The Artificial Intelligence Use in Solar Panels Proposal

Aim and Focus of the Study

The use of solar PV panels as sources of renewable energy has been gaining traction in the recent decades. As a result, there has been increased competitiveness in the installation of PV panels. Meanwhile, the growing usage of artificial intelligence continues to enhance improved performance predictions through computational power and higher data availability. The need to predict solar PV energy output remains very essential to many players in the renewable energy industry. Artificial intelligence can be leveraged to achieve this end, particularly with regard to weather input parameters such as humidity and dust rate. This study proposal aims to present an approach that can be used to predict the output of a solar PV panel based on weather input parameters through the use of artificial intelligence. Different models of artificial intelligence with the features of humidity and dust rate will be created. The datasets will be collected in different areas within the United Arab Emirates (UAE).

Research Context

The continued growth of the economy has resulted in an increase in the demand for energy across the globe. However, there are fears that the finite energy reserves are rapidly getting depleted. Besides, the Dudley [1] observes that overreliance on fossil fuel as the primary energy source is taking a catastrophic toll on the global environment, thanks to global warming and climate change. Solar energy plants are thus emerging as the most appropriate renewable energy alternatives that will help reverse these trends. In recent years, there has been an increased uptake and installation of solar PV panels around the globe. In 2019 alone, 117 gigawatts were generated from solar PV power as demostarted by Alomari [2]. Artificial intelligence is now a common phenomenon in predicting and classifying solar PV panels output against the input variables. This is because of its ability to process nonlinear and complex problems reliably.

Fuzzy logic, K-nearest algorithm, artificial neural network (ANN), decision tree-based technique, and the support vector machine are the most common AI techniques in improving the performance of photovoltaic forecasting models. Gligor et al. [3] define artificial intelligence as a technology that has the potential to make quicker, better, and more practical forecasts as compared to traditional methods. In the opinion of Son et al. [4], when predicting the solar PV panels’ output, it is important to consider the prevailing environmental conditions, such as weather, humidity, and air pollution.

Different prediction models have been advanced using weather features to estimate the solar panels’ power output. For instance, according to Nageem and Jayabarat (5), a multi-input support vector regression (SVR) model can be used to forecast the output of a solar panel connected to a grid. In arriving at this conclusion, the authors considered such weather features as temperature, the speed of wind, humidity, and pressure. From their experimental analysis, Son et al. [4] observe that the SVR model produced more effective and accurate results as compared to the analytical model. The authors also used artificial neural networks (ANN) to predict solar panels’ power output on such weather features as irradiance and temperature. A five-year dataset demonstrated that such machine learning models as K-Nearest Neighbors, Random Regression, Gradient Boosting Regressor, and Linear Regression, have the potential to produce stronger performances.

Solar panels’ power output can also be predicted by the use of pollution features, particularly the atmospheric dust rate. While studying the impacts of particulate matter on South Korea’s solar output, the authors used the concentrations of PM_2.5 and PM₁₀ for the 2015 to 2017 dataset. The authors clarified that the PMs normally decrease the output of solar power by more than 10% [6]. From their results, they recommended that the PMs have negative effects on solar panels, and this should be taken into consideration in policymaking that targets South Korea’s solar panels. In addition, the authors calculated the production of solar energy due to particulate and dust air pollution by merging global modeling and field measurements to evaluate the effect of PM and dust on the generation of solar electricity. The results show that the production of solar panels was decreased by between 17 to 25% as a result of the PM on the solar panels’ surface, as reported by Bergin et al. [6]. From these previous studies, it is apparent that weather features adversely affect the production of power using solar PV panels.

Theoretical and Historical Perspectives and Interpretations

Over the past decade, there has been an unprecedented shift around the globe toward renewable energy sources. This has led to, among other things, the reduction in costs related to electricity production using PV panels. Along with this, the efficiency of energy conversion has significantly increased as well. Specifically, Chuluunsaikhan [7] observes that there has been a decrease of 73% between 2010 and 2017 on electricity levies charged on largescale PV panels. As a result of the increased efficiency and the reduction in costs, PV panels are fast emerging as competitive alternatives in several countries across the world.

Nonetheless, it must be noted that the output of PV panel energy is dependent on conditions of weather such as humidity, temperature, and dust rate in the air. This implies that the output of energy in PV solar panels is often unstable.

Therefore, it is imperative to understand and appropriately manage the variability of energy output from these sources. Indeed, there are indications that as more countries step up their investments in renewable energy sources, solar PV panels consumption will be on the rise as well. Kurukuru et al. [8] argues that this will increase the need to appropriately predict solar PV energy output. Artificial intelligence is one of the most effective ways that can be used to achieve this. However, despite the evidence of the demand for efficient and accurate prediction of solar PV energy output, finding a solution to this is a complex undertaking.

Weather variations are a constant nuisance that poses several challenges, which interfere with accurate weather predictions. As the PV power prediction solutions increase in demand, there has also been a proportional growth in the popularity of the prediction means through the help of artificial intelligence, as observed by Mustafa et al. [9]. Although the use of AI is not new, there has been an improvement in its capacity for computation. Besides, the high-quality data availability has enhanced the AI technique resourceful for PV solar panel energy output predictions.

Research Design

Experimental research will be carried out in four regions of the UAE. These are the areas with the highest and lowest humidity and those with the highest and lowest concentration of particulate matter in the air. Information about the exact locations of these areas will be collected from the country’s meteorological department.

Research Methodology

Meteorological data, specifically humidity and dust rate in the air, will be collected from the UAE Meteorological Department. The data will include mean average, daily minimum and, maximum humidity as well as the particulate matter from January 1, 2021, to December 31, 2021. All the data will be gathered from the four areas that the meteorological department will avail. The data that then be tabulated and analyzed qualitatively

Ethical Consideration

Since this study will not involve participants, issues of confidentiality, privacy, and consent will not arise. However, there is a philosophical viewpoint that revolves around the utilitarianism ethics of solar power energy. This school of ethics states that energy policies involving solar power should benefit very many people. However, this is not the case as solar panels are still costly out of the reach of ordinary people. Besides, while solar power is billed as an efficient renewable energy source, there is no denying that the performance of a photovoltaic (PV) system is affected by various environmental factors. These include water droplets, shading conditions, birds’ droppings, and the accumulation of dust. Thus, Jogunuri [10] argue that to ensure the efficient performance of a photovoltaic (PV) system, all these must be removed. This defeats the purpose of solar power being a sustainable source of green energy. It is, hence, an ethical conflict of values or interests to delineate solar power from other energy sources.

Outline Plan

• January 2022- collecting and tabulating data from the EUA meteorological department on weather variations of the previous year.
• February 2022- analyzing the solar panel energy outputs in the four stations.
• March 2022- interpolating the results.
• April 2022-writing a report.

Justification

The prediction of the performance of solar photovoltaic systems’ output cannot be effectively achieved, because of the weather conditions such as humidity and dust rate. However, thanks to artificial intelligence, this can be efficiently and easily solved.

References

D. Dudley, “World’s largest solar power plant moves forward in Abu Dhabi with contract award.” Forbes.

M. H. Alomari, J. Adeeb and O. Younis, “Solar photovoltaic power forecasting in jordan using artificial neural networks,” International Journal of Electrical and Computer Engineering (IJECE), vol. 8, no. 1, pp. 497 – 504, 2018, doi: 10.11591/ijece.v8il.

A. Gligor, C.-D. Dumitru and H.-S. Grif, “Artificial intelligence solution for managing a photovoltaic energy production unit,” Procedia Manufacturing, vol. 22, pp. 626-633, 2018, doi: 10.1016/j.promfg.2018.03.091.

J. Son, S. Jong, H. Park and C. Park, “The effect of particulate matter on solar photo-voltaic power generation over the Republic of Korea,” EnvironmentalResearch Letters, vol. 15, no. 8, 2020.

R. Nageem and R. Jayabarat, “Predicting the power output of a grid-connected solar pane using multi-input support vector regression,” Procedia Computer Science, vol. 115, pp. 723-73, 2017.

M. H. Bergin, C. Ghoroi, D. Dixit, J. J. Schaue and D. T. Shinde, “Large reductions in solar energy production due to dust and particulate air pollution,” Environmental Science & Technology Lette, vol. 4, no. 8, pp. 339-34, 2017.

T. Chuluunsaikhan, A. Nasridinov, W. S. Choi, D. B. Choi, S. H. Choi and Y. M. Kim, “Predicting the power output of solar panels based on weather and air pollution features using machine learning,” Journal of Korea Multimedia Society, vol. 24, no. 2, pp. 222-232, 2021, doi: 10.9717/kmms, 2021.

V. S. B. Kurukuru, A. Haque, M. A. Khan, S. Sahoo, A. Malik and F. Blaabjerg, “A review on Artificial Intelligence applications for grid-connected solar photovoltaic systems,” Energies, vol. 14, pp. 1-35, 2021, doi: 10.3390/en14154690.

R. J. Mustafa, M. R. Gomaa, M. Al-Dhaifallah and H. Rezk, “Environmental impacts on the performance of solar photovoltaic systems,” Sustainability, vol. 12, no. 2, p. 2020, doi: 10.3390/su12020608, 2020.

S. K. Jogunuri, “Artificial intelligence methods for solar forecasting for optimum sizing of PV systems: A review,” Research Journal Of Chemistry And Environment, vol. 24, no. 1, pp. 174-180, 2020.

Footnotes

1. D. Dudley, “World’s largest solar power plant moves forward in Abu Dhabi with contract award.” Forbes. Web.
2. M. H. Alomari, J. Adeeb and O. Younis, “Solar photovoltaic power forecasting in jordan using artificial neural networks,” International Journal of Electrical and Computer Engineering (IJECE), vol. 8, no. 1, pp. 497 – 504, 2018, doi: 10.11591/ijece.v8il.
3. A. Gligor, C.-D. Dumitru and H.-S. Grif, “Artificial intelligence solution for managing a photovoltaic energy production unit,” Procedia Manufacturing, vol. 22, pp. 626-633, 2018, doi: 10.1016/j.promfg.2018.03.091.
4. J. Son, S. Jong, H. Park and C. Park, “The effect of particulate matter on solar photo-voltaic power generation over the Republic of Korea,” EnvironmentalResearch Letters, vol. 15, no. 8, 2020.
5. R. Nageem and R. Jayabarat, “Predicting the power output of a grid-connected solar pane using multi-input support vector regression,” Procedia Computer Science, vol. 115, pp. 723-73, 2017.
6. M. H. Bergin, C. Ghoroi, D. Dixit, J. J. Schaue and D. T. Shinde, “Large reductions in solar energy production due to dust and particulate air pollution,” Environmental Science & Technology Lette, vol. 4, no. 8, pp. 339-34, 2017.
7. T. Chuluunsaikhan, A. Nasridinov, W. S. Choi, D. B. Choi, S. H. Choi and Y. M. Kim, “Predicting the power output of solar panels based on weather and air pollution features using machine learning,” Journal of Korea Multimedia Society, vol. 24, no. 2, pp. 222-232, 2021, doi: 10.9717/kmms, 2021.
8. V. S. B. Kurukuru, A. Haque, M. A. Khan, S. Sahoo, A. Malik and F. Blaabjerg, “A review on Artificial Intelligence applications for grid-connected solar photovoltaic systems,” Energies, vol. 14, pp. 1-35, 2021, doi: 10.3390/en14154690.
9. R. J. Mustafa, M. R. Gomaa, M. Al-Dhaifallah and H. Rezk, “Environmental impacts on the performance of solar photovoltaic systems,” Sustainability, vol. 12, no. 2, p. 2020, doi: 10.3390/su12020608, 2020.
10. S. K. Jogunuri, “Artificial intelligence methods for solar forecasting for optimum sizing of PV systems: A review,” Research Journal Of Chemistry And Environment, vol. 24, no. 1, pp. 174-180, 2020.