Performance Modeling of the Weather Impact on a Utility-Scale PV Power Plant in a Tropical Region

College of Engineering, Universiti Malaysia Pahang, 26300, Gambang, Kuantan, Pahang, Malaysia Agency for New and Renewable Energy Research and Technology (ANERT), Thiruvananthapuram, Kerala, India Centre for Automotive Engineering, Universiti Malaysia Pahang, 26600 Pekan, Pahang, Malaysia Faculty of Mechanical and Automotive Engineering Technology, Universiti Malaysia Pahang, 26600 Pekan, Pahang, Malaysia Energy Centre, Maulana Azad National Institute of Technology, Bhopal, India IEEE, Kerala Section, Thiruvananthapuram, 695001 Kerala, India Department of Chemical Engineering, Manipal Institute of Technology, Manipal Academy of Higher Education, Manipal, Karnataka, India


Introduction
Solar PV has a greater role in the global energy transition for meeting the challenges of climate change. The Paris Agreement demands the global nations to limit the global temperature rise to 1.5°C. Energy-related carbon emission has to be reduced by around 3.5% per year to 2050 to address this challenge. The demand for solar PV is increasing across the world, and it becomes the most competitive option for electricity generation in many numbers of countries. Solar PV is expected to become the cheapest source of energy by 2050 in the regions where solar radiation is abundant [1]. A total cumulative solar PV capacity of 627 GW was installed up to 2019. The interest in solar PV capacity addition resulted in the addition of at least 1 GW capacity in more than 29 countries up to 2019. The major share in gridconnected renewables is from utility-scale solar power plants and also from the rooftop sector. The cost of solar PV continued to decline mainly due to the reduction in the price of PV modules [2]. According to the International Renewable Energy Agency, solar generation is increased by 28% from the previous year based on the data available from the sources [3]. Ground-mounted utility-scale solar PV projects had a major share of 70-72% in the global installed capacity. Utility-scale PV plants continue to be the major contributor to PV capacity addition in the world [4]. The performance of the solar PV power plant is influenced by various weather parameters like solar irradiation, temperature, wind speed, rain, humidity, and atmospheric pressure. The output of the PV modules depends on the solar radiation falling on the surface. The operation and maintenance of solar PV plants play a vital role in improving the performance of PV power plants. The operation and maintenance of a PV plant are extremely important for its sustainability [5].

Literature Review: Weather-Related Studies Conducted in Solar Power Plants
The performance of the solar PV power plant is influenced by various weather parameters like solar irradiation, temperature, wind speed, rain, humidity, and atmospheric pressure. The output of the PV modules depends on the solar radiation falling on the surface [6]. Parretta [12]. Vasel and Iakovidis discussed the effect of wind direction on the performance of a utility-scale PV plant [13]. The relevant performance studies conducted in different climate regions are summarized in Table 1. Based on Table 1, the performance ratio of the PV plant was found to vary between 61% and 88%. There is a large variation in the performance of the plant among different climates.
Most of the utility-scale power plants are installed with an embedded data monitoring system which includes a cellbased or thermopile-based pyranometer for solar radiation measurement. Most of the studies were conducted by utilizing the data from conventional data monitoring systems embedded with the solar power plant project when it is constructed by the developer. Sensors for measuring the ambient temperature, wind speed, and module temperature are also incorporated in most of the embedded data monitoring systems. All these studies have limitations since the data is completely dependent upon the measuring instruments of various qualities. Also, there is a chance that the instruments may not be calibrated as per the standards. However, these measuring instruments if not calibrated regularly may end up with erroneous measurement which will lead to wrong performance conclusions of the solar power plants. All the above studies were conducted based on the average monthly A comprehensive analysis of the performance modeling of PV plant generation and the influence of weather parameters like the effect of humidity, atmospheric pressure, and rain are not reported by researchers in detail. Considering the gaps in studies as found in the literature, this research work is carried out with the most accurate and reliable data from a Solar Radiation Resource Assessment (SRRA) station installed as part of the National Solar Resource Assessment Programme conducted in India.
In this paper, the influence of weather parameters like solar irradiation, air temperature, rain, wind speed, atmospheric pressure, and dew point on the performance of PV plants is studied in detail using a very sophisticated Solar Radiation Research Assessment (SRRA) station.
The objectives of this research work are as follows: (1) To analyze the weather impact of a utility-scale PV power plant in a humid tropical region utilizing SRRA data Thomas and Prasannakumar described that the state of Kerala is experiencing tropical humid climatic conditions with two monsoon periods with more than three months of heavy rainy seasons [29]. Ananthakrishnan and Soman had described a comprehensive estimation of the rain in Kerala [30]. According to Hunt and Menon, the monsoon is spread over 6 months in Kerala with an average of 300 mm rain [31]. Figure 2. This SRRA station is installed as part of a separate Solar Radiation Resource Assessment Programme in different parts of the country and installed on the same premises of the 2 MW solar PV plant which turned out to be a novel arrangement for the measurement of different weather parameters utilizing state-of-the-art measuring instruments. The methodology involves accurate weather data collection from a Solar Radiation Resource Assessment (SRRA) station. This SRRA station is used as a primary data collection source for the study. The SRRA station is operational since November 2018, and the big data is streaming since then from the SRRA station and is being stored in the central server. This data is continuously monitored in the central server. This data at a one-minute interval is collected through the above measuring sensors and transferred in realtime from the SRRA station. The most recent year of 2019 is taken for analyzing the influence of weather along with the generation data of the PV power plant which is available from the traditional embedded data monitoring system. Table 2 gives the summary of the measuring instruments in the SRRA station which are used for measuring the weather parameters at the site of the PV power plant.

Data Collection. A typical SRRA station installation is shown in
The following data is available from the SRRA station: The SRRA station provides the data of global horizontal irradiation through a pyranometer. Global tilted irradiation is calculated for determining various performance parameters like performance ratio and module temperature. The data logger and modem at the SRRA station combined with GPRS will accomplish the task of data collection and monitoring.  Kalogirou has consolidated various mathematical equations used for the determination of global tilted irradiation and module temperature [32]. The equations are given below.
3.3.1. Estimation of Global Tilted Irradiation. Global tilted irradiation (G Bt ) or the beam radiation on a tilted surface is calculated by multiplying global horizontal irradiation (GB) and beam radiation tilt factor (RB). Global horizontal irradiation is the sum of the beam (direct) and diffuse radiation and is also known as "global radiation." The beam radiation tilt factor (RB) is given by The declination angle (δ) is defined as the angle between the lines joining the centers of the sun and the earth, which also determines the direction of the direct rays coming from the sun and their projection on the equatorial plane and is given by    4 International Journal of Photoenergy The value of N is the date number which ranges from 1 to 365.
The zenith angle is the angle between the sun's rays and the line perpendicular to a horizontal plane, and the hour angle is the angular displacement of the sun from the local meridian because of the earth's rotation around its own axis [33].
The hour angle (h) is given by The latitude (N) is also an important factor determining the hour angle as given above.
Estimation of module temperature is given by (1) Detailed analysis for a selected month in a season (2) Yearly data analysis The various months that represent the typical characteristics of the season are selected for the regression modeling and arrive at the energy generation model equations.

Modeling the PV Generation with Weather Parameters.
Minitab is the statistical tool used for this analysis and for arriving at regression model equations from big data. Minitab is a statistical package developed at Pennsylvania State University, USA. It offers various computations in the field of big data analytics. Here, the weather parameters are selected as variables, and the generation is made as a response to various seasons. The outcome of this methodology is the development of a mathematical model for the 2 MWp PV power plant based on various weather parameters. Once the model is created, the generation can be theoretically calculated from the mathematical model. The important weather parameters selected for regression modeling are given in Table 3.
Minitab is also used for fitting the probability distribution plot for a typical generation (kWh) in a day against the representative average daily solar generation (G POA ) of the year.

Results and Discussion
The energy generation profile for different seasons is discussed in the following section. The plant performance results are summarized in Table 4 for arriving at conclusions. The year 2019 was faced with heavy rain in Kerala, particularly in Palakkad District where the PV plant is located. Tropical humid regions are characterised by rainy seasons, and there will be impact on the PR in these regions due to rain. Month-wise generation, global tilted generation, peak sun hours, PR, wind speed, air temperature, relative humidity, atmospheric pressure, and rain are summarized in Table 4.
It is observed that the performance ratio (PR) of the plant was reasonable during the summer season (Feb-71.67, March-71.72, April-69.75, and May-72.69), obviously due to higher energy generation. The advantage of colder temperature improves the PR of the plant further during the rainy months of June (75.95) and July (74.09) as compared to the summer season though the overall generation was lower. The highest PR of the plant was observed during September (79.50) owing to the lesser solar radiation. The lowest sun peak hours have been observed as 4.41, 4.04, 3.48, and 4.18 during June, July, August, and September of 2019, respectively, which indicate the reduction in solar insolation during the rainy season. This directly influences the performance of the plant as well.

Effect of Temperature on Generation.
Eight representative days were chosen for different seasons to study the effect of temperature on generation and are presented in Table 5. Among the eight days, selected for the study, the lowest average ambient temperature was recorded during January obviously due to the winter season. The daily energy generation for the winter days is higher than that for other days chosen for the analysis. The combined effect of optimum temperature and sunshine duration contributed to the higher  5 International Journal of Photoenergy generation which clearly shows that the winter seasons are also favorable for energy generation in tropical humid regions. The summer solstice in the tropical region falls in the southwest monsoon season, and hence, lesser generation was observed due to the cloudy and rainy conditions prevailing in the regions. This emphasizes the influence of rain in keeping the ambient temperature lower when compared to the winter months. The air temperature and module temperature show a continuous correlation throughout the seasons. On a typical summer day, the module temperature can reach above 44°C and the module temperature shows a comparatively lower level in rainy months as a general tendency in tropical regions. The module temperature follows the air temperature almost constantly throughout the summer seasons which is a common trend in tropical regions.

Effect of Irradiance on Generation.
The eight representative days were chosen for different seasons to study the effect of irradiance on energy generation as illustrated in Table 6. The highest global tilted irradiance has been observed on winter days, and correspondingly higher energy generation happened. Out of 8 days, 4 days achieved more than 8000 kWh generation which shows the correlation between higher global tilted irradiance and generation. It has been observed that the highest average global tilted irradiations happened during both summer and winter seasons. It is found that both summer and winter days provide the best solar generation among the selected days. The highest global tilted irradiance was observed mostly in summer days except for three days in monsoon. This indicates the importance of summer in tropical regions in achieving higher performance as it is evident from the higher generation values during the summer season.
Hence, it can be concluded that the typical Kerala humid climate supports good generation in the first six months of a year and dips into the lesser generation in monsoon months until it regains the status in the final months of the year.

Modeling of Solar Generation in Different Weather
Seasons. The data from the SRRA weather monitoring station is analyzed for the year 2019, and that has been used to develop a regression model for the entire four weather seasons for a 2 MWp PV power plant installed in a tropical region. The modeling output is calculated for the typical one day of the selected season as per the model equation below, and then, that value has to be multiplied by the number of days in that season to get the total generation. The weather parameters considered for the study are global tilted irradiation, wind speed, air temperature, relative humidity, atmospheric pressure, and the rain of the site. Here, the solar generation is derived from the selected weather parameters The above models are tested utilizing the 2020 SRRA data, and the results are checked and compared with the actual generation data as given in Table 7.
The above AGS model can be used to predict energy generation of a 2 MWp PV plant for a typical day in each season based on the important weather parameters in a tropical region as discussed above. The model has been tested and validated utilizing the 2020 year's measured data. The percentage average variation in the prediction of energy generation using the model is within the acceptable limits (9.87%). The novelty of this work is that this model brings important weather parameters like wind speed, air temperature, relative humidity, atmospheric pressure, and rain into this model which was not previously taken up in any research work along with seasonal solar power PV plant generation.

Probability Distribution
Model. The data set for the average daily generation (kWh) is plotted against global tilted irradiance as a scatter plot, and a line curve is fitted for the best value as given in Figure 3. The cubic model is arrived at by utilizing the Minitab statistical tool. The scatter plot is   This probability distribution predicts the values for generation in a day on varying irradiance which when multiplied by 365 days will predict the annual generation.
The validation of the probability distribution model is tested using the measured data of 2018 and 2020, and the model is also compared with the PVsyst simulated generation, and the details are summarized in Table 8. The percentage variation of predicted generation by the probability model in the years of 2018 and 2020 is within the acceptable limits with the actual generation.
This simple model predicts the average daily generation of a 2 MWp PV power plant installed in a tropical region for a complete year with a single parameter of global tilted irradiation in the equation. This model helps for a quick prediction of solar generation PV power plant sites in a tropical region. Predictive maintenance, remote sensing, and solar forecasting will help in monitoring the solar power plant performance.

Conclusion
This research work studied the influence of important weather parameters on the performance of a utility-scale power plant installed in a humid tropical region in the state of Kerala, a southern state of India. Unlike conventional data collection, big data collected from the Solar Radiation Resource Assessment (SRRA) station is used in this research work for accurate estimation. In addition to the usual weather parameters like ambient temperature, global tilted irradiance, and wind speed, other influential parameters like rain, atmospheric pressure, and humidity are also analyzed and modeled for the major weather seasons of the region. The following are the major outcomes of the study.
(i) Solar generation and its performance are affected during the rainy seasons, and it turns out to be a typical phenomenon in humid tropical regions. It is also understood that the prevailing wind during the monsoon periods does not make much influence on the performance of the PV power plant (ii) The generation of the PV plant in a tropical region strictly follows the irradiance pattern. Global horizontal irradiation and global tilted irradiation are in analogy providing comparatively higher values compared to monsoon months (iii) It is observed that higher generation happens during the period of lower atmospheric pressure in tropical regions (iv) Higher generation and reasonable PR are the highlights of summer seasons. The solar generation and performance characteristics of the winter season are almost the same as those of the summer season in tropical regions since all the performance parameters are observed to be similar. The plant performed well during the summer and winter seasons when compared with the monsoon seasons (v) The monsoon rain has an influence on the ambient temperature of the region which lowers the module temperature. As a general trend, module temperature follows the ambient temperature throughout the seasons (vi) The 2 MWp solar PV plant has got comparatively higher PR during rainy seasons due to the advantage of colder temperature during rainy seasons This research work will aid future studies on performance analysis of PV plants in other geographical regions with different climatic conditions.