Master of Science Thesis
KTH School of Industrial Engineering and Management Energy Technology TRITA-ITM-EX 2020:418
Division of Heat & Power SE-100 44 STOCKHOLM
Design of a methodology to identify optimum PV configurations for small- scale modular solutions for tropical
Paul DUPONT DE DINECHIN
Master of Science Thesis TRITA-ITM-EX 2020:418
Design of a methodology to identify optimum PV
configurations for small-scale modular solutions for tropical
Paul DUPONT DE DINECHIN
Commissioner Contact person
The development of small-scale solar photovoltaic (PV) power plants is accelerating in most isolated
regions, such as the Caribbean. Indeed, they offer a beneficial, ecological and modular alternative to reduce
the important fossil fuel dependency of these countries. However, since modules are facing a harsh
environment in these areas, the PV project configuration cannot follow the same rules as in a “classical” PV
project: the project specifications and design techniques are different as more constraints must be taken into
consideration. Therefore, designing PV configurations for such projects requires a specific methodology
adapted to the tropical and cyclonic conditions. This thesis work aims to design such a methodology for a
100 kWc solar shed: using the available literature on PV modules’ behavior under tropical conditions, a set
of adapted specifications is defined, and several technical components are selected followingly. A cost
analysis is undertaken and assesses the total CAPEX (∼220,000 €) and the OPEX (5,650 €/year) of the
project. In order to select the module that fits best the customer’s requirements, an economical and
ecological grading system is used, this choice is followed by the PV field design. The project performance
is assessed using PVSyst, and the results are used to compute the different economic indicators using a
MATLAB program. This methodology is tested on a real case study, the main outputs are total yield of
166.0 MWh, capacity factor of 0.19, project LCOE of 147.93€/MWh, IRR of 5.35% and a potential annual
rent to the customer equal to 2,900€/year. Several sensitivity analyses are conducted to judge and explore
the potential enhancements to this methodology, it appears that the ecological grading index definition
should be modified, and that orientation and inverter choice optimization could be added to the
methodology as long as they don’t add too much complexity.
Utvecklingen av småskaliga solceller (PV) kraftverk accelererar i de flesta isolerade regioner, till exempel Karibien. De erbjuder faktiskt ett fördelaktigt, ekologiskt och modulärt alternativ för att minska det viktiga fossilbränsleberoendei dessa länder. Men eftersom moduler står inför en hård miljö inom dessa områden kan inte PV-projektkonfigurationen följa de samma reglerna som i ett ”klassiskt” PV-projekt:
projektspecifikationerna och designteknikerna är olika eftersom fler tryck måste beaktas. Därför krävers en specifik metod anpassad till de tropiska och cykloniska förhållandena för att utforma PV-
konfigurationer för sådana projekt. Detta master uppsats syftar till att utforma en sådan metod för en 100
kWc solskjul: med hjälp av tillgänglig litteratur om PV-modulers beteende under tropiska förhållanden
definieras en uppsättning anpassade specifikationer och flera tekniska komponenter väljs därefter. En
kostnadsanalys genomförs och utvärderar projektets totala CAPEX (∼220 000 €) och OPEX (5 650 € /
år). För att välja den modul som passar bäst kundens krav används ett ekonomiskt och ekologiskt
betygssystem; PV-fältdesignen väljs sedan. Projektets prestanda bedöms med hjälp av PVSyst, och
resultaten används för att beräkna de olika ekonomiska indikatorerna med ett MATLAB-program. Denna
metod testas på en riktig fallstudie, huvudutgångarna är: totalavkastning på 166,0 MWh, kapacitetsfaktor
0,19, projekts LCOE på 147,93 € / MWh, IRR på 5,35% och en potentiell årlig hyra till kunden på 2 900 €
/år. Flera känslighetsanalyser genomförs för att bedöma och utforska de potentiella förbättringarna av
denna metod, det verkar som att den ekologiska graderingsindexdefinitionen bör ändras, och att
orientering och växelriktaresvaloptimering skulle kunna läggas till metoden så länge de inte komplicerar
den för mycket.
As my master thesis reaches its end, I would like to express my upmost gratitude to all the persons and institutions who made possible this thesis work and more generally these 2 fruitful years of master within the Sustainable Energy Engineering program at KTH Royal Institute of Technology. They have been crucial in my development as a renewable energy engineer, providing me the technical tools and, more importantly, an inspiring human experience, to confront the numerous challenges our world is facing, in the power generation sector, just as in all other sectors.
My first though goes to my family, and especially my parents, for their constant support and confidence. It goes without saying that I owe my entire education to them, they have always led me to be a better person with benevolence and determination.
I am also sincerely grateful to both my universities : on the one hand, CentraleSupelec, which gave me a solid scientific background, a thrilling life experience at the university campus in Chatenay-Malabry and the opportunity to conduct this double degree in Sweden, with the constant help of my coordinator Carine Morotti Delorme; thank you also to Hervé Duval, who accepted to supervise this work for CentraleSupelec.
On the other hand, KTH Royal Institute of Technology, where the whole Sustainable Energy Engineering Master Program was beyond expectations. I would like to thank my supervisor at KTH, Rafael Eduardo Guedez, firstly for his inspiring course Large scale solar power, which led me to conduct this thesis, but more importantly for his help and advice during this work. Thank you also to Björn Laumert for offering me the opportunity to conduct my thesis within the division of Heat and Power Technology, to my coordinator Elin Wiljergård, and to all the international students I met in KTH who made this experience incomparable.
Finally, I would like to express my gratitude to the whole team of Albioma Solaire Antilles. These 6 month
of thesis work within the company have been so rewarding, I could not have hoped for a better place to
learn concretely about solar energy. Many thanks to my supervisor Leslie Bocaly, to the director Sébastien
Leydier, and to the project engineer Charles Grosy, for having offered me the opportunity to work within
ASA, and for all the knowledge they shared with me on solar project development.
Table of Contents
Abstract ... 3
Sammanfattning ... 4
Acknowledgment ... 5
List of Figures ... 9
List of Tables ...10
1 INTRODUCTION ...13
1.1 Background ...13
1.1.1 Context ...13
1.1.2 Albioma Solaire Antilles presentation ...13
1.2 Objectives ...13
1.3 General methodology ...14
1.4 Scope of work ...14
1.5 Reading instructions ...15
2 THEORETICAL FRAMEWORK AND SPECIFICATIONS ...16
2.1 PV Fundamentals ...16
2.1.1 PV effect ...16
2.1.2 PV Module structure ...17
2.1.3 PV cell/module characteristics ...17
2.1.4 Maximum Power Point Tracker (MPPT) ...18
2.2 Temperature and Humidity impact on PV performance ...19
2.2.1 Temperature effect on PV performance ...19
2.2.2 Humidity effect on PV performance ...20
2.2.3 Potential Induced Degradation ...22
2.3 Solar development background in the Caribbean. ...25
2.3.1 Energy framework in the Caribbean ...25
2.3.2 Renewable energies integration framework in the French Caribbean ...25
2.4 Theoretical specifications ...26
2.4.1 Performance ...26
2.4.2 Heat resilience ...26
2.4.3 Humidity resilience ...27
2.4.4 Cyclonic and seismic conditions resilience ...27
2.4.5 Specifications table ...27
3 ASSESSMENT OF THE TECHNICAL DESIGN CHOICES ...28
3.1 Total peak power...28
3.2 PV module choice ...28
3.3 Inverter choice ...29
3.4 PV supporting structure ...29
3.4.1 Supporting structure components ...29
3.4.2 Tilt angle optimization ...30
3.5 PV shed structure ...30
4 ECONOMIC ANALYSIS ...32
4.1 Capital Expenditures (CAPEX) ...32
4.1.1 Building permit follow up ...32
4.1.2 Surveyor study ...32
4.1.3 Soil investigation ...33
4.1.4 Foundations ...33
4.1.5 Shed structure’s EPC ...34
4.1.6 PV structure EPC ...35
4.1.7 CAPEX breakdown ...36
4.2 Operational expenditures (OPEX)...37
4.3 Economic performance analysis ...37
4.3.1 Business model definition ...37
4.3.2 Inputs ...38
4.3.3 Calculation methods/Assumptions ...39
4.3.4 Outputs ...39
5 METHODOLOGY ...40
5.1 Design inputs ...40
5.2 Environment definition ...40
5.3 Technical design ...40
5.3.1 Module choice ...40
5.3.2 Photovoltaic field design...43
5.3.3 Strings breakdown on inverters ...43
5.4 PVSyst modelling ...44
5.5 Economic Performance Analysis ...44
5.7 Methodology algorithm ...45
6 CASE STUDY ...46
6.1 Customer’s inputs ...46
6.2 Environment definition ...47
6.3 Module choice ...49
6.3.1 Module grading index calculation ...49
6.3.2 Optimal module choice ...49
6.4 Photovoltaic field design ...50
6.4.1 Module in series ...50
6.4.2 Field design ...50
6.5 Strings breakdown on inverters ...51
6.6 PVSyst Modelling ...52
6.7 Economic Performance Analysis ...53
7 METHODOLOGY RESULTS, SENSITIVITY ANALYSIS ...54
7.1 Analysis of the case study’s results for different modules ...54
7.2 Economic/ecological interest ratio sensitivity analysis ...56
7.3 Inverter choice sensitivity analysis ...57
7.4 Tilt angle sensitivity analysis ...58
7.5 Shed’s orientation sensitivity analysis...60
8 CONCLUSIONS, SUGGESTIONS FOR FURTHER WORK ...61
Appendix 1 : Economic performance analysis tool code ...62
Appendix 2 : Case study PVSYST Modelling ...66
List of Figures
Figure 2-1 Valence and conduction bands of different type of materials (Guedez, 2019) ...16
Figure 2-2 Schematic of a conventional solar cell, creation of electron-hole pair. (Gray, 2011) ...16
Figure 2-3 The different components of a PV module’s structure (ECOPROGETTI, s.d.) ...17
Figure 2-4 Typical P-V curve and I-V curve of a solar cell (PVeducation, 2019) ...17
Figure 2-5 I-V curves combination (Guedez, 2019) ...18
Figure 2-6 I-V curve and P-V curve of a PV cell operation’s simulation, for different operating temperatures (A.R Amelia, 2016) ...19
Figure 2-7 Variation of irradiance level with relative humidity (S. Mekhilef, 2012) ...20
Figure 2-8 Moisture penetration into solar cells (S. Mekhilef, 2012) ...20
Figure 2-9 Example of module delamination ( (Homepower.com, s.d.) ...21
Figure 2-10 Example of module discoloration (N.C. Park, 2013) ...21
Figure 2-11 Illustration of the PID at cell-level (Advanced Energy, 2013) ...22
Figure 2-12 Damage induced by PD activity and salt-mist at the surface of PET backsheet (Jia-wei Zhang, 2019). ...22
Figure 2-13PD events occurrences with respect to salt-mist exposure time (Jia-wei Zhang, 2019) ...22
Figure 2-14 Illustration of the PID effect at string-level (darker cells are more impacted) ...23
Figure 2-15 I-V Curve evolution for a decreasing shunt resistance (Advanced Energy, 2013) ...23
Figure 2-16 Polarization Effect on c-Si modules, mitigated by reversing the potential (black vertical line) 24 Figure 2-15 Guadeloupe electricity mix in 2017 (OREC (Observatoire Régional de l'Energie et du Climat), 2017) ...25
Figure 3-1 Aluminium anchor brackets used by Albioma ...30
Figure 3-2 Design 1 : one-sided roof shed for a 100 kWp plant ...31
Figure 3-3 Design 2 : two-sided roof shed for a 100 kWp plant ...31
Figure 4-1 Building permit follow up price distribution ...32
Figure 4-2 Surveyors price distribution (box plot) ...32
Figure 4-3 Soil investigation price distribution (box plot) ...33
Figure 4-4 Different foundation types (Dr. Mohammed E. Haque, 2014)...33
Figure 4-5 Shed Structure’s EPC price distribution ...34
Figure 4-6 Total CAPEX breakdown example for the project ...36
Figure 5-1 Relationship between the rent and different module characteristics. ...41
Figure 5-2 Module Economic index ...43
Figure 6-1 Location of the project in Martinique...46
Figure 6-2 Accurate location and orientation of the project, in the near island of la Caravelle ...47
Figure 6-3 Annual global horizontal radiation in Martinique ...47
Figure 6-4 Daily sum of global irradiation ...48
Figure 6-5 Site horizon and sun path at the project location ...48
Figure 6-6 Average air temperature at the project location ...49
Figure 6-7 Modules distribution for the case study ...51
Figure 6-8 Strings’ breakdown on the inverters ...51
Figure 6-9 Normalized production of the case study project ...52
Figure 6-10 Energy loss diagram of the project ...52
Figure 7-1 Relevance of the economic index If ...54
Figure 7-2 Relevance of the ecological index ...55
Figure 7-3 Tilt angle sensitivity analysis for two-sided roof sheds ...58
Figure 7-4 Tilt angle sensitivity analysis for one-sided roof sheds ...59
Figure 7-5 Orientation sensitivity analysis for two-sided roof sheds ...60
Figure 7-6 Orientation sensitivity analysis for one-sided roof sheds ...60
List of Tables
Table 2-1 Feed-in tariff evolution in Martinique and Guadeloupe ...26
Table 2-2 HOMER survey on power output temperature coefficient (HOMER ENERGY, 2007) ...26
Table 2-3 Specifications table ...27
Table 3-1 Modules’ characteristics (STC) ...28
Table 3-2 Inverters’ characteristics (STC) ...29
Table 4-1PV structure EPC’s costs breakdown. ...35
Table 4-2 Operational expenditures’ summarizing table ...37
Table 4-3 Given parameters of the economic analysis tool. ...38
Table 5-1 Ecological indices of the different modules ...41
Table 6-1 Case study module grading score index calculation ...49
Table 7-1 Results of the methodology for different modules ...54
Table 7-2 Grading index computed in the Ecological/Economic interest ratio sensitivity analysis ...56
Table 7-3 Inverters’ sensitivity analysis results ...57
ASA Albioma Solaire Antilles
CAPEX Capital expenditures
EPC Engineering, Procurement and Construction
EVA Ethylene Vinyl Acetate
FIT Feed-in Tariff
GHG Greenhouse Gas
IRR Internal Rate of Return
I-V curve Intensity of the current – Voltage Curve
LCOE Levelized Cost of Electricity
MPP Maximum Power Point
MPPT Maximum Power Point Tracker
NPV Net Present Value
OPEX Operational expenditures
PD Partial Discharge
PID Potential Induced Degradation
PR Performance Ratio
P-V curve Power – Voltage Curve
RH Relative Humidity
ZNI Non-Interconnected Zones
Symbol Description Units
Short-circuit current temperature coefficient %
Power output temperature coefficient %
Efficiency of the module at year n %
Open-circuit voltage temperature coefficient % a, b, c Linear coefficients of the rent as a function of Pm
Module Degradation rate %
Module Degradation start (corresponds to the module efficiency at the first year)
I Module grading index
Module ecological index
Non-standardized module economic index €
Module economic index
Current Intensity at MPP A
Short-circuit current A
Maximum number of modules in series per array
Minimum number of modules in series per array Msopt
Optimal number of modules in series per array Nm
Total number of modules in the project
Unit Price of the module €/Wc
Power at MPP W
Customer ecological interest ratio rf
Customer economic interest ratio
Voltage at MPP W
Open circuit voltage V
1 INTRODUCTION 1.1 Background
As electricity and heat production account for more than 40% of the global total CO2 emissions (IEA, 2019), increasing the share of renewable energy sources in their electricity mix is a consistent and inevitable path that countries must follow if they want to fulfill their climate mitigation pledges. This statement is all the more so true when considering the case of independent isolated grids which usually rely heavily on fossil fuel importation.
Martinique, a French oversea island located in the Caribbean sea, faces this issue (energy dependency rate higher than 90% ) (French government, 2019); and aims to develop significantly photovoltaic (PV) electricity production, multiplying by 2.5 the total solar power production (French government, 2019) to reach a PV production share of 17% in 2023. On an island, space is limited, which hinders the development of large- scale ground PV plants. Therefore, it becomes essential to make the most out of roof occupation: most of the PV potential of the island lies in small-scale (< 1 MW) roof-integrated PV projects.
However, since modules are facing a harsh environment, the PV project configuration cannot follow the same rules as in a “classical” PV project: the project specifications and design techniques are different as more constraints must be taken into consideration. Therefore, designing PV configurations for such projects requires a specific methodology adapted to the tropical and cyclonic conditions.
1.1.2 Albioma Solaire Antilles presentation
Albioma is an independent French energy producer with most of its plants located in French overseas territories. Though it launched its activities in 1982 by developing coal power plants, Albioma shifted severely its strategy this past 20 years to embrace energy transition by replacing coal by biomass in their thermal plants and increasing their solar capacity. With this new strategy, they doubled the renewables share in their electricity mix (from 36 % to 62 %) between 2013 and 2018 and aim to exceed 80% by 2023.
Solar projects development in the Caribbean zone is covered by its subsidiary Albioma Solaire Antilles (ASA), the company that hosts this master thesis. ASA is a PV project developer which designs, owns and operates its power plants ; among all the steps of a PV project development, the only ones that are subcontracted are the construction works : in this case, Albioma supplies the subcontractor with all the different components and the technical drawings. Albioma has a strong local footprint in the French Caribbean and foster local companies through the construction phase. Though ASA develops large scale solar power plant (>1MW), most of its business lies in small-scale roof integrated projects, specifically existing roofs where Albioma offers an annual fee in exchange of the use of the roof to operate a PV plant.
Nevertheless, as the competition is tough and the market of existing roofs limited, ASA would like to develop an offer for small-scale PV integrated structure, such as PV sheds, adapted to cyclonic and tropical conditions. In such projects, a specific remuneration scheme must be designed as ASA can either bear the construction costs or give an annual fee to the customer. Since ASA has no record for this kind of project, they are willing to design their offer with an innovative methodology, in order to be sure that it fits perfectly the local conditions and constraints, and the costumer’s requirements.
The overarching aim of this study is to design a suitable methodology for identifying optimum PV
configurations, for small-scale modular solutions such as PV sheds, for tropical and cyclonic locations. This
methodology must ensure that an offer is adapted from a technical point of view, and is also consistent
from an economic point of view. This methodology needs to consider all the factors that are typical of the
Caribbean islands. Given this global approach, the following intermediate objectives will be targeted:
• Investigate the effect of tropical conditions on PV panels.
• Investigate the structural needs for a PV structure to withstand cyclonic conditions.
• Develop a set of specifications consistent with the local environment and local regulations.
• Investigate and Analyze the technical design of small-scale modular solutions in such conditions.
• Investigate and Analyze the economic structure of small-scale modular solutions.
• Develop a methodology that permits to design the PV configuration of such projects.
• Perform this methodology on a case-study and analyze the sensitivities of the result.
The expected deliverable of this master thesis is an innovative methodology to design small-scale modular PV projects in the Caribbean islands, tested on a case-study and based on a real need for Albioma Solaire Antilles as they are committed to develop such projects for their customers.
1.3 General methodology
In order to carry this master thesis work, a scientific approach will be chosen, and the following steps will be undertaken:
1. Literature review on PV fundamentals, with a special focus on temperature and humidity impacts on PV performances.
2. Solar development background in the Caribbean.
3. Choice of the specifications for the PV structure.
4. Assessment of the technical design choices of a PV structure for small-scale modular PV projects under tropical conditions.
5. Cost study assessment of the different expenditures of these projects.
6. Business models identification, fitting the needs of the local customers.
7. Development of an innovative methodology to find the best PV configuration for such projects, using personalized indicators.
8. Results of the methodology applied to the case study: Definition of the case study environment, modelling of the PV configuration obtained.
9. Sensitivity/ optimization analysis of the project.
The data collection for the case study environment’s definition will be extracted from SOLARGIS, a reliable and detailed database used by Albioma Solaire Antilles. The graphic design of the PV structure will be performed through Sketch’Up. The technical performance modelling will be performed with PVSyst, and the economic analysis will be performed using MATLAB and Excel.
As every thesis work, the different steps and requirements of the methodology were specified during the literature review, it revealed that one main leverage to design a modular solution was the module choice.
The fact that the solution must also be adapted to the customer’s expectations brought the idea of grading indices.
1.4 Scope of work
As designing a PV project offers endless possibilities, the methodology developed will be limited to these aspects of the project’s design:
• Module choice.
• Shed structure choice.
• String and PV field design.
• Financial offer to the customer.
Therefore, some design aspects will not be considered, such as:
• The total peak power will be equal to 100 kWc, this choice is justified in the paragraph 3.1.
• Inverter choice (which is supposed to be limited to the different designs of the HUAWEI SUN2000), will not be part of the methodology for the moment., the basic inverter design will be 2 SUN2000-36KTL and 1 SUN2000-20KTL
• Tilt angle optimization.
• The PV supporting structure choice.
• Earthquake-resistant design.
Nevertheless, sensitivities analysis will be conducted in order to assess the impact of such aspects on the plant performance, to identify which enhancements could be made in a future version of the methodology.
1.5 Reading instructions
References appears in the main report as (Last name, Year), following the Harvard method. These are collected in a bibliography at the end of the report.
Figures and tables are numbered according to the chapter in which they occur. Therefore, the first figure in chapter 3 has number 3-1, the second figure has number 3-2 and so on. These are listed at the beginning of the report. Describing text for figures and tables is placed beneath the given figures and tables and the reference is also specified. If the reference is not specified, then the figures and tables are made by the author itself.
Equations are specified by a number in a bracket and they are numbered by chronological order. Therefore, the first equation has number (1), the second equation has number (2), no matter in which chapter they occur in.
Abbreviations and Symbols are also listed at the beginning of the report. They are always specified the first
time they occur in the report.
2 THEORETICAL FRAMEWORK AND SPECIFICATIONS
This section aims to provide the necessary theoretical notions on solar power while focusing on the ones that concern tropical and cyclonic conditions, it is not meant to be exhaustive.
2.1 PV Fundamentals
2.1.1 PV effect
Photovoltaic cells are a rather simple device, they have the capacity to absorb light from the sun and to transmit a portion of the energy absorbed to carriers of electrical current: electrons and holes. This capacity is based on the electronic properties of semiconductors materials, as shown in the figure 2-1, they possess an energy gap between their valence band and conduction band.
When a light photon of sufficient energy encounters a semiconductor, a valence band electron can absorb part of its energy to reach the conduction band, leaving a hole in the valence band. Using two opposite types of semiconductors: N-type (with increased electron carrier concentration) and P-type (with increased hole carrier concentration), one can create a PN junction, which is the basic conventional solar cell of which the structure is given in the figure 2-2 :
Figure 2-1 Valence and conduction bands of different type of materials (Guedez, 2019)
Figure 2-2 Schematic of a conventional solar cell, creation of electron-hole pair. (Gray, 2011)
-17- 2.1.2 PV Module structure
A module consists in several strings of PV cells in parallel (depending on the desired power, voltage and current characteristics). These strings are vacuum encapsulated together, usually using an encapsulant material called EVA (Ethylene Vinyl Acetate), this process, called lamination is critical in PV modules’
manufacturing as the encapsulant quality affects the light transmission. The module is protected from weather and moisture by the back-sheet at the back and by the front glass on top. This latter can be treated with an antireflective layer to increase the module efficiency. All these parts are assembled in the frame, usually made of aluminum. Finally, the junction box contains the protection
Figure 2-3 The different components of a PV module’s structure (ECOPROGETTI, s.d.)
2.1.3 PV cell/module characteristics
PV cells operation can be described with the Power-Voltage Curve (P-V curve) and Intensity-Voltage Curve (I-V Curve). They display the power/intensity of the current through the cell with respect to the voltage.
Figure 2-4 Typical P-V curve and I-V curve of a solar cell (PVeducation, 2019)
-18- The key operating data are:
• The short circuit current ISC
when there is no voltage.
• The open circuit voltage VOC
when there is no current.
• The current IMP
, voltage VMP
, and power PMP
at Maximum Power Point (MPP), which is the operating point to aim for in order to achieve optimal performance of the module.
As highlighted in the graph below, adding cells in series will add their voltage, adding them in parallel will add their currents.
2.1.4 Maximum Power Point Tracker (MPPT)
In order to drive the operating point of a module or a string towards the MPP, PV arrays are often connected to maximum power point trackers. These devices are small DC-DC converters, they utilize different types of control circuit or logic to search for the MPP and modify accordingly the impedance seen by the panel which determines the operating point of the solar panel.
It is critical to connect only similar panels to the same MPPT : if two panels are not on the same side of the roof, they won’t have the same operating I-V curve, so their MPP won’t be similar, so the MPPT won’t be able to achieve the most efficient operation.
Figure 2-5 I-V curves combination (Guedez, 2019)
2.2 Temperature and Humidity impact on PV performance
2.2.1 Temperature effect on PV performance
Temperature influences the energy carriers’ behavior which leads to a shorter bandgap for higher temperatures. This dependency is well modeled by the following equation (where𝛼 and 𝛽
are constant specific to each semiconductor) :
(𝑇) = 𝐸𝐺
The bandgap directly influences the open circuit voltage VOC
, which thus varies roughly linearly with temperature. As current increases only slightly with temperature, a temperature increase will lead to a power output decrease. This linear relationship is usually given in PV modules specification through the short circuit temperature coefficient 𝛼𝐼𝑆𝐶
and the open circuit voltage temperature coefficient 𝜇𝑉𝑂𝐶
being the cell temperature):
[ 1 − 𝛼𝐼𝑆𝐶
)] 𝑎𝑛𝑑 𝑉𝑂𝐶@𝑇𝐶
[1 − 𝜇𝑉𝑂𝐶
Usually, these two coefficients are combined in the power output temperature coefficient, 𝛼𝑃
[ 1 − 𝛼𝑃
This dependency is precisely discussed in (A.R Amelia, 2016), the graphs below display the I-V curve and the P-V curve simulations of a module for different operating temperatures :
These figures observed that an increase in 10 °C of temperature causes a decrease of the output power about 5W or 5%. As high solar irradiance usually comes with higher ground temperature, temperature regulation in PV projects in tropical regions is critical and will be considered during the design of our solution, therefore consistent values for the temperature coefficients will be chosen.
Figure 2-6 I-V curve and P-V curve of a PV cell operation’s simulation, for different operating temperatures (A.R Amelia, 2016)
-20- 2.2.2 Humidity effect on PV performance
Humidity affects PV cells in two different ways: it affects the irradiance level of sunlight with water vapor particles and it can enter the solar cell enclosure. In the first scenario, water vapor will cause a non-uniform distribution, which will affect light diffraction, as the figure 2-6 displays, as the relative humidity (RH) exceeds 20%, the irradiance drops significantly. However, no technological solution can counteract this phenomenon as it directly influences the sun irradiance.
In the second approach, humidity can foster moisture penetration. It increases the risks of cell cracking and corrosion at cell interconnection, two of the main failure mechanisms of PV cells. In order to hinder this phenomenon, edge sealants, or low ionic conductive material for encapsulant can be used.
Figure 2-8 Moisture penetration into solar cells (S. Mekhilef, 2012) Figure 2-7 Variation of irradiance level with relative humidity (S. Mekhilef, 2012)
The issue is not moisture itself, but the phenomena it causes. In fact, studies have stated that it can induce discoloration and delamination of the cell’s encapsulant. (N.C. Park, 2013), as presented in the figure below.
Figure 2-9 Example of module delamination ( (Homepower.com, s.d.)
Figure 2-10 Example of module discoloration (N.C. Park, 2013)
These phenomena decrease the reflectance et transmittance of the module, which causes a decrease in the ISC
, thus reducing the module performance. They also affect the way module degradation must be computed.
In this thesis work, degradation is computed using a classical linear model, but it could be interesting to
consider other models which take into account these potential phenomena.
-22- 2.2.3 Potential Induced Degradation
Potential Induced Degradation (PID) is an undesirable effect that concerns some PV modules, it occurs when the module’s voltage potential and leakage current foster ion movement within the module structure itself : between the semiconductor material and the other components of the module ( frame, mount, glass, anti-reflective coating…) which causes the module’s power output capacity to degrade. Such an effect can be mitigated and even reversed if treated properly.
Figure 2-11 Illustration of the PID at cell-level (Advanced Energy, 2013)
As ion mobility accelerates with humidity, temperature and voltage potential, several factors affect PID:
• Environmental factors: an increase of temperature and/or relative humidity intensifies PID as it fosters ion mobility. Moreover, conditions on coastal areas can also spur PID, as highlighted in the image and the graph below which underline that Partial Discharge (PD), a phenomenon induced by PID, increases with salt-mist exposure (Jia-wei Zhang, 2019).
Figure 2-13PD events occurrences with respect to salt-mist exposure time (Jia-wei Zhang, 2019) Figure 2-12 Damage induced by PD activity and salt-mist at the surface of PET backsheet (Jia-wei Zhang, 2019).
• System factors: The module’s voltage potential and sign have a significant impact on PID.
Therefore, it depends on the system grounding topology (symmetric/positive/negative grounded) and the module’s position in the array. Indeed, modules with a high negative potential with respect to earth are more likely to develop PID. Then, as shown below in the illustration, the degradation is stronger in panels that are closer to the negative side of the string.
Figure 2-14 Illustration of the PID effect at string-level (darker cells are more impacted)
• Module and factors: The choice of the different components of a PV module: glass, encapsulation, frame… has an impact on PID. For instance, encapsulants that have superior moisture permeability properties compared with EVA, the most common encapsulant, have been shown to decrease susceptibility to PID. Also, the anti-reflective coating is considered a causative factor in PID.
This effect leads to a reduction in shunt resistance which drives down the module’s MPP, as highlighted below in the simulation. It can be noticed through unexplainable yield losses, or high voltage difference between modules at different string positions.
Figure 2-15 I-V Curve evolution for a decreasing shunt resistance (Advanced Energy, 2013)
The PID effect can be either irreversible, when electrochemical reactions lead to electro-corrosion and/or
film delamination in the modules, or reversible, then called the Polarization Effect. Indeed, studies have
found that the Polarization Effect can be mitigated by reversing the polarity of the module. This process is
still under test but might be implemented in the next generation of PV modules.
Figure 2-16 Polarization Effect on c-Si modules, mitigated by reversing the potential (black vertical line)
As PID can affect significantly the power output of PV modules, especially in warm and humid
environment, the designed methodology must take the PID-resistance of the modules in its criteria,
especially on coastal areas.
2.3 Solar development background in the Caribbean.
2.3.1 Energy framework in the Caribbean
Energy project development faces mostly island-related constraints in the Caribbean:
• Energy dependency, with the exception of Trinidad and Tobago, the islands are net importers of hydrocarbons for electricity generation. (Charles, 2012). In Martinique and Guadeloupe, the two French Caribbean islands, the energy dependency was respectively 94% and 95% in 2016.
(OREC (Observatoire Régional de l'Energie et du Climat), 2017)
• Moreover, as islands are usually isolated, companies must consider higher shipping and delivery costs. For specialized work, labor cost is also increased as foreign experts must be brought on site.
• Land use is also a key constraint in these regions. Indeed, islands must manage carefully the way they handle land availability, and often, large space-requiring projects are not accepted. As an example, Large scale ground based solar projects are not accepted in the French Caribbean if they are not located on a degraded ground.
• Soil quality could also be problematic on islands, however, as the Caribbean are volcano-formed islands, the soil is usually good at a low depth.
Besides, despite high solar and wind potential, the electricity mix in the Caribbean usually relies heavily on fossil fuels. For instance, above is given the electricity mix of Guadeloupe in 2017. The black and brown parts represent respectively coal and oil while the other colored parts represent renewables sources. This graph highlights that more than three quarters of the electricity generation is fossil-fuel powered.
Islands of the Caribbean will probably be among the most impacted territories regarding climate change, they will be affected through beach erosion, depletion of fishery resources, tropical storms and floods increase… (Charles, 2012). Combined with high oil prices, this should act as a strong motivation for the Caribbean to spur their energy transition.
2.3.2 Renewable energies integration framework in the French Caribbean The French Caribbean Islands (Martinique and Guadeloupe) are part of the Non-Interconnected Zones (ZNI), which follow a different grid policy from the national French grid. Indeed, their renewables’
integration objectives are more ambitious (Comission de Régulation de l'Energie, 2018) :
• to reach a 50% share of renewable energy in their energy mix by 2020
• to reach energy independence by 2030.
Considering these two ambitious milestones, solar energy development is based on two different schemes:
public tenders for installations of which total peak power is higher than 100 kWc and feed-in tariff (FIT).
(Légifrance, 2017) for smaller installations. Given this master thesis context, feed-in tariff fits best to the project characteristics, firstly as it applies to all project which fulfills the conditions without any selection;
secondly given the use of the PV project (a solar shed), it is more likely to be smaller than 100 kWc.
Therefore, the size of the PV plant will be limited to 100 kW.
Figure 2-17 Guadeloupe electricity mix in 2017 (OREC (Observatoire Régional de l'Energie et du Climat), 2017)
Total peak power group Feed-in Tariff (ct€/kWh)
Q4 2018 Q1 2019 Q2 2019 Q3 2019
≤ 3 kWc 22.34 22.34 22.19 21.84
> 3 kWc and ≤ 9kWc 19.86 19.85 19.73 19.42
> 9 kWc and ≤ 36 kWc 18.2 18.2 18.08 17.8
> 36 kWc and ≤ 100 kWc 16.55 16.54 16.44 16.18
Table 2-1 Feed-in tariff evolution in Martinique and Guadeloupe
The FIT is degressive with respect to the total peak power of the project, as displayed in the table above.
Moreover, for a given project, this FIT is fixed for 20 years as soon as the grid connection demand is sent to the grid operator.
2.4 Theoretical specifications
The methodology that is developed in this work must be defined considering a set of specifications. These latter will permit to preselect the different technical components before refining the technical choices with the methodology. Therefore, the PV module, the inverter, the modules’ supporting structure and the whole system’s structure must comply with the following specifications in terms of corrosion, heat resilience, or stress resistance.
The system has to be efficient and reliable, so the following minimum values, which correspond to efficient inverter and module (Baumgartner, 2017) will be selected : for the inverter, an efficiency greater or equal to 97%, calculated following the European norms, and an efficiency greater or equal to 18% for modules.
Moreover, in order to avoid a complete shutdown of the plant in the case of an inverter failure, Albioma chose to design its plants with at least 2 different inverters, so this project will fulfill this requirement.
2.4.2 Heat resilience
Resilience to hot temperature will affect two design choices: Firstly, the PV supporting structure will need to allow air flow under the panels in order to maximize heat dissipation. Secondly, the module chosen must show good performance with temperature increase. HOMER energy, a microgrid modelling and optimization software, realized a quick survey to compare power output temperature coefficients for different module types which results are shown below.
Table 2-2 HOMER survey on power output temperature coefficient (HOMER ENERGY, 2007)
The average power coefficient is around 0.40%/°C; therefore, the chosen module must not exceed
-27- 2.4.3 Humidity resilience
Similarly, humidity resilience orientates the choices for both the supporting structure and the electric devices (modules and inverters). Firstly, the structure’s material must be stainless or galvanized steel which guarantees that the material can withstand high relative humidity. Indeed, stainless steel contains chromium which prevents it from corroding, and galvanized steel is commonly coasted with layers of zinc which enhances the corrosion withstanding ability. Secondly, all electric components must meet the IP65 requirements: being completely protected against dust and water jets, from any directions.
2.4.4 Cyclonic and seismic conditions resilience
The Caribbean is famous for being one of the most cyclone-prone area, given these harsh conditions, all building must be able to withstand strong winds. In the French Caribbean, the corresponding norm specifies the minimum pull-out loads that a construction must withstand: 1 200 Pa in normal conditions, and 2 100 Pa in extreme conditions. (QUISTIN, 2019) In order to comply with this norm, the chosen module must be able to withstand a load superior or equal to the extreme conditions limit : 2 100 Pa.
As the Caribbean host a powerful volcano activity, through the volcano arc of the West Indies, seismic measures and norms can apply to construction works. However, in the case of an agricultural shed, no earthquake-resistant specifications apply in the French Caribbean islands, therefore this will not be considered in our specifications.
2.4.5 Specifications table
Stainless or galvanized steel IP 65
Allowing air flow under the PV modules
Power output temperature coefficient <
Maximum load > 2 100 Pa Efficiency > 18 %
Efficiency > 97 % IP65
Minimum 2 inverters
Table 2-3 Specifications table
3 ASSESSMENT OF THE TECHNICAL DESIGN CHOICES 3.1 Total peak power
This main characteristic will affect all the technical design choices. As most energy production projects, solar plants’ profitability increases with their size, because some investment costs are fixed, regardless of the total peak power, while the revenues are an increasing function of the capacity. Therefore, considering paragraph 2.3.2, the most suitable total peak power is 100 kWc, as it is the maximum power for which the FIT exists. It will be the total peak power of the designed project.
3.2 PV module choice
Given the harsh conditions the module will be facing, it is important to choose a reliable and resistant one, though the cost is likely to be higher. Albioma Solaire Antilles has been working with several different brands and technologies: First Solar with Cadmium telluride modules, Yingli solar, General Electric …. For the upcoming projects, they have selected five high-end manufacturers: REC, Longi Solar, Sunpower, Qcell and Voltec as they offer modules that meet the specifications decided in 2.4.5. These companies have a really strong warranty policy which underlines the reliability of their modules. Their modules’ characteristics are displayed in the table below:
Q-Peak DUO G5 330
SunPower E20-435- COM
TARKA 60 VSMS Characteristics Unit 300
Longi Solar, China
Nominal Power Wp 360 310 330 435 300
Efficiency % 21.7 19.0 19.6 20.1 18.0
VOC V 45.5 40.3 40.66 85.6 33.00
ISC A 10.26 9.98 8.22 6.43 9.09
%/K -0.26 - 0.37 - 0.37 - 0.35 - 0.375
Product warranty 20 years 10 years 12 years 25 years 20 years Power output Warranty 25 years 25 years 25 years 25 years 25 years
Degradation rate (Dr) %/year 0.50 0.55 0.54 0.4 0.7
Degradation start (Ds) % 98.5 98 98 98 96
Maximum load (push) Pa 7000 5400 5400 5400 5400
Maximum load (pull) Pa 4000 2400 4000 2400 2400
PID resistance Not
Solid PID resistance
Anti PID Technology
PID tested Not specified
Certification IP67 IP67 IP65 IP65 IP65
Carbon footprint gCO2eq/Wc 400 450 450 359 300
Price c€/Wc 29 27 34 35.5 30
Table 3-1 Modules’ characteristics (STC)
The choice of the module will be the main objective of the developed methodology. Indeed, given the requirements of the project, different modules can be favored. For instance, if the performance prevails in the project, the efficiency and power output warranty will be the decisive factors. On the other side, if the environmental impact of the project prevails, then the carbon footprint will be the decisive factor.
3.3 Inverter choice
As for the modules’ choice, choosing an optimal inverter out of the endless possibilities of brands and designs requires a preselection. ASA has chosen HUAWEI SUN2000 inverters for their reliability but also their multiple MPPTs which permit to optimize the production. The main advantage of such configuration is that, thanks to the several input connections, no junction box is needed to connect the strings in parallel which highly simplifies the design of the plant : the only optimization to conduct is the one on string voltage to assess the number of modules in series. The main characteristics of these inverters are listed below:
HUAWEI SUN2000 design 10KTL 20KTL 36KTL 60KTL
Nominal Power kVA 10 20 36 60
Maximum input power (cosφ = 1)
kW 11 22 40 66
Efficiency % 98.0 98.3 98.6 98.7
Operating voltage interval
V 200-1000 200-1000 200-1100 200-1100
Nominal input voltage V 620 620 620 600
Maximum input current per MPPT
A 18 18 22 22
Number of MPPT 2 3 4 6
Number of inputs connections
4 6 8 12
Warranty 5 years 5 years 5 years 5 years
Certification IP65 IP65 IP65 IP65
Price € 1000 1400 2050 3180
Table 3-2 Inverters’ characteristics (STC)
For the moment, inverter choice’s optimization will not be part of the methodology. The basic inverter design will be the one chosen by ASA, which is 2 SUN2000-36KTL and 1 SUN2000-20KTL. A sensitivity analysis, computing the performance of one project with different inverter designs could be conducted later in order to support this choice.
3.4 PV supporting structure
3.4.1 Supporting structure components
In the case of a solar shed, for the sake of simplicity, solar panels will be mounted parallel to the roof,
therefore the mounting structure design will be rather simple. ASA have been using the same supporting
structure for the past two years, supplied by EXTOL, a Spanish aluminium profile manufacturer. It allows
the air to flow as there is a 10 cm gap between the roof and the panels. This structure will be the one chosen for the project; it consists in:
• an anchor bracket screwed directly on the roof purlins.
• Aluminium rails screwed on the anchor brackets.
• Panel-fasteners which bonds the solar panels to the rails.
Figure 3-1 Aluminium anchor brackets used by Albioma
3.4.2 Tilt angle optimization
Since panels are parallel to the roof, the panels’ tilt angle is equal to the roof tilt angle. Therefore, this parameter is chosen when designing the shed structure.
3.5 PV shed structure
In order to minimize the shed structure costs, two standardized designs are chosen (designs below). The first design is a one-sided roof, more economical, while the second design is a more classical two-sided roof design, slightly more expensive but which better sheds the sun. It is important to offer also this latter type of design since this type of project is usually used as farm sheds.
For both selected designs, the tilt angle of the roof will be 8.5°. It is relevant to notice that this tilt angle is
not equal to the latitude (between 14-16°), which is supposed to be approximately equal to the optimal tilt
angle for a single-sided roof, therefore the project’s performance could be enhanced by choosing a shed
structure with an optimized tilt angle, the impact of such measure could be studied in the sensitivity analysis.
Figure 3-2 Design 1 : one-sided roof shed for a 100 kWp plant
Figure 3-3 Design 2 : two-sided roof shed for a 100 kWp plant
4 ECONOMIC ANALYSIS
4.1 Capital Expenditures (CAPEX)
The first step to undertake this economic analysis is to list all the different expenditures of the project. A chronologic approach is chosen to underline the different steps. ASA has made the choice to design, build, operate and possess its plants, therefore they do not call in any EPC (Engineering, Procurement and Construction) company to manage the whole construction process, they would rather subcontract key services to specialized companies while managing the whole project.
To start with, the project begins with a lot of administrative procedures: on the one hand, it is necessary to contract an architect to the building permit of the structure. On the other hand, a surveyor must also be contracted to create a new parcel for the project, this will be necessary to sign a long-term leasehold which only concerns the newly built structure.
Simultaneously, a soil investigation is required, as this will significantly impact the foundation design.
Foundations come next in the project’s development and represent a non-negligible expenditure.
Lastly, key expenditure items are the EPC of the shed structure, and the EPC of the PV structure.
Several providers were contacted in order to assess the cost of each expenditure. This work is detailed in the following part, for every item, a price distribution (displayed through a box chart) is given. For the economic analysis of the project, the selected price will always be the first quartile price, in order to optimize the project’s costs while keeping conservative values.
4.1.1 Building permit follow up
A milestone of the project is the procurement of the building permit by the municipality which is essential before sending the grid connection demand or starting any construction work. Such permits can easily be refused if the claim is not perfectly justified, thus they require a certain knowledge of urbanism and specific paperwork, therefore it is advised, though not absolutely necessary, to contract an architect to undertake the procedure.
In ASA’s case, though these competencies must be internalized in the end, the choice of contracting an architect for the first project of this offer has been made, therefore a cost study has been undertaken. For the economic analysis of the project, the selected price is 4,000 €
4.1.2 Surveyor study
In order to operate the solar plant, ASA needs to sign a long term leasehold with the landlord that grants it the right to use the roof for the plant operation. As the building is to be constructed, the cadastral plan must be modified to create a specific parcel for the shed only. Such procedure can only be undertaken by an expert surveyor.
As the box plot on the left underlines, the price of surveyors’
study for this kind of work is rather fixed. For the economic analysis of the project, the selected price is 4,400 €
Figure 4-2 Surveyors price distribution (box plot) Figure 4-1 Building permit follow up price distribution
-33- 4.1.3 Soil investigation
The soil test will determine the properties of the soil on the site. It will describe the reactivity of the ground: whether it expands or contracts considerably: and the compaction rates: how dense it is. It will permit to design the foundations which can have a significant effect on the total expenditures as foundations’ costs vary a lot given the foundation type.
For the economic analysis of the project, the selected price is 2,900 €.
There are several types of foundations. As described in the scheme below, they depend on the depth of the good (solid or dense) soil.
Figure 4-4 Different foundation types (Dr. Mohammed E. Haque, 2014)
In the case of shallow foundations, which is the simpler and thus more cost-effective case, the basic component is called a footing, it is a structural element fixed in the soil (usually in concrete) which transfers loads to the soil from columns, walls or lateral loads from earth retaining structures, depending on the structure, different types of footings can be used :
• Isolated spread footings, under individual columns. Usually rectangular, they are the most common type of foundations used.
• Wall footing is a continuous slab strip along the length of wall. It can be a spread wall footing if it is wider than the wall, to transfer a higher load.
• Combined footings and Cantilever or strap footings are used when two columns can share one footing.
Figure 4-3 Soil investigation price distribution (box plot)
• Finally, raft or mat foundations consist in a large continuous footing supporting all the columns of the structure. This is used when soil conditions are poor, but piles are not used.
Deep foundations are used when the soil bearing capacity near the surface is too low. The load is transferred through end bearing (pier foundations) but also skin friction (pile foundations).
Knowing the exact type of foundations adapted to our project requires a soil investigation. However, in order to assess the foundations’ cost in the methodology, it will be considered that, given its environment, a project has either a rather good soil where shallow foundations are sufficient, or it has a bad soil, which requires deep foundations. The assumed costs for shallow foundations will be 10,000 € while the cost for deep foundations will be 20,000€.
4.1.5 Shed structure’s EPC
The shed structure’s EPC is one of the biggest expenses of the project, it consists in 3 main works:
• Engineering: sizing and design of the shed structure, taking into account the PV structure loads and the environment related specifications, supply of the assembly drawings.
• Procurement: supply of the shed structure: metallic frame, roof cover and wall cover if needed.
• Construction: assembly of the shed structure
The price range is quite broad depending on how personalized the work must be. Indeed, one can either choose a shed “off-the-shelf” (with engineering cost included in the procurement) which will reduce significantly the total cost, or more personalized products (engineering plans undertaken by a design office…).
Also, the 3 different parts of the EPC can be appointed to different companies as they require different skills. Therefore, the price distribution has been given for each part of the EPC in order to compare companies that only offer one service to companies that offer a complete EPC package. These prices concern two types of shed (one or two-sided) as it is a technical choice offered to the customer.
Figure 4-5 Shed Structure’s EPC price distribution
For the economic analysis, the selected costs will be for the two-sided roof : 4,600 € for the engineering,
38,000 € for the procurement and 36,000 € for the construction, and for the one-sided roof : 4,400 € for
the engineering, 31,000 € for the procurement and 28,000 € for the construction.
-35- 4.1.6 PV structure EPC
Thanks to ASA’s experience in the photovoltaic business, these costs have been identified precisely; using the data gathered in the different projects of ASA, the following table displays the calculations’ assumptions and the total cost per item. The costs in italic (PV panels and Grid connection) are assumed in this table to calculate the total cost but can differ for every project.
Item Assumptions Total cost for a
100kWp project Components procurement
Depends on the panel,assumed 0.30 €/Wc , 100 kWc 30,000 €
Inverters 1400 €/u for SUN2000-20KTL : 1 unit
2050 €/u for SUN2000-36KTL : 2 units 5,500 €
PV supporting structure 0.06 €/Wc , 100 kWc 6,000 €
components 0.07 €/Wc , 100 kWc 7,000 €
Shipment 0.05 €/Wc , 100 kWc 5,000 €
Construction works PV structure
construction 0.32 €/Wc , 100 kWc 32,000 €
Depends on the project location,assumed 15,000 €
for a 100 kWc installation 15,000 €
Engineering 6,000 € for 100 kWc 6,000 €
Bank fees Approximately 2% of the total project's cost 4,000 €
Insurance 70 €/kWc, 100 kWc 7,000 €
Miscellaneous 5,000 € per project 5,000 €
TOTAL 122,500 €
Table 4-1PV structure EPC’s costs breakdown.
-36- 4.1.7 CAPEX breakdown
The different data presented above are summarized in the following pie chart to display an example of the total breakdown of CAPEX for a 100 kWp project (with two-sided shed’s structure and shallow foundations).
Figure 4-6 Total CAPEX breakdown example for the project
As one can see, the PV structure-related costs (blue part) account for more than half of the total CAPEX while the shed EPC (red part) account for approximately one third of it. It is also interesting to underline that the cost of the core components of the PV structure : the modules, represents less than 15% of the CAPEX, though it is a significant share, it confirms that a difference in the module’s unit price of 1 or 2 cts per Wc is acceptable and won’t skyrocket the total CAPEX.
2% Surveyor 2%
Soil investigation 1%
Shed - Engineering 2%
Shed - Procurement 17%
Shed - Construction 16%
PV - Modules 14%
PV - Inverters 2%
PV - Other procurement
PV - Construction 21%
PV - Project Development