Design, Greenhouse Emissions, and Environmental Payback of a Photovoltaic Solar Energy System

Abstract

This study aims to design a 16.4 MW photovoltaic solar system located in the Brazilian Northeast and quantify the associated greenhouse gas emissions and environmental payback. The energy system was designed to minimize the Levelized Cost of Energy. The greenhouse gas emissions were quantified with the Life Cycle Assessment methodology, expressing the environmental impact in terms of generated energy (kg CO₂-eq/kWh) and following ISO 14040 and 14044. The environmental payback considered the Brazilian electricity mix and degradation of the panels. The results indicated a system capable of producing 521,443 MWh in 25 years, with an emission factor of 0.044 kg CO₂-eq/kWh and environmental payback of five years and eight months. The emission factor is at least ten times lower than thermoelectric natural gas power plants. The solar panels were the main contributors to the greenhouse gas emissions, representing 90.59% of overall emissions. It is concluded that photovoltaic energy systems are crucial in the search for emissions mitigation, even in a country that presents a predominantly renewable electricity matrix, with demonstrated environmental benefits.

1. Introduction

Global warming and climate change are current and growing concerns. Fast economic development and technological progress have increased energy demands worldwide [1]. Traditionally, fossil fuels have been employed for energy purposes since the beginning of the industrial revolution, being the driving force of the industrialized world, contributing in 2019 to more than 81% of world energy production [2]. Technological development and the progressive growth of energy demands culminated in a significant increase in greenhouse gas (GHG) emissions, which can lead to adverse effects such as changes in rainfall, snow, and ice cover patterns and sea level rise [3]. The concentration of carbon dioxide (CO₂) in the atmosphere increases due to the combustion of fossil fuels, enhancing the natural greenhouse effect and global warming [4].

Mitigating these ever-increasing GHG emissions can encompass low carbon technologies, renewable energy resources, and energy efficiency strategies. Growing awareness of the importance of environmental protection and the potential environmental impacts associated with products and services have been in the research spotlight, attracting interest to develop methods to understand better and address or mitigate these environmental impacts.

The Brazilian electricity matrix is predominantly comprised of renewable energy resources, with hydroelectricity contributing 63.81% to the overall generation mix in 2020 [5]. Hydroelectricity is directly related to the levels of reservoirs and the incidence of rainfall. Brazil suffered a water crisis between 2013 and 2015 [6], leading to an increase in fossil fuel-based thermoelectric generation to meet the dispatchable demand. However, even after the hydric reservoirs returned to their original levels, fossil fuel-based thermoelectric generators were maintained operational to supply non-dispatchable demands. Natural gas thermoelectric generation increased 10.7% from 2018 to 2019 [7] because increases in energy demands occur at higher rates than the expansion of hydroelectric generation capacity (for the same period, the increase in hydroelectric generation was only 2.3%).

Renewable energy generation systems, such as solar and wind, require specific conditions for generation (solar radiation or wind) and cannot be directly manipulated by man, so they are considered non-dispatchable systems. Renewable energies can supply a non-dispatchable portion and contribute to reducing greenhouse gas (GHG) emissions in the process of burning fossil fuels. Renewable generation systems can save the levels of hydroelectric reservoirs by supplying energy to the grid during their favorable generation period. With higher reservoir levels, the need to operate fossil fuel-based thermoelectric power plants (TEPPs) is reduced, thereby reducing GHG emissions from fuel burning.

Life Cycle Assessment (LCA) can be employed to compare the GHG emission levels of these power generation systems. LCA comprises the inventory of relevant inputs and outputs, the assessment of possible environmental impacts, and the interpretation of results [8]. Furthermore, LCA is often used to analyze renewable energy alternatives to conventional energy systems, especially for estimating GHG emissions [9].

The state of Paraiba, Brazil, has excellent potential for solar power generation. The Sertão Paraibano region in the Northeast has one of the country’s highest rates of solar irradiation, between 5750 and 6250 Wh/m²/day [10]. According to the Brazilian Agency of Electricity (ANEEL) [11], there are eight plants in operation in the state of Paraíba (Northeast Brazil), with a total granted capacity of 136 MW and five units under construction with a full capacity of 135 MW. The state’s transmission and distribution networks are being expanded/adapted to absorb these new solar and wind farms better.

It is known that fossil fuel-based thermoelectric systems present the highest share of GHG emissions associated with the operation phase; photovoltaic systems, in turn, have low use-phase emissions and more significant emissions associated with the pre- and post-operational phase (e.g., raw material extraction, panel production, disposal) [12]. Mac Kinnon et al. [13] conducted a literature review to compare GHG emissions from electricity generation, obtaining a range of 0.306–0.681 kg CO₂-eq/kWh for thermoelectric systems (natural gas, combined cycle) and 0.020–0.104 kg CO₂-eq/kWh for multicrystalline photovoltaic systems. Using LCA can help equalize these different characteristics of the systems, resulting in a similar analysis of the parameters. In a comparative study, the equivalence of the systems being compared should be evaluated before interpreting the results [14].

Recognizing the importance of contributing to the analysis of generation systems and their respective GHG emissions, this study aimed to design a theoretical 16.4 MW (this is to maintain the same generation potential of already existing TEPPs in the state) solar photovoltaic system located in Patos (Paraíba state, Northeast Brazil). The LCA methodology is then applied to quantify the GHG emissions associated with its lifetime. The time required to amortize the emissions generated in the construction of the system (environmental payback) was also calculated.

A significant innovation of this study is the inclusion of Operation and Maintenance activities (e.g., vegetation management, washing of panels), which are not present in most LCA studies. Usually, only the replacement of equipment (inverters and panels) is considered. Regarding the environmental payback, this parameter reveals the relevance of the growth of the photovoltaic sector to produce electricity with lower carbon levels.

2. Material and Methods

This section initially designs the photovoltaic solar generation system and then focuses on the LCA.

2.1. Photovoltaic Solar Generation System

2.1.1. Initial Definitions

The first step is to estimate the area required for a 16.4 MW monocrystalline photovoltaic panel system. The capacity of the energy system was determined according to the capacity of existing TEPPs in the region (considering that photovoltaic electricity can displace the generation of TEPPs in the same location).

Following [15], an initial area of 1 km² will be predefined for this project. This value will be recalculated after the definition of the number and types of equipment required.

(a)

Available Area

The study area was selected based on its availability, proximity to the high voltage electric network, water supply, and paved roads. The study area should present relatively flat terrain and be close to the existing weather station.

An area located beside the BR-230 highway, 10 km away from downtown Patos, was delimited (6.973794° S; 37.355265° W), as shown in Figure 1.

(b)

Solar Resource

The selected region has a high incidence of solar radiation, ranging between 5750 and 6250 Wh/m²/day. The historical average of hourly solar irradiation was available [17], encompassing the period between 21 July 2007 and 31 December 2020. Data were collected by the Patos-A321 weather station, less than 15 km from the location chosen for the photovoltaic generation plant. The annual average solar radiation is 1887 kWh/m²/year.

(c)

Local Temperature

The average temperature at the installation site was obtained from weather station data (1 January 1975 to 31 December 2020) at 9:00 a.m. and 3:00 p.m. (local time) [17]. This parameter will be used to calculate the photovoltaic module temperature in the design phase.

(d)

Accessibility

The chosen site has road access by BR-230 highway, having good conditions for the supply of materials and flow of people. Because no new roads are required, road construction was not included in the LCA.

(e)

Connection to the Electricity and Water Network

The site is less than 70 km away from the LT 230 kV Milagres/Coremas transmission line and 10 km away from the urban area of the city of Patos. Infrastructure for the connection to the electric grid and water supply was not considered herein.

(f)

Structure of panels

For this project, the panels were considered to be installed with fixed support, tilt angle of 9°, and azimuth of 0°.

2.1.2. Calculation Methodology

The procedure presented by [18] was followed, who proposed the design and optimization of large photovoltaic plants with minimum Levelized Cost of Energy. In this section, the type of technology is selected along with its nominal power.

(a)

Selection of photovoltaic panels, inverters, and junction boxes.

Pieces of equipment were selected based on commercial availability in the region. The following technologies were included in the energy system: Canadian Solar Photovoltaic Panel CS6K-280M, ABB ULTRA-1500.0-HD-TL-US Inverter, and Sungrow PVS-24MH junction box. Appendix A presents the technical specifications of the selected models.

(b)

Pre-calculation of the number of panels ($N_{PVP}$) follows Equation (1).

$$N_{PVP}=\frac{P_{Design}}{P_{M,STC}}$$

(1)

where $P_{Design}$ [W] is the design capacity of the plant and $P_{M,STC}$ is the nominal power [W] of the photovoltaic module. Herein, P_Design = 16,400,000 W and P_M,STC = 280 W (Appendix A).

(c)

Pre-calculation of the area occupied by the modules ($A_{PTOT}$) involves calculating the area of each panel using Equation (2).

$$A_{PVP}=l·w$$

(2)

where $A_{PVP}$ [m²] is the calculated area of each panel, and l is length [m] and w is width [m] are dimensions for the selected photovoltaic module. Herein, l = 1.650 m and w = 0.992 m (Appendix A).

Equation (3) yields the total area occupied by the modules $A_{PTOT}$ [m²]:

$$A_{PTOT}=A_{PVP}·N_{PVP}$$

(3)

(d)

Calculation of the maximum number of photovoltaic modules in series and in parallel ($N\_S_{max}$, $N\_P_{max}$) encompasses the calculation of the average voltage of the inverter ($V_{inv,avg}$), from Equation (4).

$$V_{inv,avg}=\frac{V_{inv,max}·V_{inv,min}}{2}$$

(4)

$V_{inv,max}$ [V] is the maximum input voltage for the MPP (Maximum Power Point), $V_{inv,min}$ [V] is the minimum input voltage for the MPP.

The result of Equation (4) is an input to Equation (5), for the calculation of the maximum number of panels in series ($N\_S_{max}$).

$$N\_S_{max}=\frac{V_{inv,avg} }{V_{mod,max}}$$

(5)

$V_{mod,max}$ [V] is the maximum voltage of the module at the MPP.

Equation (6) is used to calculate the maximum number of photovoltaic modules in series N_S_max. The highest number of modules connected in series must be found without exceeding the $V_{DC inv,max}$ [V], which is the maximum input voltage allowed by the inverter.

$$N\_S_{max}·V_{OC mod,max}<V_{DC inv,max}$$

(6)

$V_{OC mod,max}$ [V] maximum open circuit voltage of the module.

If the above condition is not met, the number of modules in a series must be reduced until this condition is met.

To calculate the maximum number of modules in parallel $N_{P,max}$, Equation (7) is used. The highest number of modules connected in parallel must be found, without exceeding the $I_{DC inv,max}$ [A], which is the maximum input current allowed by the inverter.

$$N\_P_{max}=\frac{I_{DC inv,max} }{I_{mod,max}}$$

(7)

$I_{DC inv,max}$ [A] is the maximum direct current of the inverter and $I_{mod,max}$ [A] is the maximum current of the module. The number of panels calculated must be rounded down so that the maximum parameters are not exceeded.

(e)

Equation (8) determines the number of inverters ($N_{inv})$.

$$N_{inv}=\frac{N_{PVP} }{N_S·N_P}$$

(8)

$N_S$ is the maximum number of modules connected in series adopted in the calculations and $N_P$ is the maximum number of strings connected in parallel adopted in the calculations.

(f)

Equation (9) calculates the total number of panels ($N_{PVP,final}$) required.

The total number of panels must be calculated from the integers adopted for the number of photovoltaic modules in series and parallel and for the number of inverters.

$$N_{PVP,final}=N_S·N_P·N_{inv}$$

(9)

(g)

The Installed Capacity ($P_{installed}$, in W) is given by Equation (10), after defining the total number of panels.

$$P_{installed}=N_{PVP,final}·P_{M,STC}$$

(10)

(h)

The Final area occupied by the panels ($A_{PTOT, final}$, in m²) is given by Equation (11).

$$A_{PTOT, final}= A_{PVP}·N_{PVP,final}$$

(11)

(i)

Calculation of module temperature ($T_M$) follows Equation (12).

$$T_M=T_{amb}+\frac{G_t}{800}·\left(N_{OCT}-20\right)$$

(12)

$T_{amb}$ [°C] is the ambient temperature, $G_t$ [W/m²] is the incident solar radiation and $N_{OCT}$ [°C] is the nominal operating cell temperature.

The average temperature of the module was calculated hourly, using the historical average temperature ($T_{amb}$) and incident solar radiation ($G_t$) for the location stipulated for the photovoltaic plant [17]. The values entered in the spreadsheet refer to the average of the hours contained in a year.

(j)

Maximum Power Point (MPP) of each photovoltaic module after losses due to temperature ($P_{MPP}$) is determined by Equation (13).

$$P_{MPP}=P_{M,STC}·\frac{G_t}{800}·\left[1+\frac{Ƴ·\left( T_M-25\right)}{100}\right]$$

(13)

$Ƴ$ [%/°C] is the temperature parameter of the photovoltaic module at the MPP, specified in the selected PV module manufacturer’s catalog [19]. Herein Y = −0.41.

(k)

Equation (14) yields the actual power of each photovoltaic module after generation losses due to dust (F_r), shading (E_s), and temperature ($P_{mod}$).

$$P_{mod} = \left(1-\frac{F_r}{100}\right)·\left(1-\frac{E_s}{100}\right)·P_{MPP}$$

(14)

Representative values for the region are $F_r$ = 6.9% [20] and $E_s$ = 3% [15]. Pavan et al. [20] analyzed the effect of soil on the energy production of large-scale pv plants, resulting in Fr factors of 6.9% for sandy soils and 1.1% for compacted soils. The Fr factor of 6.9% was adopted due to the similarity of the plant’s soils. The factor Es adopted was the same used by [15], requiring a 3D model of the photovoltaic plant or real field measurements to improve accuracy.

(l)

Energy losses ($P_{mod,losses}$) are calculated by Equation (15).

$$P_{mod,losses}=P_{MPP}-P_{mod}$$

(15)

$P_{mod,losses}$ [W] represents the losses of each panel due to dirt and shading.

(m)

The output power of each photovoltaic set ($P_{PVSet}$) is given by Equation (16).

$$P_{PVSet}=N_S·N_P·\frac{{\mathsf{\eta }}_{MPPT}}{100}·\left(1-\frac{{\mathsf{\eta }}_{DC}}{100}\right)·\left(1-\frac{{\mathsf{\eta }}_{mismatch}}{100}\right)·P_{mod}$$

(16)

${\mathsf{\eta }}_{MPPT}$ [%] is the MPP efficiency of the DC/AC inverter, ${\mathsf{\eta }}_{DC}$ [%] is the DC cable voltage drop, and ${\mathsf{\eta }}_{mismatch}$ represents losses related to a slight difference in the manufacture of interconnected PV modules. According to [21], for most commercial PV inverters, ${\mathsf{\eta }}_{MPPT}$ is above 99% in most of the ac power-dc voltage range, and so ${\mathsf{\eta }}_{MPPT}$ [%] = 99% was used herein. Normally, the voltage drop must be less than 3%, and cable losses of less than 1% are achievable [22]. ${\mathsf{\eta }}_{DC}$ = 1.5% was the same adopted by [15]. The ${\mathsf{\eta }}_{mismatch}$ for this project was estimated at 2%, the standard value used by [23].

Figure 2 shows the configuration of the eight photovoltaic sets, composed of 23 panels interconnected in series. These strings are interconnected in parallel, resulting in 322 parallel connections for each inverter.

(n)

Equation (17) expresses the total output power of each DC/AC inverter ($P_{output,inv}$).

$$\begin{array}{c}\mathrm{If}, P_{PVSet} \le P_{i,adm}\to P_{output,inv}=\frac{{\mathsf{\eta }}_{inv}}{100}·P_{PVSet}\\ \mathrm{If}, P_{PVSet} \ge P_{i,adm}\to P_{output,inv}=\frac{{\mathsf{\eta }}_{inv}}{100}·P_{i,adm}\\ \mathrm{If}, P_{PVSet} \le P_{i,cons}\to P_{output,inv}=0\end{array}$$

(17)

$P_{i,adm}$ [W] is the maximum permissible power level of the inverter, ${\mathsf{\eta }}_{inv}$ [%] is the power conversion efficiency of the inverter and $P_{i,cons}$ is the energy consumption of the inverter. This factor is determined so that the panels do not send higher power than that supported by the inverter, as it can be damaged by operating above the limit.

(o)

The land occupied by the photovoltaic solar power plant ($A_{land}$, in km²) is given by Equation (18).

$$A_{land}= max (P_{output,inv})·{10}^{-6}·N_{inv}·Re{l}_{land}$$

(18)

$Re{l}_{land}$ was assumed to be the weighted average capacity presented by [24] for small PV plants (less than 20 MW), which is 0.028 km²/MW.

(p)

The power that the photovoltaic plant can supply to the grid ($P_{PLANT}$, in MW) is given by Equation (19).

$$P_{PLANT}=\frac{{\mathsf{\eta }}_T}{100}·\frac{{\mathsf{\eta }}_{cable}}{100}·P_{output,inv}·{10}^{-6}·N_{inv}$$

(19)

${\mathsf{\eta }}_T$ [%] accounts for the losses arising from the interconnection transformers and ${\mathsf{\eta }}_{cable}$ [%] for the efficiency of the alternating current cable connections. The ${\mathsf{\eta }}_T$ value is set at 99% [18] and ${\mathsf{\eta }}_{cable}$ is set to 99.5% [25].

(q)

The total energy supplied to the grid from the photovoltaic plant ($E_{PLANT,TOT}$) is expressed by Equation (20). The availability of generation of the plant will be predicted firstly, discounting the necessary maintenance stops.

$$E_{PLANT}=\frac{Avai{l}_{energy}}{100}·P_{PLANT}·\Delta t$$

(20)

$Avai{l}_{energy}$ [%] is the energy availability factor of the system considering the stops for maintenance of its components; this parameter is set at 99.5%, as presented by [15]. Garcia et al. [26] surveyed six PV plants in Spain, these plants were considered to have careful layout design and excellent maintenance services, and the obtained energy availability factors were above 99.7%. The $Avai{l}_{energy}$ was set at 99.5% to be conservative. $\Delta t$ [h] is the fraction of time used to calculate the energy supplied to the grid, and the value of $\Delta t$ is 1 h.

Equation (21) accounts for the total energy supplied to the grid throughout one year.

$$E_{PLANT,TOT}=\frac{Avai{l}_{energy}}{100}·{{\displaystyle \sum }}_{t=1}^n{P}_{PLANT}·\Delta t$$

(21)

where n ranges from 1 to 2190; this variation refers to the total hours of solar radiation available in the year, considering the period of solar incidence from 9:00 a.m. to 3:00 p.m. (Figure 2).

(r)

Degradation rate ($R_{degrad}$)

The degradation rates provided by the manufacturer of the solar panel are $R_{degrad1}$ = 3.0% in the first year, $R_{degrad2to6}$ = 0.7% from the second to the sixth year, and $R_{degrad7to25}$ = 0.5% from the seventh year onwards.

(s)

The total power supplied in 25 years ($E_{PLANT,TOT25}$) is given by Equation (22) in MWh. The annual contribution of the energy supplied to the grid is added, considering the accumulated degradation rates of the panels year after year during its lifetime (from n = 1 to n = 25).

$$E_{PLANT,TOT25}={\displaystyle \sum }_1^n{E}_{annual\left(n\right)}$$

(22)

(t)

The annual energy production ($E_{annual}$) depends on the degradation rate employed (R_degrad, which in turn depends on the operational year of the system). For the first operational year, E_annual(a) is employed; from years 2 to 6, E_annual(b) is used, and from year 7 onwards, E_annual(c) is used.

$$\begin{array}{c}{E}_{annual\left(a\right) }=E_{PLANT,TOT}·\left(1-R_{degrad1}\right)\\ E_{annual\left(b\right) }=E_{annual\left(b-1\right)}·\left(1-R_{degrad2to6}\right)\\ E_{annual\left(c\right) }=E_{annual\left(c-1\right)}·\left(1-R_{degrad7to25}\right)\end{array}$$

(23)

(u)

The total nominal energy of the photovoltaic plant ($E_{PLANT,NOM TOT}$) is given by Equation (24).

$$E_{PLANT, NOM TOT}=P_{installed}·\frac{G_{t PLANT}}{G_{t STANDARD}}·\Delta t$$

(24)

$E_{PLANT,NOM TOT}$ [MWh] is the nominal energy supplied by the plant. $G_{t PLANT}$ is the average radiation and $G_{t STANDARD}$ is 1000 W/m².

(v)

The performance rate ($PR$) is given by Equation (25), which expresses the relationship between the actual and nominal performances of the photovoltaic solar plant.

$$PR=\frac{E_{PLANT,TOT}}{E_{PLANT, NOM TOT}}·100$$

(25)

(w)

Equation (26) expresses the Capacity factor (CF, in %), the relationship between the actual and nominal energy productions of the photovoltaic plant throughout one year.

$$CF=\frac{E_{PLANT,TOT}}{P_{installed}·8760}·100$$

(26)

(x)

The amount of Full-Load Hours ($FLH$) is given by Equation (27).

$$FLH=\frac{ E_{PLANT,TOT}}{P_{installed}}$$

(27)

$FLH$ [h] or [kWh/kWp] expresses the number of hours the photovoltaic panel would need to operate at its nominal power to generate the same energy.

(y)

The total number of junction boxes ($J{B}_{inv,tot}$) is calculated by Equation (29) and requires the calculation of $J{B}_{inv}$ by Equation (28). This number has been rounded up to avoid overload at the junctions.

$$J{B}_{inv}=\frac{ N_P}{Ma{x}_{string}}$$

(28)

$$J{B}_{inv,tot}=J{B}_{inv}·N_{inv}$$

(29)

(z)

Specification of medium- and high-power transformers (CA)

The medium- and high-power transformers were specified according to the maximum power of the DC/AC inverters.

2.2. Calculation of Greenhouse Gas Emissions: LCA

Life Cycle Assessment (LCA) is an established, consolidated methodology to quantify the environmental loads associated with a product, service, or activity. LCA is standardized by ISO 14040 [27] (principles and framework) and ISO 14044 [14] (specification of requirements and provision of guidelines). ISO 14040 [27] is an overarching standard that encompasses the four phases of LCA: definition of objective and scope, analysis of inventory, environmental impact assessment, and interpretation of results.

The analysis developed herein accounted for the equipment (extraction of raw materials, manufacture, distribution, and transportation), use phase (operation and maintenance of the energy system), and final disposal (municipal landfilling, common practice of the region).

The result of Section 2.1 (photovoltaic solar generation system) provided the information necessary for preparing the life cycle inventory (equipment and materials), which is complemented by data on O&M and the construction of an administrative building. Appendix B presents the detailed life cycle inventory.

Connection to the transmission line and the water supply network were not included. A lifetime of 25 years is considered for the energy system, and transportation distance is considered as 2500 km for equipment and components (road transportation).

The materials necessary to build the support structure of panels and the number of cables and conduits were quantified based on [28], who presented a detailed quantitative table of materials needed per MW generated. The data is presented in Appendix B.

For the O&M processes, vegetation management, panel washing, and replacement of DC/AC inverters and photovoltaic panels were considered. To control vegetation growth, two mowing operations per year are considered in the corridors between the panel sets. Perpendicular and parallel tractor displacements in all corridors are considered. Monthly washing operations are considered for the panels, accounting for the displacement of the washing vehicle in the parallel corridors. These operations account for diesel consumption associated with the displacement and the tractor itself, with a lifetime of 30 years. For the replacement of inverters, 10% are replaced every ten years [28]. For the photovoltaic panels, a replacement rate of 0.05% per year was considered [29]. Appendix B contains a detailed description of the data involved.

SimaPro v.9.1.0.8 software [30] was utilized to develop the LCA using the Ecoinvent 3.5 database [31]. Due to environmental concerns regarding the emission of greenhouse gases (GHG), the environmental impact assessment method selected is the IPCC 2013 GWP 100y method [32], which groups the emissions of GHG gases and expresses the impact in terms of a common metric, kg CO₂-eq, throughout a time horizon of 100 years.

The functional unit used herein was the amount of GHG emissions associated with the energy system throughout its lifetime (kg CO₂-eq).

The total emissions of the photovoltaic solar generation system during its lifetime ($Emission{s}_{PV25}$) are calculated using Equation (29).

$$Emission{s}_{PV25}=\sum Emission{s}_{Installation}+\sum Emission{s}_{O\&M}$$

(30)

$\sum Emission{s}_{Installation}$ [kg CO₂-eq] is the sum of GHG emissions associated with the construction phase and $\sum Emission{s}_{O\&M}$ [kg CO₂-eq] represents the emissions associated with O&M. Appendix C presents a detailed account of the items considered in this calculation.

2.3. Determination of the Environmental Payback (T_payback) of the Photovoltaic Solar Generation System

After the LCA yielded the emissions associated with the energy system throughout its lifetime ($Emission{s}_{PV25}$), this amount is divided by the electricity generated in the same time period, and the emission factor ($E{F}_{PV}$, in kg CO₂-eq/kWh) is obtained, as shown by Equation (31).

$$E{F}_{PV}=\frac{Emission{s}_{PV25}}{E_{PLANT,TOT25}}$$

(31)

The emission factor of the Brazilian electricity mix (EF_Mix) was calculated following the methodology presented by Carvalho and Delgado [8] and considering the generation mix for the year 2019: 66.67% hydro, 9.28% natural gas, 9.15% wind, 8.25% sugarcane bagasse, 2.79% nuclear, 1.62% coal, 1.55% oil, and 0.69% solar [33].

Equation (32) defines the environmental payback (T_payback), and the return time (the breakeven point) is obtained when the equation equals zero.

$$\begin{array}{c}{T}_{payback1}=Emission{s}_{PV25}-E_{annual1}·1000·\left(E{F}_{\mathrm{Mix}}-E{F}_{PV}\right)\\ T_{payback \left(n\right)}=T_{payback \left(n-1\right)}-E_{annual \left(n-1\right)}·1000·\left(E{F}_{\mathrm{Mix}}-E{F}_{PV}\right)\end{array}$$

(32)

$T_{payback}$ [year] and n ranges from 2 to 25. $T_{payback}$ was evaluated from year to year because its value depends on $E_{annual}$, which also varies yearly. The 1000 factor is to convert the energy from MWh to kWh.

3. Results and Discussion

3.1. Sizing of the Photovoltaic Solar Generation System

Table 1. Equipment and components of the photovoltaic solar generation system.

16.4 MW Photovoltaic Plant
Photovoltaic Panel	Canadian Solar CS6K-280M
Inverter	ABB ULTRA-1500.0-HD-TL-US
$\mathrm{Panels} \mathrm{in} \mathrm{Series} (N_{S,max}$)	23
$\mathrm{Panels} \mathrm{in} \mathrm{Parallel} (N_{P,max})$	322
$\mathrm{Total} \mathrm{of} \mathrm{Inverters} (N_{inv})$	8
$\mathrm{Total} \mathrm{of} \mathrm{Panels} (N_{PVP,final}$)	59,248
Junction Box Model	Sungrow PVS-16/20/24MH
$\mathrm{Total} \mathrm{of} \mathrm{Junction} \mathrm{Boxes} (J{B}_{inv,tot}$)	112
1560 kVA–0.69 kV–33 kV Step-up Transformer	8
12480 kVA 33 kV–230 kV Step-up Transformer	2

presents a description of the technologies and components obtained from the design procedure of Section 2.1.

Table 2. Photovoltaic solar generation system production parameters.

Production Parameters
Pdesign [MW]	16.40
Pinstalled [MW]	16.59
Performance Rate—PR [%]	72.59
Capacity Factor—CF [%]	15.84
Full-Load Hours—FLH [h]	1388
Total Energy Production—EPLANT,TOT25 [MWh/25 years]	521,443

presents the main electricity production parameters for the proposed photovoltaic solar generation system.

Figure 3 shows a Sankey diagram of energy flows to visualize all losses accounted for in the calculations. The value of 23,022 MWh represents the electricity generated during the system’s first year of operation. The energy losses obtained herein agree with those presented by [34,35].

Figure 4 presents the effect of degradation on energy production over the 25-year lifetime of the system. The degradation rate depends on photovoltaic technology and climatic factors, such as solar irradiation intensity, temperature, wind speed and direction, dust, precipitation, and humidity.

The value of the total energy supplied throughout 25 years ($E_{PLANT,TOT25}$) is 521,443 MWh; if degradation were not included, this value would be 575,558 MWh. In other words, consideration of degradation represented a decrease of 54,115 MWh in total production (−9.4%). During the first year of operation, generation is 22,332 MWh, 3% lower than the nominal design production, due to the degradation of the panels. Energy production will decrease to 19,602 MWh in the last year of operation.

Appendix D shows, year by year, the values of the degradation rates used, the energy produced, and the accumulated value of energy produced during the system’s lifetime.

3.2. LCA

After the energy system was fully specified and with information on the material composition of equipment (along with transportation, O&M, and disposal), the GHG emissions were quantified.

Table 3. GHG emissions for the photovoltaic solar generation system.

Item	Emissions (kg CO2-eq)	Description	Emissions (kg CO2-eq)
Photovoltaic cells (monocrystalline silicon)	15,289,615	Silicon product	38,774
Aluminum alloy (AlMg3)	2,284,673	Polyvinyl Fluoride (PVF) Residues	24,901
Solar Glass	1,064,295	Heat (Natural Gas)	17,481
Electricity (Medium Voltage)	353,504	Drawn Copper	8309
Ethylene-Vinyl Acetate (EVA)	289,755	Brazing (Cadmium free)	5291
Polyvinyl Fluoride (PVF)	205,624	Propanol	4863
Tempered Glass	167,554	Acetone (Liquid)	3226
Glass Fiber Reinforced Plastic	163,223	Tap Water	2293
Manufacture of Photovoltaic Panels	123,456	Wastewater	1159
Polyethylene Terephthalate (PET)	113,859	Lubricating Oil	216
Plastic Residues	91,332	Nickel	216
Copper	87,462	Mineral Oil Residues	205
Corrugated Cardboard Box	72,607	Methanol	138
Total			20,414,032

presents the GHG emissions associated with the photovoltaic solar generation system (cradle-to-grave).

Table 4. GHG emissions associated with inverters.

Description	Emissions (kg CO2-eq)
Inverters—1560 kW	392,018
Transportation of Inverters—1560 kW	432
Total	392,450

and

Table 5. GHG emissions associated with transformers.

Description	Emissions (kg CO2-eq)
12480 kVA 33 kV–230 kV Step-up Transformers	253,089
Transportation of 12480 kVA 33 kV–230 kV Step-up Transformers	20,727
1560 kVA–0.69 kV-33 kV Step-up Transformers	151,948
Transportation of 1560 kVA–0.69 kV-33 kV Step-up Transformers	12,620
Total	438,384

show the contributions of GHG emissions related to inverters, and transformers (high- and medium- voltage), respectively.

Table 6. GHG emissions related to support structure and cabling.

Description	Emissions (kg CO2-eq)
Copper	406,505
Reinforced Steel	171,239
Low-alloy Steel	161,578
Drawn Copper	72,124
Polyvinyl Chloride (PVC)	70,490
Concrete	39,504
Zinc	28,287
Aluminum Production	15,050
Aluminum	6847
Total	971,623

and

Table 7. GHG emissions associated with junction boxes.

Description	Emissions (kg CO2-eq)	Description	Emissions (kg CO2-eq)
Polyethylene	8630	Copper Scrap	137
Polyethylene Residues	2639	Brass	48
Copper	1725	Zinc	44
Nylon	766	Polyvinyl Chloride (PVC) Residues	41
Polyvinyl Chloride (PVC)	612	Polycarbonate	6
Steel	561	Epoxy Resin	3
Electrical Wiring Residues	360	Steel Scrap	3
Drawn Copper	164
Total			15,458

show the contributions of GHG emissions related to the support structure, cabling, and junction boxes, respectively.

The emissions related to the construction of the administrative building are presented in

Table 8. GHG emissions for the administrative building.

Description	Emissions (kg CO2-eq)	Description	Emissions (kg CO2-eq)
Window Frame, aluminum	66,779	Plasterboard	285
Clay Brick	16,869	Alkyd Paint	241
Ceramic Tile	15,061	Inert Residues	123
Aluminum Scrap	3772	Diesel	110
Coated Glass	3165	Glass Residues	54
Fan + Heat Exchange Unit	2344	Tap Water	27
Concrete	2052	Wall Paint Residues	26
Wooden Doors	1325	Wastewater	23
Ceramic	824	Electricity (Low Voltage)	18
Concrete Residues	646	Plasterboard Residues	8
Total			113,004

.

Table 9. GHG emissions for O&M.

Description	Emissions (kg CO2-eq)
Replacement—Panels	255,136
Replacement—Inverters	98,113
Mowing	59,957
Washing of Panels	58,089
Total	471,294

shows the emissions associated with O&M throughout the lifetime of the energy system, based on the inventory presented in Appendix B.

Finally, Figure 5 shows the overall result of the LCA ($Emission{s}_{PV25}$), with a total of 22,817,275 kg CO₂-eq emitted over 25 years, accompanied by the respective percentages of contribution.

It can be verified that the installation phase is responsible for 97.9% of emissions and that emissions related to panels account for 89.47% of the total in this phase. The O&M phases contributed 2.1% to total emissions. It is essential to highlight that, of these 2.1% of emissions, 1.12% are related to the replacement of panels.

From the point of view of raw materials, silicon, and aluminum within the panels account for 78.3% of total emissions. When accounting for the replacement of panels, this value increases to 79.3%.

After calculating the value of total energy production (E_PLANT,TOT25) and total emissions ($Emission{s}_{PV25}$), the emission factor of the electricity generated by the photovoltaic plant was calculated ($E{F}_{PV}$), yielding 0.044 kg CO₂-eq/kWh. If the degradation rate of the panels were not taken into account, the value of $E{F}_{PV}$ would be 0.040 kg CO₂-eq/kWh, i.e., the degradation of the panels was responsible for the 10% increase in this factor. Degradation contributes to increasing the system’s emission factor, but its inclusion results in values closer to the actual ones.

Deterioration in crystalline silicon modules can vary from −0.6 to −5%/year [35,36]. This range encompasses the values employed herein: 3.0% for year one, 0.7% from year two to six, and 0.5% onwards.

More specifically to monocrystalline silicon, linear regression was employed to determine degradation, yielding 0.53 ± 0.01%/year [37]. In the study by Bansal et al. [38], a 5 MW grid-connected crystalline silicon solar photovoltaic power plant was analyzed between 2013 and 2019: the degradation rate exceeded 1% in 2016 and reached around 2–2.5% in 2019, raising concern. Jordan et al. [39] reported that the average degradation rate for crystalline silicon modules is 0.8–0.9%/year; in addition, hotter climates and mounting configurations can lead to higher degradation in some, but not all, products. As Ameur et al. [37] mentioned, the degradation observed in fielded PV modules is highest in hot climatic zones. Environmental factors, such as temperature, humidity, and UV radiation, are the main factors affecting the aging of PV modules [40].

These data alert the possibility that the degradation rate of a working energy system can be higher than the degradation rate provided in manufacturer catalogs. This can lead to a higher $E{F}_{PV}$ value for the designed system.

Al-Otab et al. [41] investigated the degradation rates for different types of photovoltaic modules with varying periods of exposure in German weather conditions. For outdoor tests, the overall efficiency drop for polycrystalline, monocrystalline, and amorphous thin-film modules was 14.3%, 37.7%, and 22.09%, respectively, for larger units, and 19%, 19.4%, and 21.7% respectively, for smaller units. These results suggest that different materials can considerably affect the design parameters of photovoltaic systems.

The emissions of grid-connected photovoltaic power generation in China were evaluated by [42]; for the same type of panels used herein (monocrystalline silicon), the results were 0.065 kg CO₂-eq/kWh for large-scale applications, slightly higher than those presented in this study, which was 0.044 kg CO₂-eq/kWh. The main factor contributing to this difference is the site’s annual solar radiation, 1600 kWh/m²/year used by [42] against 1887 kWh/m²/year used in this study.

3.3. Environmental Payback Time

Considering the Brazilian electricity mix in 2019, its emission factor ($E{F}_{Mix}$) is 0.227 kg CO₂-eq/kWh, approximately five times higher than 0.044 kg CO₂-eq/kWh obtained for the photovoltaic solar generation system designed herein.

Applying Equation (32) yields a payback time of 5 years, 8 months, and 1 day. Figure 6 shows the moment when the environmental payback $(T_{payback}$) equation equals zero, which is when environmental payback time is determined. Figure 6 also shows that, at the end of 25 years, 72,733,030 kg CO₂-eq is avoided compared to consuming electricity from the grid. This parameter reveals the importance of further disseminating photovoltaic solar electricity to mitigate emissions.

Even with Brazil’s energy mix predominantly comprised of renewable energy generation, the PVS showed much more advantageous emission values.

This environmental payback analysis provides a basis so that environmental aspects such as GHG emissions are also considered when choosing the type of generation to adopt.

4. Conclusions

This work quantified the greenhouse gas (GHG) emissions related to using a photovoltaic system to generate electricity. The analysis encompassed the total energy production and total emissions throughout the system’s lifetime. Application of the Life Cycle Assessment (LCA) methodology enabled the comparison of the emissions associated with the solar photovoltaic system and the Brazilian electricity mix.

The innovations of this study are the inclusion of maintenance processes in LCA; consideration of panel degradation throughout 25 years of operation (lifetime of the system); and determination of the environmental payback time when compared with the Brazilian electricity mix.

The pre-operational phases presented a high load of GHG emissions, with silicon and aluminum being the main contributors to the emissions. Inclusion of Operation and Maintenance increased GHG emissions by 2.1%, of which 1.12% was related to the replacement of panels. Overall, photovoltaic panels were the main contributors to emissions, representing 90.6% of the overall emissions. Combined, the silicon and aluminum required for producing the panels represent 79.3% of the total emissions.

The consideration of panel degradation represented a decrease of 54,115 MWh in total production over the system’s lifetime, equivalent to 9.4% less energy produced.

Although renewable sources mainly constitute the Brazilian electricity mix, the photovoltaic generation system presented low emissions (0.044 kg CO₂-eq/kWh) and yielded an environmental payback of five years and eight months, compared to the Brazilian mix (0.227 kg CO₂-eq/kWh), leading to a reduction of 72,733,030 kg CO₂-eq in GHG emissions over 25 years.

The comprehensive LCA developed herein demonstrated that the 16.4 MW photovoltaic generation system located in Northeast Brazil presented significantly lower GHG emissions than the Brazilian electricity mix.

Future studies can evaluate the implementation of tracking systems on photovoltaic panels. Although it is known that tracking devices increase the generation capacity, it is necessary to include the equipment and structures related to the tracker in the LCA inventory so that a more detailed analysis of the increase in emissions is also performed.