Real yields and PVSYST simulations: comparative analysis based on four photovoltaic installations at Ibn Tofail University

1 Introduction

Energy is often simply defined as the ability to do work or generate heat. Human beings con-sume energy in almost every aspect of life, and the source of energy varies with the degree of human civilization. In early human history, humans depended on the muscular strength of the human body, the energy of animals, the sun, water and air. However, with the discovery of fire, people began to use biomass as an energy source. Energy from the sun, wind, biomass, hydroelectricity, and geothermal energy are called renewable energy sources that naturally renew themselves in a short period of time. As human civilization progressed, coal was dis-covered and it largely replaced biomass, especially in industrialized countries. The discovery of oil and natural gas is significantly displacing coal, especially in developed countries. Coal, uranium, oil and natural gas are called non-renewable energy sources because they are de-pletable and take a long time to form (What is Energy). There is empirical evidence for links between energy use and population growth, and between energy use and economic growth.

The relationship between energy consumption and economic growth in China, India, and G7 countries from 1965 to 2017 is studied by applying multi-spatial fusion cross-mapping (CCM). For research purposes, the two sets of data, total energy consumption (EC) per capita in million tons of oil equivalent and real gross domestic product (GDP) calculated at constant prices in 2010 are from BP Statistical Yearbook of World Energy and World Development Indicators (WDI). A multi-spatial aggregated cross-mapping (CCM) analysis of the collected data revealed positive correlations between energy consumption and economic growth in the United States (US), Japan, Italy and Canada, China and India.

Photovoltaic self-consumption, energy efficiency and electric vehicles are topics of particular interest to ITU. Projects on these topics include green energy and the environment, photovoltaic systems. Several software programs optimize the simulation results of photovoltaic systems, PVSYST software is one of the most commonly used (El Mokhi et al. 2022). Therefore, efficiency and accuracy are keywords that every software user hopes to obtain, especially when the richness of the software confuses them. Morocco’s commitment to international efforts to combat climate change has been translated over time into concrete actions and plans in different sectors (Choukai et al. 2022). Therefore, efficiency and accuracy are keywords that every software user hopes to obtain, especially when the richness of the software confuses them. Morocco’s commitment to international efforts to combat climate change has been translated over time into concrete actions and plans in different sectors (Ait Errouhi et al. 2022). To implement the Royal Vision, Moroccan universities, especially Ibn Tofail University, are aware of climate change issues, especially since researchers at these universities have sounded the alarm and have the mechanisms and resources to assess the state of the Earth and in the context of climate change From the perspective of providing solutions for adaptation and mitigation, Ibn Tofail University (UIT) is committed to providing solutions through the development of scientific research and the accompanying public policy deadlines (Author Anonymous). UIT fulfills its mission to contribute to Morocco’s sustainable development, higher education advancement and economic growth by adopting and implementing several promising projects in the fields of development and environment. The University strives to further develop its contribution to research and its relationships with various stakeholders. UIT is firmly committed to environmental protection and actively participates in the implementation of sustainable green projects (Présentation de l’UIT).

The parameters describing energy photovoltaic systems and their components were developed by the International Energy Agency (IEA) Photovoltaic Systems Program and described in IEC standards (Ekeocha, Penzin, and Ogbuabor 2020). In short, the general definition of a power plant production factor describes how much of the energy produced during a power plant’s operation covers the energy used to build the plant.

Since the PVSYST simulator is the most famous software among researchers and PV experts, we decided to compare its results with those of manual methods. Therefore, the aim of this paper is to determine the degree of similarity between the PVSYST simulation results and the manual calculation results for the same parameters, based on a study of photovoltaic systems at the University of Ibn Tofail (Chandrika et al. 2020). Four photovoltaic systems were installed at the later university. These installations are installed in the Faculties of Letters, Science, President and CFC. The ones used later in our study have different parameters (orientation, electrical installation, type of photovoltaic panels …). Therefore, in order to collect these parameters and their various data, we used several sources (Chioncel et al. 2010; Kasim, Hussain, and Abed 2019). The data sheet for each component of a photovoltaic system contains parameters that differentiate the device, so it is imperative to look up this information as it will aid our research. The meteorological database PVGIS is our reference for obtaining temperature and sunshine data (Verma, Yadav, and Sengar 2021). In addition, PVSYST software facilitates the simulation of photovoltaic systems, thereby providing detailed reports with a wealth of information. All this data collection should help us compare the results of yield and other parameters with simulations in PVSYST through manual calculations (Mermoud and Wittmer 2014).

2 Detailed design and methodology

2.1 Collection of data

We will first determine the energy produced from E _ac and E _dc for each installation. To obtain the E _ac and E _dc energies, we had to retrieve them from the Solar–Log or Solar–Web™ website, put them in an Excel file and thus, calculate the average which will be used in the calculations.

2.1.1 Faculty Letters

The average E _dc = 227.706 and E _ac = 222.876 are calculated from the values of 03, 27, 28 June 2020 and 03, 10 June 2020 for the summer. For winter, the average E _dc = 185.59 and E _ac = 181.326 are calculated from the values of 11, 12, 30 January 2020 and 01, 03 February 2020 (literature university) Table 1.

Table 1:

E _ac and E _dc of 4 installations in summer and winter.

		FL (Letters Faculty)	FS (Sciences Faulty)	Presidency	CEC
	E dc (kWh)	227.70	274.42	106.96	220.16
Summer	E ac (kWh)	222.87	263.03	102.90	216.20
	System size (kWp)	46.8	60	21.06	34.56
	E dc (kWh)	185.59	219.45	84.95	121.82
Winter	E ac (kWh)	181.32	210.70	81.83	119.82
	System size (kWp)	46.8	60	21.06	34.56

2.1.2 Faculty Sciences

The average E _dc = 274.4225 and E _ac = 263.03 are calculated from the values of 01, 19 July 2021 and 05, 23 June 2021 for the summer. Regarding the winter, the average E _dc = 219.4525 and E _ac = 210.7075 are calculated from the values of 11, 12 January 2021 and 01, 03 February 2021 (Faculty of Science) Table 1.

2.1.3 Presidency

The average E _dc = 106.96 and E _ac = 102.8966667 are calculated from the values of 07, 12, 11 July 2021 and 05, 28, 29 June 2021 for the summer. For winter, the average E _dc = 84.95833333 and E _ac = 81.83666667 are calculated from the values of 04, 12 January 2021 and 10, 14, 23 February 2021 (Presidency) Table 1.

2.1.4 Center for continuing education

The average E _dc = 220.1575 and E _ac = 216.2 are calculated from the values of 07, 20 July 2021 and 01, 05 June 2021 for the summer. For the winter, the average E _dc = 121.825 and E _ac = 119.825 are calculated from the values of 03, 18 January 2021 and 15, 16 February 2021 (Center of Continuing Education CFC) Table 1.

Then, we will have to extract the irradiation of the two months January and February of the winter season, and June and July of the summer season. PVGIS is a web site containing the meteorological database such as irradiation and temperature. Choosing the location and month of study and integrating it into PVGIS is essential to have the exact irradiation values since they change depending on the geographical location and month. The following figures and tables represent the irradiation of the months January, February, June, and July.

1. The curves in Figure 1 represent the daily irradiation pattern of the two months January and February of the winter season where the irradiation decreases due to climatic conditions.

2. The average of the daily values for the month of February equals 342.22 W/m², and for January, it equals 295.781 W/m² (Table 2).

3. The curves in Figure 2 represent the daily irradiation of the two months June and July of the summer season where we notice a considerable increase.

4. The average of the daily values for June equals 547.81 W/m², and for July, it is 552.67 W/m² (Table 3).

Figure 1:

Irradiation of February and January-PVGIS meteorological database.

Table 2:

Irradiation: January and February-PVGIS meteorological database.

	January W/m2	February W/m2
	59.83	111.44
	204.9	265.85
	331.87	394.56
	427.42	495.58
	472.92	554.94
	472.93	558.90
	417.65	522.49
	328.55	432.08
	196.7	295.62
	45.04	132.3
	–	0.66
Average irradiation W/m 2	295.78	342.22

1. Bold values represent the daily irradiation in W/m².

Figure 2:

Irradiation of July and June-PVGIS meteorological database.

Table 3:

Irradiation: June and July-PVGIS meteorological database.

		June W/m2	July W/m2
		85.29	60.60
		247.50	217.03
		424.43	401.48
		593.60	598.86
		742.47	757.92
		875.16	886.27
		931.25	956.86
		933.49	951.70
		877.74	893.41
		755.76	772.21
		588.96	604.67
	387.17		404.41
		194.21	199.23
		32.34	32.73
Average irradiation W/m2		547.81	552.67

A third step consists in listing the components of each photovoltaic installation. The orientation and inclination of each system are essential for the simulation of the pre-existing systems. The following table summarizes the components, their types, the number of modules for each installation and the installed power (Table 4).

Table 4:

The components of each installation.

Data	Faculty Letters	Sciences Faculty	Presidency	CEC
Fabricant	JA SOLAR	JINKO SOLAR	JINKO SOLAR	BISOL
Type de PV	JAM60S01-315Wp	EAGLE DUAL 60 280Wc	EAGLE 60 270Wc	BISOL XL SERIES
Number of modules	156	216	78	108
Field power	46.8	60.48	21.06	34.56
Type of inverter	SUN2000-20KTL	SYMO 20.0-3-M	SYMO 20.0-3-M	DELTA ENERGY RPIM30A
Power of inverter kWc	20	20	20	30
Inverter efficiency	98.65 %	98.10 %	98.10 %	98.50 %
Angle	15	13	15	15
PV power rating (kWp)	0.315	0.28	0.27	0.32
Power kWp under field	(17 modules/4 chaînes) 21.42	(18 modules/12 chaînes) 60.48	(20 modules/2 chaînes) 10.8	–
Area m2	111	353.54	65.47	–
Power kWp under field	(22 modules/4 chaînes) 27.72	–	(19 modules/2 chaînes) 10.26
Area m2	143.89	–	62.1984	210.41
Orientation	158	−152	−150	−99

After the collection of data, we will proceed to evaluate PVSYST by comparing the results of two different methods:

1. Manuel method

2. Simulated method

2.3 Simulation method

2.3.1 PVSYST simulation

When using software, one always looks for an optimal result and minimum margin of error for the obtained results. PVSYST facilitates the task of dimensioning photovoltaic installations and allows to obtain various information such as energy production, necessary surface, and annual energy production (Chandrika et al. 2020).

To simulate the 4 photovoltaic installations, we need to:

1. First, choose the irradiation and temperature from the PVGIS weather database (Figure 3).

2. Second, the choice of tilts and orientations differ from one geographical location to another (Figure 4).

Figure 3:

Geographic parameters of the site.

Figure 4:

Orientations.

1. Third, the characteristics of each PV system (inverters, PV module type …) are entered into the software.

Concerning the installations of the Faculty of Letters (Figures 5 and 6) (power 46.8 kW and 156 number of modules), and for the Presidency (Figures 9 and 10) of the University Ibn Tofail (power 21.06 kW and 78 number of modules) their simulation was in the form of sub-fields contrary to the CFC (Figure 8) (power 34.56 kW to the number 108 modules) and faculty of Sciences (Figure 7) (power 60.48 kW and 216 number of modules) which constitutes a single field (Figures 8 –10).

Figure 5:

PVSYST project design of the F. Letter.

Figure 6:

PVSYST project design the F. Letter.

Figure 7:

PVSYST project design of the Science Faculty.

Figure 8:

PVSYST project design of the CFC.

Figure 9:

PVSYST project design of the Presidency.

Figure 10:

PVSYST project design of the Presidency.

2.3.2 Results of PVSYST simulation

The following figures determine 4 parameters which are: the performance index, the collection and system losses and the useful energy produced.

After the simulation of the installation of the Faculty of Letters, the normalized production (per kWp Installed) equal to 49.1 kWp does not differ strongly from the real value which equals 46.8 kWp. The performance index equals 0.854 considering all the months of the year (Figure 11).

Figure 11:

Nominal power et PR of Letters Faculty.

After the simulation of the installation of the Faculty of Science, the normalized production (per kWp Installed) equal to 60.5 kWp does not differ strongly from the real value which equals 60.48 kWp. The performance index equals 0.848 considering all the months of the year (Figure 12).

Figure 12:

Simulation PVSYST nominal power et PR of Science Faculty.

After the simulation of the installation of the Presidency, the normalized production (per kWp Installed) equal to 21.06 kWp equals the real value. The performance index equals 0.850 considering all the months of the year (Figure 13).

Figure 13:

Simulation PVSYST nominal power et PR of the Presidency.

After the simulation of the CFC installation, the normalized production (per kWp Installed) equal to 34.6 kWp does not differ strongly from the real value which equals 34.56 kWp. The performance index equals 0.848 considering all the months of the year (Figure 14).

Figure 14:

Simulation PVSYST nominal power et PR of the CFC.

The following diagrams summarize the different losses of the system such as the losses of inverters and cables to finally determine the value of energy injected into the network:

1. The nominal field energy (according to STC) of the first simulation of the Faculty of Letters is equal to 77.5 MWh/year which gives a value of energy injected into the grid of 69.4 MWh/year after losses at the irradiation levels, camp temperature, module quality as well as cable and inverter losses (Figure 15).

2. The nominal field energy (according to STC) of the first simulation of the Faculty of Science equals 97.2 MWh/year which gives a value of energy injected into the grid of 86.1 MWh/year after losses at irradiation levels, camp temperature, module quality as well as cable and in-verter losses (Figure 16).

3. The energy injected into the grid equals 29.6 MWh/year after losses at irradiation levels (−0.3 %), field temperature (−7.7 %), module quality (+0.7 %) as well as cable and inverter losses (Figure 17).

4. The energy injected into the grid equals 54.5MWh/year after losses at irradiation levels (−2.1 %), field temperature (−8.1 %), module quality (+0.7 %) as well as cable and inverter losses (Figure 18).

Figure 15:

Loss diagram: Faculty of Letters.

Figure 16:

Loss diagram: Faculty of Science.

Figure 17:

Loss diagram: Presidency.

Figure 18:

Loss diagram: CFC.

3 Comparison and discussion

When comparing the performance index obtained by calculation and by simulation, the latter is between −1.5 % and 13.4 % Table 17.

To better understand this comparison, taking as an example a 35 kWp installation connected to the grid and with a performance index of 86 % and considering the comparison interval already made, the PR of this installation obtained via PVSYST will be between 84.71 % and 97.5 %. In other words, the performance index of the PVSYST software of the installation is close to the real value while respecting the interval previously determined.

The accuracy of this index is even more accurate when the following parameters are taken into consideration:

1. Degree of dust presence.

2. The length of the cables that induce losses.

3. Degree of shading.

4. The uncertainty of PVGIS.

5. Annual degradation (1 %).

It should be noted that in the calculation made, unlike the PVSYST calculation that considers all the months of the year, only specific days are considered.

The following Table 18 are a summary and an overview of the calculated and simulated values of the parameters Y _f, L _c and L _s in order to calculate their performance.

After the simulations of the four installations on the PVSYST software, the results are as follows:

1. The specific yield for Faculty of Letters equals 3.87 kWh/kWp/dr, L _c equals 0.6 kWh/kWp/dr and L _s equals 0.07 kWh/kWp/dr.

2. The specific efficiency for the Faculty of Science is 3.9 kWh/kWc/dr, L _c is 0.61 kWh/kWc/dr and L _s is 0.09 kWh/kWc/dr.

3. The specific yield for the Presidency equals 3.88 kWh/kWc/dr, L _c equals 0.6 kWh/kWc/dr and L _s equals 0.09 kWh/kWc/dr.

4. The specific output for Continuing Education Center equals 4.23 kWh/kWp/dr, L _c equals 0.66 kWh/kWp/dr and L _s equals 0.09 kWh/kWp/dr.

The tables below summarize the results of the parameters Y _f, L _c and L _s obtained by calculation.

1. The averages of the specific yield Y _f = 4.318 kWh/kWp/dr, L _c = 1.413 kWh/kWp/dr, and L _s = 0.466 kWh/kWp/dr encompass both the summer and winter season values of the FLK PV plant.

2. The averages of the two winter and summer seasons of FSK and of the specific yield and losses L _c and L _s are consecutively equal 3.9155 kWh/kWc/dr, 6.682 kWh/kWc/dr, 0.166 kWh/kWc/dr.

3. The averages of the two winter and summer seasons of the Presidency and of the specific yield and losses L _c and L _s are consecutively equal to 4.385 kWh/kWc/dr, 1.6375 kWh/kWc/dr, 0.17 kWh/kWc/dr.

4. The averages for the two winter and summer seasons of the CFC Training Center and of the specific efficiency and the losses L _c and L_s are consecutively equal to 4.858 kWh/kWc/dr, 0.8175 kWh/kWc/dr, 0.0885 kWh/kWc/dr.

The results in (Table 19) show the difference between the calculated and simulated values. The difference between calculated and simulated Y _f does not exceed 13 %.

The results in (Table 19) show the difference between the calculated and simulated values. The difference between calculated and simulated Y _f does not exceed 13 %.

In the case of the CFC, the efficiency of the losses L _s is equal to −2 %. This is an exception, which shows that the calculated values are considerably greater than the values obtained by simulation.

The parameters not taken into consideration generate losses, hence their yields are between a minimum value equal to −2 % and a maximum value equal to 91 %. This is another indication that affirms the hypothesis that the losses not taken into consideration can affect the results. Therefore, the more losses are taken into consideration, the more accurate the results.

4 Conclusions

The abundance of simulation and sizing software for PV systems as well as the user’s requirement for accuracy, efficiency and speed can lead the researcher to evaluate them. Since in my academic environment in the use of PVSYST is favorable, the choice of the title of the subject is the comparison between PVSYST simulations and real PV production. The objective was to evaluate the accuracy of the PVSYST software. First, it was necessary to make a bibliography of articles with the intention of validating the equations to be used. Secondly, determining the components of the installations was essential for the simulation. Another task to validate was the choice of meteorological databases since the calculations depended on them. At the end, a data analysis was done whose main goal is to evaluate the software. The evaluation of the software was based on the uncertainty rate of the performance index which is between −1.5 % and 13.4 %. The latter depends on several parameters such as the temperature, the degree of shading, the length of the cables, the degree of the presence of dust, and not forgetting the annual degradation which is 1 % and the uncertainty of the values of the irradiation collected from the PVGIS database. The result of the simulations depends on these parameters; therefore, losses are generated. Hence, the interest in analyzing the difference between the estimated and simulated parameters which provides values up to 91 % for the L _c losses of the FLK, as an example. PV installations require precise dimensioning and PVSYST is a quite powerful software. The more one takes into consideration the parameters and the real conditions, the closer the PR index and the losses of the simulation are to those calculated.

The objective was to evaluate the accuracy of the PVSYST software. First, it was necessary to make a bibliography of articles with the intention of validating the equations to be used. Secondly, determining the components of the installations was essential for the simulation. Another task to validate was the choice of meteorological databases since the calculations depended on them. At the end, a data analysis was done whose main goal is to evaluate the software. The evaluation of the software was based on the uncertainty rate of the performance index which is between −1.5 % and 13.4 %. The latter depends on several parameters such as the temperature, the degree of shading, the length of the cables, the degree of the presence of dust, and not forgetting the annual degradation which is 1 % and the uncertainty of the values of the irradiation collected from the PVGIS database. The result of the simulations depends on these parameters; therefore, losses are generated. Hence, the interest in analyzing the difference between the estimated and simulated parameters which provides values up to 91 % for the L _c losses of the FLK, as an example. PV installations require precise dimensioning and PVSYST is a quite powerful software. The more one takes into consideration the parameters and the real conditions, the closer the PR index and the losses of the simulation are to those calculated.