and without PCMs – within a co-simulation
The processes involved in manufacturing and assembling setup. While the co-simulation was managed by BCVTB,
the system and further details on the thermo-physical it was launched from MATLAB sequentially for the 128
properties of the components can be found from published location, updating the BCVTB xml file with the
literature (Sharma et al., 2016). The results showed an appropriate weather file before each simulation. This way,
efficiency improvement in the CPV panel with the use of an autmated co-simiulation workflow was created, which
PCM at all the tested irradiance (500, 750, 1000 and 1200 in future studies could be easily expanded to consider
Wm−2) levels (Figure 1). The relative electrical efficiency different weather locations, building types and
improvements were 1.15%, 7.7%, 4.20% and at 6.80% at experimental data.
500, 750, 1000 and 1200 Wm−2 respectively using PCM For each timestep (10min), EnergyPlus calculates the heat
for passive cooling. A reduction in average temperature transfer for the interaction between the PV surface and the
was noticed (6.2°C) in the centre of the CPV panel weather conditions. The simulated PV surface
integrated with PCM compared to the naturally ventilated temperature dictates the PV efficiency which is read in
system. MATLAB based on the linear relationships between PV
surface temparature and efficiency from labwork (Figure
2). Thus, with this scheme, the heat transfer model of
EnergyPlus is coupled with the lab data that qauntifies the
effect of PCM on PV efficiency, whilst taking into
account weather conditions for the typical reference year.
It sould be noted from Figure 2 that PV efficiency only
improves when the PV surface temperature is above
33.3°C. This is important because for any timepoint
where the surface temperaure is below this, PCM will
actually reduce PV efficiency. Thus, simulating for the
Figure 1: Temperature at the centre of the module fluctuating weather conditions over a year would help
without and with PCM at various irradiance levels. The answer if PCMs could improve PV efficiency.
average temperature reduction for the complete dataset is A limitation of this study is that indoor laboratory data
6.2 °C was only available for a limited temperature range (15 <
T < 55.2). An analysis of the 128 Global South weather
files revealed that less than 1% of the total hours lie above
55.2°C but a substantial amount (40%) of data is at a PV
surface temperature below 15°C. Therefore, the results of
the colder climates within the global south for this study
may underestimate the reduction in PV efficiency from
PCM integration for such climates.
Results and Discussion:
Based on the hourly simulations for the typical year,
Figure 3 shows the boxplots for each location for the
baseline no-PCM case. For this particular CPV used in the
laboratory, the average PV efficiency across the 128
locations ranged between 9-11% approximately. This data
was based on efficiency for daylight hours only, as this is
when the PV will actually operate. The actual locations
corresponding to their serial numbers in Figure 3 and the
remaining paper is provided in the Appendix.
Figure 2: Effect of PCM on PV efficiency based on
indoor lab experiments. Note that at a surface
temperature greater than 33.3C, usage of PCM results in
higher efficiency.
The PV average surface temperaure was plotted against
the efficiency (Figure 2). A linear fit resulted in a R-
squared value of 0.9034 and 0.9022 for PV with and
without PCM respectively, to be used in the co-simulation
setup. Figure 3: No-PCM PV baseline case annual efficiency
boxplots for 128 GS locations
In order to assess the impact of integrating PCMs into efficiency reduction, up to 0.53% in location #52 (Dong
PVs, the difference in average annual efficiencies Ujimqin Qi, China,. This was a reduction from 11.41% to
between with and without PCM cases was calculated for 10.88%. On the other hand, for a hotter (location #109
the 128 locations. This absolute difference in average ,Hissar, India), the absolute annual average PV efficiency
efficiencies was then plotted as a scatter chart, and the increased by 0.66% (from 9.26% to 9.92%). This is a
data colour coded using the baseline PV surface average relative increase in efficiency of 7.13% over the naturally
annual temperature (Figure 4). The figure highlights the cooled PVs. However, from Figure 4, we observe that
fact that the usefulness of PCM does indeed improve as most locations had a lower absolute improvement in the
the PV average surface temperature (and by association range of 0.2% - 0.4%.
the climate) gets hotter. To better understand the results, the best and worst
performing locations have been analysed in greater detail
in Figures 5 and 6. The daytime profile (Figure 5(b))
shows that the amplitude of the PCM based PV is smaller
for a period in April. This is also evident from the yearly
data (Figure 5(a)) where PCM integration results in a
more compressed histogram as compared to the no-PCM
case. Qualitatively, one may say that integration of PCMs
result in more consistent efficiency performance of PV by
preventing the efficiency from excessively dropping
during hot summer afternoons. Note that this is the
timeframe when the maximum amount of solar radiation
will be incident on the PV. In comparison, the worst
performing location (#52) results in consistently worse off
performance when PCMs are present (Figure 6). Here, the
PV surface temperature stays below 33.3°C, where PCM
results in lower efficiency. This is evident from the yearly
Figure 4: Effect of PCM on efficiency for the selected histogram as well as daily PV surface profile during April
128 Global south climates. in Figures 6(a) and 6(b).
In the colder locations, especially when the average
annual PV surface is below 15°C, PCMs result in PV

(b)
(a)
Figure 5: Example location where PCM integration into PV is an advantage. Figure (a) is a yearly histogram showing
that PCM integration results in an overall higher average efficiency. Figure (b) shows how the efficiency does not drop
as low when PCM are present.
 (a) (b)
Figure 6: Example location where PCM integration into PV is a disadvantage. Figure (b) shows the daytime profile in
April when PCM lead to significantly lower efficiency, also evident from the yearly results (Figure (a)).
Conclusion Dados, N., & Connell, R. (2012). The global south.
Contexts, 11(1), 12–13.
Based on the 885 Global South locations considered, 128
were identified with warm/hot summers where PCM DOE. (2020). Commercial reference buildings.
could sufficiently charge and discharge over the daily https://www.energy.gov/eere/buildings/commercia
cycle. Among these 128 locations, integration of PCMs l-reference-buildings
into PVs resulted in a relative improvement of Elarga, H., Fantucci, S., Serra, V., Zecchin, R., & Benini,
approximately 2% - 7% in average annual efficiency over E. (2017). Experimental and numerical analyses on
naturally cooled PV in 36 locations. Here, the PCM thermal performance of different typologies of
reduced the amplitude of oscillations in PV efficiency PCMs integrated in the roof space. Energy and
over the daily cycle by reducing the daily maximum and Buildings, 150, 546–557.
minimum efficiencies achieved, resulting in a more https://doi.org/10.1016/j.enbuild.2017.06.038
consistent performance as compared to the non-PCM EnergyPlus. (2012). Energy Simulation Software Version
case. For the remaining locations, integration of PCM 8.1 [Online]. US department of energy.
resulted in overall reduced performance, owing to an http://apps1.eere.energy.gov/buildings/energyplus/
average elevated PV average surface temperature. [Accessed: 6 December 2013]
While the methodology presented in this study is an EnergyPlus. (2022). Weather Data.
automated co-simulation approach which can utilize any https://energyplus.net/weather-
other experimental data on PV performance, and it can be region/europe_wmo_region_6
used for buildings energy modelling, it is cumbersome
and slow to simulate. An improvement would be to couple Gaglia, A. G., Lykoudis, S., Argiriou, A. A., Balaras, C.
the EnergyPlus PCM model into its PV model, which A., & Dialynas, E. (2017). Energy efficiency of PV
would remove the requirement for co-simulation. panels under real outdoor conditions–An
Nonetheless, the results of this study can be used to decide experimental assessment in Athens, Greece.
if such an effort is warranted. Renewable Energy, 101, 236–243.
Ghadikolaei, S. S. C. (2021). An enviroeconomic review
References of the solar PV cells cooling technology effect on
Aelenei, L., Pereira, R., Gonçalves, H., & Athienitis, A. the CO2 emission reduction. Solar Energy, 216,
(2014). Thermal performance of a hybrid BIPV- 468–492.
PCM: modeling, design and experimental Kant, K., Shukla, A., Sharma, A., & Biwole, P. H. (2016).
investigation. Energy Procedia, 48, 474–483. Heat transfer studies of photovoltaic panel coupled
Amirahmad, A., Maglad, A. M., Mustafa, J., & with phase change material. Solar Energy, 140,
Cheraghian, G. (2021). Loading PCM into 151–161.
buildings envelope to decrease heat gain- Khattak, S. H., Natarajan, S., & Chung, W. J. (2020). Out
performing transient thermal analysis on nanofluid with the power outages : Peak load reduction in the
filled solar system. Energy Res, 9, 727011. developing world. Windsor 2020 Resilient Comfort,
Curpek, J., & Hraska, J. (2017). Simulation study on 904–922.
thermal performance of a ventilated PV façade https://core.ac.uk/download/pdf/323491371.pdf
coupled with PCM. Applied Mechanics and Lazaro, A., Dolado, P., Marín, J. M., & Zalba, B. (2009).
Materials, 861, 167–174. PCM–air heat exchangers for free-cooling
 applications in buildings: Experimental results of photovoltaics considering the annual weather
two real-scale prototypes. Energy Conversion and conditions. Solar Energy, 105, 561–574.
Management, 50(3), 439–443. Sharma, S., Tahir, A., Reddy, K. S., & Mallick, T. K.
Mathworks. (2021). MATLAB. (2016). Performance enhancement of a Building-
https://uk.mathworks.com/products/matlab.html Integrated Concentrating Photovoltaic system using
Nada, S. A., & El-Nagar, D. H. (2018). Possibility of phase change material. Solar Energy Materials and
using PCMs in temperature control and Solar Cells, 149, 29–39.
performance enhancements of free stand and https://doi.org/https://doi.org/10.1016/j.solmat.201
building integrated PV modules. Renewable 5.12.035
Energy, 127, 630–641. Wetter, M., Haves, P., & Coffey, B. (2008). Building
Park, J., Kim, T., & Leigh, S.-B. (2014). Application of a controls virtual test bed. Lawrence Berkeley
phase-change material to improve the electrical National Lab.(LBNL), Berkeley, CA (United
performance of vertical-building-added States).

Appendix

Table 1: Location and EPW file names considered in this study

Location # Location and EPW name Location # Location and EPW name
1 ARE_Abu.Dhabi.412170_IWEC.epw 40 CHN_Henan.Lushi.570670_CSWD.epw
2 ARG_Buenos.Aires.875760_IWEC.epw 41 CHN_Jilin.Linjiang.543740_CSWD.epw
3 ARG_Chaco_Resistencia.Intl.AP_871550_ArgTMY.epw 42 CHN_Kinjiang.Uygar.Kashi.517090_SWERA.epw
4 ARG_Chaco_Saenz.Pena.AP_871490_ArgTMY.epw 43 CHN_Kinjiang.Uygur.Kuqa.516440_SWERA.epw
5 ARG_Corrientes_Corrientes.Intl.AP_871660_ArgTMY.epw 44 CHN_Kinjiang.Uygur.Yining.514310_SWERA.epw
6 ARG_Corrientes_Paso.de.los.Libres.Intl.AP_872890_ArgTMY.epw 45 CHN_Kinjiang.Uyhur.Hotan.518280_SWERA.epw
7 ARG_Entre.Rios_Concordia_873950_ArgTMY.epw 46 CHN_Liaoning.Chaoyang.543240_CSWD.epw
8 ARG_Entre.Rios_Gualeguaychu.AP_874970_ArgTMY.epw 47 CHN_Liaoning.Kuandian.544930_CSWD.epw
9 ARG_Formosa_Formosa.Intl.AP_871620_ArgTMY.epw 48 CHN_Nei.Mongol.Abag.Qi.Hot.531920_CSWD.epw
10 ARG_Misiones_Puerto.Iguazu.Intl.AP_870970_ArgTMY.epw 49 CHN_Nei.Mongol.Bairin.Zuo.Qi.540270_CSWD.epw
11 ARG_Santa.Fe_Ceres_872570_ArgTMY.epw 50 CHN_Nei.Mongol.Bayan.Mod.524950_CSWD.epw
12 ARG_Santa.Fe_Rosario.Intl.AP_874800_ArgTMY.epw 51 CHN_Nei.Mongol.Chifeng.542180_CSWD.epw
13 ARG_Santa.Fe_Sauce.Viejo.AP_873710_ArgTMY.epw 52 CHN_Nei.Mongol.Dong.Ujimqin.Qi.509150_CSWD.epw
14 BRA_Brasilia.833780_IWEC.epw 53 CHN_Nei.Mongol.Ejin.Qi.522670_CSWD.epw
15 BRA_Brasilia.833780_SWERA.epw 54 CHN_Nei.Mongol.Ejin.Qi.522670_SWERA.epw
16 BRA_Campo.Grande.836120_SWERA.epw 55 CHN_Nei.Mongol.Erenhot.530680_CSWD.epw
17 BRA_DF_Brasilia-Kubitschek.Intl.AP.833780_TRY.1962.epw 56 CHN_Nei.Mongol.Hailisu.532310_CSWD.epw
18 BRA_PR_Londrina.837680_INMET.epw 57 CHN_Nei.Mongol.Haliut.533360_CSWD.epw
19 BRA_SP_Campinas.837210_INMET.epw 58 CHN_Nei.Mongol.Hohhot.534630_CSWD.epw
20 CHL_Santiago.855740_IWEC.epw 59 CHN_Nei.Mongol.Jartai.535020_CSWD.epw
21 CHN_Beijing.Beijing.545110_CSWD.epw 60 CHN_Nei.Mongol.Jartai.535020_SWERA.epw
22 CHN_Beijing.Beijing.545110_IWEC.epw 61 CHN_Nei.Mongol.Jarud.Qi.540260_CSWD.epw
23 CHN_Beijing.Beijing.545110_SWERA.epw 62 CHN_Nei.Mongol.Jurh.532760_CSWD.epw
24 CHN_Gansu.Jiuquan.525330_CSWD.epw 63 CHN_Nei.Mongol.Linxi.541150_CSWD.epw
25 CHN_Gansu.Lanzhou.528890_CSWD.epw 64 CHN_Nei.Mongol.Linxi.541150_SWERA.epw
26 CHN_Gansu.Lanzhou.528890_IWEC.epw 65 CHN_Nei.Mongol.Otog.Qi.535290_CSWD.epw
27 CHN_Gansu.Lanzhou.528890_SWERA.epw 66 CHN_Nei.Mongol.Tongliao.541350_CSWD.epw
28 CHN_Gansu.Yumenzhen.524360_CSWD.epw 67 CHN_Nei.Mongol.Tongliao.541350_SWERA.epw
29 CHN_Gansu.Yumenzhen.524360_SWERA.epw 68 CHN_Nei.Mongol.Xi.Ujimqin.Qi.540120_CSWD.epw
30 CHN_Hebei.Fengning.543080_CSWD.epw 69 CHN_Nei.Mongol.Xilinhot.541020_CSWD.epw
31 CHN_Hebei.Huailai.544050_CSWD.epw 70 CHN_Ningxia.Hui.Yinchuan.536140_CSWD.epw
32 CHN_Heilongjiang.Anda.508540_CSWD.epw 71 CHN_Shaanxi.Yanan.538450_CSWD.epw
33 CHN_Heilongjiang.Harbin.509530_CSWD.epw 72 CHN_Shaanxi.Yulin.536460_CSWD.epw
34 CHN_Heilongjiang.Harbin.509530_IWEC.epw 73 CHN_Shaanxi.Yulin.536460_SWERA.epw
35 CHN_Heilongjiang.Harbin.509530_SWERA.epw 74 CHN_Shandong.Weifang.548430_CSWD.epw
36 CHN_Heilongjiang.Huma.503530_CSWD.epw 75 CHN_Shandong.Yanzhou.549160_CSWD.epw
37 CHN_Heilongjiang.Mudanjiang.540940_CSWD.epw 76 CHN_Shanxi.Datong.534870_CSWD.epw
38 CHN_Heilongjiang.Nenjiang.505570_CSWD.epw 77 CHN_Shanxi.Jiexiu.538630_CSWD.epw
39 CHN_Heilongjiang.Tonghe.509630_CSWD.epw 78 CHN_Shanxi.Taiyuan.537720_CSWD.epw
Table 1: Location and EPW file names considered in this study
79 CHN_Shanxi.Yuanping.536730_CSWD.epw
80 CHN_Shanxi.Yushe.537870_CSWD.epw
81 CHN_Sichuan.Batang.562470_SWERA.epw
82 CHN_Xinjiang.Uygur.Bachu.517160_CSWD.epw
83 CHN_Xinjiang.Uygur.Fuyun.510870_CSWD.epw
84 CHN_Xinjiang.Uygur.Hami.522030_CSWD.epw
85 CHN_Xinjiang.Uygur.Hotan.518280_CSWD.epw
86 CHN_Xinjiang.Uygur.Jinghe.513340_CSWD.epw
87 CHN_Xinjiang.Uygur.Kashi.517090_CSWD.epw
88 CHN_Xinjiang.Uygur.Kuqa.516440_CSWD.epw
89 CHN_Xinjiang.Uygur.Minfeng.518390_CSWD.epw
90 CHN_Xinjiang.Uygur.Tikanlik.517650_CSWD.epw
91 CHN_Xinjiang.Uygur.Turpan.515730_CSWD.epw
92 CHN_Xinjiang.Uygur.Turpan.515730_SWERA.epw
93 CHN_Xinjiang.Uygur.Yining.514310_CSWD.epw
94 CHN_Yunnan.Mengla.569690_CSWD.epw
95 DZA_Algiers.603900_IWEC.epw
96 EGY_Aswan.624140_ETMY.epw
97 EGY_Aswan.624140_IWEC.epw
98 EGY_Asyut.623930_ETMY.epw
99 EGY_Cairo.623660_IWEC.epw
100 EGY_Cairo.Intl.Airport.623660_ETMY.epw
101 EGY_El.Arish.623370_ETMY.epw
102 EGY_Ismailia.624400_ETMY.epw
103 EGY_Kharga.624350_ETMY.epw
104 EGY_Luxor.624050_ETMY.epw
105 IND_Akola.429340_ISHRAE.epw
106 IND_Bhopal.426670_ISHRAE.epw
107 IND_Bikaner.421650_ISHRAE.epw
108 IND_Gwalior.423610_ISHRAE.epw
109 IND_Hissar.421310_ISHRAE.epw
110 IND_Indore.427540_ISHRAE.epw
111 IND_Jabalpur.426750_ISHRAE.epw
112 IND_Jagdelpur.430410_ISHRAE.epw
113 IND_Jaipur.423480_ISHRAE.epw
114 IND_Jodhpur.423390_ISHRAE.epw
115 IND_Lucknow.423690_ISHRAE.epw
116 IND_Pune.430630_ISHRAE.epw
117 IND_Sholapur.431170_ISHRAE.epw
118 IRN_Shiraz.408480_ITMY.epw
119 IRN_Tabriz.407060_ITMY.epw
120 KEN_Garissa.637230_SWERA.epw
121 KEN_Makindu.637660_SWERA.epw
122 KWT_Kuwait.Intl.AP.405820_KISR.epw
123 LBY_Tripoli.620100_IWEC.epw
124 MAR_Casablanca.Nouasser.601560_IWEC.epw
125 PRY_Asuncion.862180_IWEC.epw
126 SAU_Riyadh.404380_IWEC.epw
127 SYR_Damascus.400800_IWEC.epw
128 ZWE_Harare.677750_IWEC.epw