See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/346039220

The effect of clearance height, albedo, tilt and azimuth angle in bifacial PV
energy estimation using different existing algorithms

Conference Paper · November 2020

CITATION READS

1 729

4 authors:

Hugo Sánchez-Ortiz Carlos Meza

Costa Rican Institute of Technology (ITCR) Costa Rican Institute of Technology (ITCR)
7 PUBLICATIONS   7 CITATIONS    63 PUBLICATIONS   737 CITATIONS   

SEE PROFILE SEE PROFILE

Sebastian Dittmann Ralph Gottschalg

Hochschule Anhalt Fraunhofer Institute for Microstructure of Materials and Systems IMWS
38 PUBLICATIONS   225 CITATIONS    364 PUBLICATIONS   3,171 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Uncertainty of measurement of PV modules View project

Reconfigurable photovoltaic array View project

All content following this page was uploaded by Hugo Sánchez-Ortiz on 16 February 2021.

The user has requested enhancement of the downloaded file.

Proceedings of the III Ibero-American Conference on Smart Cities (ICSC-2020) Page 315

The effect of clearance height, albedo, tilt and

azimuth angle in bifacial PV energy estimation
using different existing algorithms

Hugo Sánchez1,2 , Carlos Meza1,2 , Sebastian Dittmann2,3 , and Ralph

Gottschalg2,3
1
Electronics Engineering School, Costa Rica Institute of Technology, Costa Rica
[email protected]
2
Anhalt University of Applied Sciences, Köthen, Germany
3
Fraunhofer-Center for Silicon Photovoltaics CSP, Halle, Germany

Abstract. Bifacial photovoltaic is a promising technology that is being

consider more in solar installations worldwide. However, there is still a lot
of open questions regarding the estimation of the energy generation for
this technology. The present paper presents an analysis of the variables
involved in the modelling of the energy yield, as well as a quantitative
analysis for some of the most common models found in the literature. In
this regard, a series of simulations were made in order to compare the bi-
facial PV model energy yield in Bernburg, Germany and Cartago, Costa
Rica. A statistical comparison has been performed to analyse the accu-
racy and response for each variable. Also the algorithms are compared
with respect to the computational resources they use. The work presented
might serve as a base for validation with real data to contribute with the
accuracy of the estimation for the energy yield in bifacial modules.

Keywords: bifacial PV · photovoltaic energy simulation, photovoltaic

systems · Energy Rating

1 Introduction
It is well known that bifacial photovoltaic (PV) modules can utilise incident
sunlight via front and rear side absorption. This allows higher energy produc-
tion under various installation conditions such as large solar farms or roof-top
installations. However, the bifacial energy gain depends highly on, e.g., mod-
ule design, mounting method, and reflectivity of the background. However, the
bifacial energy gain depends highly on, e.g., module design, mounting method,
and reflectivity of the background. [1, 2, 4, 7, 12, 25]. Figure 1 shows a qualitative
comparison between the monofacial and bifacial modules.
Nowadays, the PV bifacial technology has been getting much attention from
stakeholders given the increased capacity of generation with practically the same
amount of area and system costs as monofacial PV modules [13]. However, there
is not yet a standard method to estimate the real energy output as it is the case
in monofacial PV technologies. The energy estimation on bifacial PV modules

ISBN 978-9930-541-79-1
Proceedings of the III Ibero-American Conference on Smart Cities (ICSC-2020) Page 316

2 Hugo Sánchez et al.

Fig. 1. Comparison between bifacial and monofacial module

presents even more challenges due to reflected rear side irradiance, which highly
depends on the location and the design of the PV system. To reach confident
estimations, it is necessary to consider several variables such as the height from
the floor of the module, the tilt angle, the background reflectance, sun elevation,
and diffuse irradiation [23]. The present paper compares different PV bifacial
power estimation models located in two different climates, i.e., equatorial fully
humid (Af) and warm temperate fully humid warm summer (Cfb) according
to the Köppen-Geiger Climate Classification [14]. The comparative analysis is
based on simulations executed with five different programs. The objective of
the present paper is to determine the different results obtained and how that
changes for different climates. The rest of the paper is structured as follows, first,
a literature review about the methods for estimation of energy yield in bifacial
modules as well as the main variables that affect the energy production in bifacial
models are presented. Then, a description of the main simulation programs used
in the comparison, methodology, and other considerations are shown. After that,
the paper describes the simulation results for the two places selected (Bernburg,
Germany, and Cartago, Costa Rica) and compares the energy yield between the
models. Finally, the synthesis of the results and conclusions are shown.

2 Methods for estimation of Energy Yield in the

Literature
Energy estimation model for bifacial modules have been studied by several au-
thors and research groups developing diverse methods with its own considera-
tions and assumptions ( [1,3,4,15–17,22,24]). However, it is possible to describe

ISBN 978-9930-541-79-1
Proceedings of the III Ibero-American Conference on Smart Cities (ICSC-2020) Page 317

Variables involved in the generation of bifacial PV modules 3

a general approach when the different energy estimation methods are compared.
Figure 2 shows a flow diagram of this general structure together with their tools
founded in the literature, which is further detailed in Table 1.

Fig. 2. General modelling flow for the estimation of energy in bifacial modules

Rear side irradiance analysis is a process that it is essential for the power
estimation of bifacial modules, and that haven’t been analysed in detailed until
recently. The next section presents the main considerations that need to be taken
into account for this case.

2.1 Estimation of the rear side irradiance

The front side irradiance estimation for bifacial PV modules can be adapted from
monofacial PV modules, which is already standardised and validated in several
models. [8]. However, the estimation of the rear side irradiance is relatively new
and several approaches can be found in the literature. In addition to the variabil-
ity of the solar resource during the day, the irradiance reflection on the surface
generates a more complicated phenomenon to analyse. Several studies founded
in the literature show that the energy output is dependent on the installation
conditions (e.g. elevation, orientation, tilt angle), as well as the environment con-
ditions (irradiance intensity, ground reflectively of different substrates, seasonal
variation) [9]. This situation opens the discussion about the optimisation in the
energy estimation for a bifacial module.
Most of the models found in the literature can be classified in three cate-
gories:: Empirical models, View Factor Models and Ray Trace Models. Figure 3

ISBN 978-9930-541-79-1
Proceedings of the III Ibero-American Conference on Smart Cities (ICSC-2020) Page 318

4 Hugo Sánchez et al.

Table 1. Description for the simulations steps in the energy estimation for bifacial
modules

Stage Description Output Results

This is the most common methods used – Sun Path
because it is completely studied and – Direct Irradiance POA
Modelling front
analyzed for the estimation of energy – Diffuse Irradiance POA
side irradiance
in monofacial modules. Also it is stan- – Albedo Irradiance POA
darized process.
– Direct Irradiance POA
Then, the effects of the reflected irradi- – Diffuse Irradiance POA
Modelling rear
ance is estimated. As a result, the the – Albedo POA
side irradiance
bifacial relatively gain is obtained. – Bifacial Gain

– Power conversion
– Efficiency
In this step, the monofacial module en-
Electro- – Thermal Losses
ergy is estimated. For this step the rules
Thermal Model – NOCT
already analyzed and tested are used
– Monofacial Energy Yield

A final corrections are applied according – Corrections

Final System the suppositions made in all the models. – Electrical Losses
Performance There is a difference in the methods used – Mismatch
for every model found.

shows an overview and a classification of nine energy yield models for bifacial
PV modules.
The Empirical Models corresponds to the group of models that were ob-
tained using short or long therm experimental data. The View Factor Models
correspond to the utilisation of geometrical construction as a definition of ther-
modynamics. Ray Tracing models corresponds to the models that use software
for the estimation of the light rays through a determined conditions in a 3D
environment.

2.2 Variables that affects the energy production in bifacial models

Several variables affect the energy production of bifacial PV modules but also
the energy yield estimation. The present section gives a primary definition of
these variables. Figure 4 shows the main variables which affects the energy yield
estimation of bifacial PV modules. Namely,

– Albedo (albedo): is the diffuse reflectivity of a horizontal surface which is

defined as the ratio of the reflected irradiance by surface and the received
irradiance. It is measured on a scale from 0 to 1.
– Rear/front-side irradiance ratio (α): is the diffuse reflectivity of a tilted
surface which is defined as the ratio of the reflected irradiance by surface

ISBN 978-9930-541-79-1
Proceedings of the III Ibero-American Conference on Smart Cities (ICSC-2020) Page 319

Variables involved in the generation of bifacial PV modules 5

Modelling rear- Irradiance

Empirical Model View Factor Model Ray Trace Model

Prism Solar [3] PV Syst [17] NREL Bifacial [5]

Solar World [15] NREL Bifacial VF Fraunhofer ISE [20]

[16]

BSolar [2] RWTH and ISC Kon-

stanz [24]

Purdue University
[22]

Fig. 3. Comparison for modelling rear-irradiance

and the received irradiance on the rear side of the bifacial PV module. It is
measured on a scale from 0 to 1.
– Clearance Height (CH): is defined as the distance from the lowest point of
the solar module to the ground. This influences the collection of the rear-side
irradiance.
– Tilt Angle (β): refers to the inclination, measured in degrees, that the edge
of the PV module forms with a parallel line to the ground. This factor will
influence both front and rear side irradiance collection in the solar module.
– Module Height(M H): refers to the dimension of the PV module, seen in
the lateral plane. It is dependent on the orientation of the module (landscape
or portrait). This factor has an impact on the self-shadowing of the solar
module and the possibility to collect rear-side irradiance.
– Azimuth(γ): refers to the compass direction in which the solar module or
array is installed. It affects both front and rear side energy production due
to the changing position of the sun during the day.
– Ground Coverage Ratio(gcr): it is defined as the ratio of the PV mod-
ules’area to the used land area. This is an important aspect when the analysis
considers more than one array. It can be also defined as the ratio of array
length to row-to-row pitch. A definition and discussion about this term can
be found in [6, 7, 24]

3 Bifacial PV power estimation programs

The models considered for comparisons are listed next:

ISBN 978-9930-541-79-1
Proceedings of the III Ibero-American Conference on Smart Cities (ICSC-2020) Page 320

6 Hugo Sánchez et al.

Fig. 4. Factors that affect the bifacial gain energy

– Purdue University Bifacial: this is a model that uses View Factor concept
from thermodynamics developed by researchers at Purdue University for the
estimation of bifacial energy generation. The information on the model can
be found in [22]
– Prism Solar: This model was developed by the company Prism Solar. It
consist in a general linear regression that allows predicting the behaviour of
the solar bifacial tools. The linear regression was validated with real data
for a small range of values. The model and its validation can be found in [3]
– Solar World: The model was developed by the company Solar World, using
data analytic from three different systems installed and then obtaining a
mathematical regression from the data. The description of model as well as
the experiment can be found in [15]
– NREL Radiance Bifacial: It consists in a open source software that uses
an integration of the back-end software Radiance for ray tracing estimation.
The considerations for the software as well as the validation of the model
can be found in [18].
– PVSyst: Since 2017 the software includes the feature for simulation for bifa-
cial modules. The software uses the view factor geometry for the installation.
The details of the models can be found in [17].

The models of Prism Solar, Solar World and Purdue University Bifacial
were implemented in Python v3.7 with the integration of the library PVLib [21].
For the case of PV Syst, the version used was the v6.8.6. In the case of NREL
Radiance Bifacial, the release used was the 5.2 that is available in [5]. In the
next section the simulation study and the comparison between the models are
presented.

4 Simulation study
To compare the bifacial PV module power estimation results of the considered
algorithms (Purdue University Model, Prism Solar Model, Solar World Model,

ISBN 978-9930-541-79-1
Proceedings of the III Ibero-American Conference on Smart Cities (ICSC-2020) Page 321

Variables involved in the generation of bifacial PV modules 7

Radiance Bifacial Model and PVSyst Model ) a simulation case study has been
designed. The places for study correspond to the location of the PV Outdoor
Characterisation of Anhalt University of Applied Sciences at Bernburg, Ger-
many, and the experimental PV plant from the Costa Rica Institute of Tech-
nology in Cartago, Costa Rica so that in the future it will be possible to run
experiments to validate the simulated results. The parameters used in the sim-
ulation study are shown in Table 2.

Table 2. Main Parameters for the sites of analysis

Module Module
Latitude Longitude Tilt Angle
Location Landscape Portrait Bifaciality [%]
[◦ ] [◦ ] [◦ ]
Height [m] Height [m]

Bernburg 51.773 11.763 35 1 2 90

Cartago 9.853 -83.912 15 1 2 90

The module height for both configurations (landscape and portrait) is as-
sumed for the standard dimensions for a 72 solar cell PV module. The tilt angle
is defined to maximise the power energy production of a monofacial PV mod-
ules in each site. The bifaciality is chosen in 90% as it is common to find in
commercial modules. The 22 year meteorological database from NASA is used.
Additional parameters used in the simulation are shown in Table 3.

Table 3. Main physical variables used in the

Parameter Value Unit

◦
Azimut Angle 180
Front Site Efficiency 18 %
Bifaciality 90 %
Temperature Coefficient -0.4319 %/K
U0 Constant heat transfer 22.7 W/m2 /K
U1 Convective heat transfer 6.84 W.s/m3 /K

Only the landscape configuration was used given that this is the case from
which more energy can be obtained. As a metric to compare the simulation re-
sults the same methodology for the analysis of the energy performance of the
models in a Round Robin campaigns is used [8]. These metrics consider the de-
viation with respect to the average value. In this regards, for all the variables
analysed the median value is calculated and then the Root Mean Square De-
viation (RMSD) and the Mean Bias Deviation(MBD) are estimated using as a
reference value the median for the all measurement.

ISBN 978-9930-541-79-1
Proceedings of the III Ibero-American Conference on Smart Cities (ICSC-2020) Page 322

8 Hugo Sánchez et al.

5 Results
The results of the simulation study are shown in figures 5, 6, 7 and 8 and are
discussed in the following sections.

5.1 Clearance Height

It is known that there is a dependence between the clearance height of the module
and the reflected irradiance. The parameters used to evaluate the effect of the
clearance height in the models are shown in Tables 2 and 3 and the following
values have remained constant for all simulations:
– Rear/front-side irradiance ratio: 25%
– GCR: 50%
– Clearance Height: from 0.5m to 4m, in steps of 0.1m.
Figure 5 shows the results of the effect of the clearance height on the energy
for the selected models, where it can be seen that as the height of the module
increases the energy yield increases. The slope of the dependence is larger below
1,5 meters of height. Except for the Prism Solar model, the models follow the
same trend for the height variation. The most conservative model for this vari-
able is the Solar World model. The most optimistic model corresponds to the
Radiance Bifacial for both locations. The curve from Prism Solar is expected
because its authors stated that is only valid in a range from 0.5m and 1.2m. The
calculation of RMSD and MBD gives a statistical comparison between modules.
In this case, the average value of all models is estimated and this value is used
as the reference for the estimation of the errors. Table 4 shows the calculation
of RMSD and MBD.

Fig. 5. Comparison height effects for different models Bernburg and Cartago

Based on the results shown in 4 the Radiance model has the closest approx-
imation in the case of Bernburg. On the other hand, the PUB model has the
closest approximation for the case of Cartago. The model of Prism Solar is the
farthest approximation in both cases.

ISBN 978-9930-541-79-1
Proceedings of the III Ibero-American Conference on Smart Cities (ICSC-2020) Page 323

Variables involved in the generation of bifacial PV modules 9

Table 4. Summary of comparison clearance height effects between different models

Site Bernburg Cartago

Model RMSD [%] MBD [%] RMSD [%] MBD [%]
PUB 2.666 -2.331 2.2270 -1.935
Prism Solar 12.925 10.851 10.324 7.643
SolarWorld 6.019 -5.619 4.936 -4.2932
Radiance 1.748 0.505 3.923 3.890
PVSyst 3.938 -3.405 4.564 -4.087

5.2 Rear/front-side irradiance ratio

Rear/front-side irradiance ratio has an almost linear influence on the reflected
irradiance and therefore in the bifacial energy estimation. In addition to the
parameters given in Tables 2 and 3, the next values were set constant for all
simulations:

– Rear/front-side irradiance ratio: from 10% to 70%, in 10% steps.

– GCR: 50%
– Clearance Height: 1.5m

Figure 6 shows the simulation results between the selected models.

Fig. 6. Comparison Rear/front-side irradiance ratio effects for different models Bern-
burg and Cartago

In this case, it is possible to identify a linear relationship between the energy

yield and the rear/front-side irradiance(RFI) value. For higher RFI more energy
at the rear is expected. Except for the Prism Solar model, the models follow the
same trend for the RFI variation. The most conservative model for this variable
is the Pv Syst model. The most optimistic model corresponds to the Radiance
Bifacial model for both locations. However the model from Prism Solar also has

ISBN 978-9930-541-79-1
Proceedings of the III Ibero-American Conference on Smart Cities (ICSC-2020) Page 324

10 Hugo Sánchez et al.

a slope but it is to small that graphically looks like constant value. Using the
same methodology from this analysis, the RMSD and MBD is estimated. Table
4 shows the calculation of RMSD and MBD.

Table 5. Summary of comparison rear/front-side irradiance ratio effects between dif-

ferent models

Site Bernburg Cartago

Model RMSD [%] MBD [%] RMSD [%] MBD [%]
PUB 4.594 3.085 3.764 2.361
Prism Solar 6.517 0.787 7.439 -2.597
SolarWorld 2.749 -1.534 1.916 -0.343
Radiance 7.012 5.766 11.483 9.658
PVSyst 8.715 -8.104 9.944 -9.080

The obtained results shown that the Prism Solar model has the closest ap-
proximation for the case of Bernburg and for the Solar World model in the case
of Cartago. On the other hand, the PV Syst is the farthest approximation in
Bernburg and the Radiance model is the farthest approximation in Cartago.

5.3 Tilt Angle

Due to the movement of the Sun during the year the tilt angle has an influence on
the performance of the installation. This situation is well analysed for monofacial
PV modules. For the tilt angle analysis, in addition to the parameters given in
Tables 2 and 3, the next values are maintain fixed for all simulations:

– Rear/front-side irradiance ratio: from 25%

– GCR: 50%
– Clearance Height: 1.5m
– Tilt Angle: from 20◦ to 50◦ in Bernburg. From 0◦ to 30◦ in Cartago.

Figure 7 shows the comparison between the selected models.

In contrast previous analysis all the models present more variability in the
energy yield with respect to changes of the tilt angle. The most conservative
model for this variable is the Solar World model in both sites. The most opti-
mistic model corresponds to the Prism Solar model in Bernburg and the Ra-
diance model in Cartago. Using the same methodology from this analysis, the
RMSD and MBD are estimated. Table 6 shows the calculation of RMSD and
MBD.
The PUB model has the closest approximation in the Bernburg case and the
PV Syst model has the closest approximation for the case of Cartago. The Prism
Solar model is the furthest in the case of Bernburg and the Radiance model is
the furthest in the case of Cartago.

ISBN 978-9930-541-79-1
Proceedings of the III Ibero-American Conference on Smart Cities (ICSC-2020) Page 325

Variables involved in the generation of bifacial PV modules 11

Fig. 7. Comparison tilt angle effects for different models Bernburg and Cartago

Table 6. Summary of comparison tilt angle effects between different models

Site Bernburg Cartago

Model RMSD [%] MBD [%] RMSD [%] MBD [%]
PUB 0.756 -0.727 1.332 -1.332
Prism Solar 5.902 5.825 2.066 1.571
SolarWorld 4.130 -3.969 3.548 -3.463
Radiance 2.463 2.408 3.560 3.528
PVSyst 3.585 -3.536 0.326 -0.313

5.4 Azimuth
For the azimuth angle simulation the following variables are kept constant:

– Rear/front-side irradiance ratio: from 25%

– GCR: 50%
– Clearance Height: 1.5m
– Azimuth: from 90◦ to 270 ◦

Figure 8 shows the comparison between the selected models. It important

to highlight that the response for this variable is similar than the monofacial
modules. The models have the same trend of variation for both sites. However
the sensibility to the azimuth angle is bigger in the case of Bernburg where the
optimal angle is close to 180◦ . The most conservative model for this variable
is the PV Syst model. The most optimistic model corresponds to the Radiance
Bifacial model for both locations. However the model from Prism Solar also has
a slope but it is to small that graphically looks like constant value. Using the
same methodology from this analysis, the RMSD and MBD is estimated.
Table 4 shows the calculation of RMSD and MBD.
Compared to the average value, the PV Syst model has the closest approx-
imation in both places. On the other hand, the Solar World has the farthest
approximation in both places.

ISBN 978-9930-541-79-1
Proceedings of the III Ibero-American Conference on Smart Cities (ICSC-2020) Page 326

12 Hugo Sánchez et al.

Fig. 8. Comparison azimuth effects for different models Bernburg and Cartago

Table 7. Summary of comparison azimuth effects between different models

Site Bernburg Cartago

Model RMSD [%] MBD [%] RMSD [%] MBD [%]
PUB 3.367 3.321 1.189 -1.167
Prism Solar 3.260 2.8795 1.640 1.632
SolarWorld 6.722 -6.651 3.431 -3.428
Radiance 1.873 1.093 3.131 3.125
PVSyst 1.094 -0.642 0.473 -0.161

5.5 Computational Resources

In the previous section and analysis for different models under different variables
was performed. This analysis gives an idea of the accuracy of the different models
in every case. When simulation methods are compared, it is also necessary to
have a look at the computational resources and performance for every algorithm.
For this purpose, Table 8 shows a quantitative comparison for the result of every
method available. The table is made using as a reference the work presented
in [19].
From the previous table, the models from PV Syst and Bifacial Radiance
from NREL that can estimate more features regarding the simulation. Looking
for the time consuming aspect, the Bifacial Radiance model is the most time-
consuming algorithm, due to the ray-tracing solving process. In the case of Prism
Solar and Solar World, the running time is faster. These models are simplified
for a simple equation. Nevertheless, these models are more liable to underfitting
or overfitting the energy yield estimation in bifacial models.

A balance between accuracy and use of the computational resources has to

be consider for the creation of suitable tool for the estimation of energy yield in
bifacial models.

ISBN 978-9930-541-79-1
Proceedings of the III Ibero-American Conference on Smart Cities (ICSC-2020) Page 327

Variables involved in the generation of bifacial PV modules 13

Table 8. Quantitative comparison between models for the energy estimation in bifacial
models

Prism Solar RTWH

Quantity PUB PV Syst Radiance
Solar World & ISC
Albedo, tilt, h, gcr D1 D D D D D
Panel Size X2 D3 D3 D3 D D
BGE resolution Yearly Yearly Hourly Monthly Hourly Hourly
Running time for
< 1s < 1s < 60s < 60s < 5s < 60s
yearly simulation4
Irradiance profiles - - - D D D
D D D D5
Ground incident irra- 5 5 6
- -
diance
D
Glass-Air Transmis-
- - - - -
sion Loss
D D
Light transmitted
- - - -
between cells
Shading Losses - - D - D 6
D
D D D
Specific nuMBDr &
- - 1
size rows
D D
Edge effects mod-
- - - -
elled
1-axis tracking - - - - D D
1
Row spacing is assumed to be equal to or larger than the typical row spacing at noon
on DeceMBDr 21st .
2
The model is designed for a specific commercial modules.
3
Geometry is scaled relative to the collector width.
4
Simulations performed on a ASUS PREDATOR computer equipment with a Intel
Core 17-7700HQ processor and 16 GB of RAM
5
Perez diffuse sky irradiance
6
Isotropic diffuse sky irradiance
7
Consider all the shadows with a shadow factor

6 Synthesis of results

The present paper contains an analysis of the influence of different variables

in the energy yield estimation for bifacial modules. The factor analysed are
clearance height, Rear/front-side irradiance ratio, tilt angle, and azimuth. The
second part contains a comparison for every variable consider in the mathemat-
ical models found in the literature. Using Round Robin methodology presented
in [8, 10, 11] a summary of the errors obtained are shown in Figure 9.
As was studied in the present chapter and making an analysis of Figure 9,
Table 9 shows a summary of the comparison obtained.
The previous table shows the typical curve for every variable. Also, it shows
which is the most accurate prediction as well as the less accurate prediction
model for every variable. This accuracy is estimated compared to the mean val-

ISBN 978-9930-541-79-1
Proceedings of the III Ibero-American Conference on Smart Cities (ICSC-2020) Page 328

14 Hugo Sánchez et al.

Fig. 9. Comparison albedo effects for different models Bernburg and Cartago

Table 9. Summary of the comparison in energy yield estimation models

Site Bernburg Cartago

13 23
Variable Curve MAA LAA MAA13 LAA23
Clearance Height Bounded Exponential Radiance Prism Solar PUB Prism Solar
α Linear increasing Prism Solar PV Syst Solar World Radiance
Tilt Angle Polynomial PUB Prism Solar PV Syst Radiance
Azimuth Parabola PV Syst Solar World PV Syst Solar World
1
More Accurate Algorithm.
2
Less Accurate Algorithm.
3
Compared to the mean value of the simulations.

ues for all models. When the RMSD and MBD are estimated, all the models
report less than 10 % of error. This value indicates that all the models could fit
the estimation. However, there are some atypical cases such as the Prism Solar
model. This model has a different response when the clearance height and the
albedo change. According to the error estimated, the more acceptable models for
both places are PUB model and PV Syst model. In Bernburg, Radiance model
has acceptable accuracy.

Another important point to consider is the computational performance of

the algorithms. On this topic, it has to be considered the balance between the
computational resources and the possible outputs obtained. Table 8 presents an

ISBN 978-9930-541-79-1
Proceedings of the III Ibero-American Conference on Smart Cities (ICSC-2020) Page 329

Variables involved in the generation of bifacial PV modules 15

analysis for all the models used. From this table, the PV Syst model and the Ra-
diance model give output variables and information. Nevertheless, the Radiance
model is one of the most time-consuming algorithms to make and estimation
along with the PUB model. Despite their differences in assumptions and com-
plexity, there is a good agreement between all models.

7 Conclusions and outlook

The present work shows how the estimation of bifacial modules is affected by
several factors such as Clearance Height, Albedo, Tilt Angle, and Azimuth. Also,
the practical comparison of different models found in the literature. Due to
the penetration in the market of bifacial modules it is necessary to elaborate
on accurate tools for the process, planning and design. In the case of the Ray
Trace Models it is still not clear the accuracy of the model compared to the
View Factor Models (PUB, PVSyst), that present less deviation in the statistical
analysis. On the other hand, the Ray Trace Models requires more time and more
computational resources to converge into a result.
It is necessary to carry out analyses with field validation data to get a more
realistic analysis, as well as considering the changes that the variables can present
during the year. Bifacial technology offers interesting applications and benefits
compare to the monofacial modules. For this reason, it is necessary to develop
accurate tools to help other PV professionals and stakeholders in the making-
decision process. And with that, contribute to the transition to a clean energy
age.

References
1. Appelbaum, J.: Bifacial photovoltaic panels field. Renewable Energy 85, 338–343
(2016)
2. Bsolar: Commercial test sites and outdoor field results. http://www.b-solar.com/
Technology.aspx?Sel=Field%20Results, accessed: 2020-04-26
3. Castillo-Aguilella, J.E., Hauser, P.S.: Multi-variable bifacial photovoltaic module
test results and best-fit annual bifacial energy yield model. Ieee Access 4, 498–506
(2016)
4. Cuevas, A., Luque, A., Eguren, J., del Alamo, J.: 50 per cent more output power
from an albedo-collecting flat panel using bifacial solar cells. Solar Energy 29(5),
419–420 (1982)
5. Deline, C., Ayala, S.: Bifacial radiance. [Computer Software] https://doi.org/10.
11578/dc.20180530.16 (12 2017), https://doi.org/10.11578/dc.20180530.16
6. Deline, C., Dobos, A., Janzou, S., Meydbray, J., Donovan, M.: A simplified model
of uniform shading in large photovoltaic arrays. Solar Energy 96, 274–282 (2013)
7. Deline, C., MacAlpine, S., Marion, B., Toor, F., Asgharzadeh, A., Stein, J.S.: Eval-
uation and field assessment of bifacial photovoltaic module power rating method-
ologies. In: 2016 IEEE 43rd Photovoltaic Specialists Conference (PVSC). pp. 3698–
3703. IEEE (2016)

ISBN 978-9930-541-79-1
Proceedings of the III Ibero-American Conference on Smart Cities (ICSC-2020) Page 330

16 Hugo Sánchez et al.

8. Dittmann, S., Friesen, G., Williams, S., Betts, T., Gottschalg, R., Beyer, H.,
de Montgareuil, A.G.: Results of the 3rd modelling round robin within the eu-
ropean project „performance”–comparison of module energy rating methods. In:
Proceedings of the 25th European Photovoltaic Solar Energy Conference, Valen-
cia, Spain. pp. 6–10 (2010)
9. Dittmann, S., Sanchez, H., Burnham, L., Gottschalg, R., Oh, S.Y., Benlarabi, A.,
Figgis, B., Abdallah, A., Rodriguez, C., Rüther, R., et al.: Comparative analysis of
albedo measurements(plan-of-array and horizontal at multiple sites worldwide. In:
36th European Photovoltaic Solar Energy Conference (EU PVSEC). pp. 1388–1393
(2019)
10. Friesen, G., Gottschalg, R., Beyer, H., Williams, S., van Sark, W., Guérin de Mont-
gareuil, A., Van Der Borg, N., Huld, T., Müller, B., De Keizer, A., et al.: Inter-
comparison of different energy prediction methods within the european project"
performance"-results of the 1st round robin. In: 22nd European Photovoltaic Solar
Energy Conference. pp. 2659–2663. WIP-Renewable Energies (2007)
11. Friesen, G., Dittmann, S., Williams, S., Gottschalg, R., Beyer, H., de Montgareuil,
A.G., Van Der Borg, N., Burgers, A.R., Kenny, R.P., Huld, T., et al.: Inter-
comparison of different energy prediction methods within the european project
„performance”-results of the 2nd round robin. In: Proceedings of the 24th Euro-
pean Photovoltaic Solar Energy Conference. pp. 3189–3197 (2009)
12. Guerrero-Lemus, R., Vega, R., Kim, T., Kimm, A., Shephard, L.: Bifacial solar
photovoltaics–a technology review. Renewable and sustainable energy reviews 60,
1533–1549 (2016)
13. ITRPV: International technology roadmap for photovoltaic (itrpv)-2019 results
(2020)
14. Kottek, M., Grieser, J., Beck, C., Rudolf, B., Rubel, F.: World map of the köppen-
geiger climate classification updated. Meteorologische Zeitschrift 15(3), 259–263
(2006)
15. Kutzer, M., Fülle, A., Jahnke, A., Hahn, J., Wendt, S., Neuhaus, D., Witzig,
A., Kutzer, K.: Ertragssteigerung durch bifaciale modultechnologie. In: Proc. 31st
Symp. Photovolt. Sol. Energy. pp. 1–10 (2016)
16. Marion, B., MacAlpine, S., Deline, C., Asgharzadeh, A., Toor, F., Riley, D., Stein,
J., Hansen, C.: A practical irradiance model for bifacial pv modules. In: 2017 IEEE
44th Photovoltaic Specialist Conference (PVSC). pp. 1537–1542. IEEE (2017)
17. Mermoud, A., Wittmer, B.: Bifacial shed simulation with pvsyst. In: Bifacial Work-
shop. pp. 25–26 (2017)
18. Pelaez, S.A., Deline, C., Greenberg, P., Stein, J.S., Kostuk, R.K.: Model and vali-
dation of single-axis tracking with bifacial pv. IEEE Journal of Photovoltaics 9(3),
715–721 (2019)
19. Pelaez, S.A., Deline, C., MacAlpine, S.M., Marion, B., Stein, J.S., Kostuk, R.K.:
Comparison of bifacial solar irradiance model predictions with field validation.
IEEE Journal of Photovoltaics 9(1), 82–88 (2018)
20. Reise, C., Schmid, A.: Realistic yield expectations for bifacial pv systems—an
assessment of announced predicted and observed benefits. In: Proc. 31st Eur. Pho-
tovolt. Sol. Energy Conf. Exhib. pp. 1775–1779 (2015)
21. SANDIA Labs: Pvlib tool box. https://pvlib-python.readthedocs.io/en/stable/,
accessed: 2020-03-15
22. Sun, X., Khan, M.R., Deline, C., Alam, M.A.: Optimization and performance of
bifacial solar modules: A global perspective. Applied energy 212, 1601–1610 (2018)

ISBN 978-9930-541-79-1
 Proceedings of the III Ibero-American Conference on Smart Cities (ICSC-2020) Page 331

Variables involved in the generation of bifacial PV modules 17

23. Wang, S., Wilkie, O., Lam, J., Steeman, R., Zhang, W., Khoo, K.S., Siong, S.C.,
Rostan, H.: Bifacial photovoltaic systems energy yield modelling. Energy Procedia
77, 428–433 (2015)
24. Yusufoglu, U.A., Pletzer, T.M., Koduvelikulathu, L.J., Comparotto, C., Kopecek,
R., Kurz, H.: Analysis of the annual performance of bifacial modules and optimiza-
tion methods. IEEE Journal of Photovoltaics 5(1), 320–328 (2014)
25. Yusufoglu, U.A., Lee, T.H., Pletzer, T.M., Halm, A., Koduvelikulathu, L.J., Com-
parotto, C., Kopecek, R., Kurz, H.: Simulation of energy production by bifacial
modules with revision of ground reflection. Energy Procedia 55, 389–395 (2014)

ISBN 978-9930-541-79-1

View publication stats