A PREDICTIVE MODEL FOR ENERGY EFFICIENCY OF SOLAR POWER PLANTS USING ARTIFICIAL INTELLIGENCE
Abstract
In this study, a prototype artificial intelligence model was developed to predict energy production based on weather parameters, using data from a 100 kW subset of the 510 kW small-scale solar power plant launched under Qarshi State University in Qarshi city.
Among the five solar panel arrays of the station, the most reliable dataset from one array was selected as the basis for analysis. The study integrated the plant’s recorded production data throughout 2024 with corresponding meteorological indicators — temperature, humidity, wind speed, and atmospheric pressure.
The data were cleaned and merged in a Python environment, then trained using the Random Forest Regression algorithm [1]. The model was used to forecast power generation for January–February 2025.
According to the results, the model achieved a Mean Absolute Error (MAE) of 17.5 kW, a Root Mean Square Error (RMSE) of 23.3 kW, and a coefficient of determination (R²) of 0.043.
Humidity (34%) and temperature (33%) were identified as the most influential factors. The findings confirm that meteorological variables significantly affect solar energy generation under Karshi’s climatic conditions and establish a scientific foundation for the development of high-accuracy forecasting systems in future research.
Keywords
Artificial Intelligence, Solar Energy, Qarshi, Random Forest Regression, Energy Forecasting, Weather Data, Data Integration, Meteorological Factors, Data Analysis, Photovoltaic System, AI Modeling.
References
- L. Breiman, “Random Forests,” Machine Learning, vol. 45, no. 1, pp. 5–32, 2001.
- A. Mellit and S. A. Kalogirou, “Artificial intelligence techniques for photovoltaic applications: A review,” Progress in Energy and Combustion Science, vol. 34, pp. 574–632, 2008.
- C. Voyant, G. Notton et al., “Machine learning methods for solar radiation forecasting: A review,” Renewable Energy, vol. 105, pp. 569–582, 2017.
- G. Li and J. Shi, “Application of artificial neural networks for forecasting solar radiation data,” Renewable Energy, vol. 35, no. 12, pp. 2801–2812, 2010.
- M. Khandakar, M. R. Islam, and M. Alam, “Hybrid AI models for solar power prediction based on weather data,” Energy Reports, vol. 6, pp. 1040–1051, 2020.
- A. Diagne et al., “Review of solar irradiance forecasting methods and a proposition for small-scale insular grids,” Renewable and Sustainable Energy Reviews, vol. 27, pp. 65–76, 2013.
- T. Chen and C. Guestrin, “XGBoost: A scalable tree boosting system,” Proc. 22nd ACM SIGKDD, 2016.
- S. Hochreiter and J. Schmidhuber, “Long short-term memory,” Neural Computation, vol. 9, no. 8, pp. 1735–1780, 1997.
- M. Ahmed et al., “Predicting PV output using hybrid deep learning approaches,” Energy Conversion and Management, vol. 252, 2022.
- K. Kim and S. Park, “Optimization of photovoltaic output prediction using ensemble learning,” Energies, vol. 14, no. 3, 2021.
- N. Ozturk and E. Yetis, “Impact of temperature and humidity on PV panel efficiency,” Solar Energy Materials and Solar Cells, vol. 174, pp. 119–127, 2018.
- A. A. Abdullaev and B. S. Kholmatov, “Climate analysis of solar energy in the southern regions of Uzbekistan,” UzEnergy Journal, vol. 4, no. 2, pp. 33–41, 2022.
- Ministry of Energy of Uzbekistan, “Renewable Energy Development Strategy 2030,” Tashkent, 2023.
- Python Software Foundation, “Python Documentation – Pandas, Scikit-learn, Matplotlib,” 2024.
- OpenAI Research, “GPT-5 Technical Capabilities in AI-assisted scientific analysis,” 2025.