What is it about?

This study uses artificial intelligence and machine learning to predict how much solar power a building’s photovoltaic (PV) system will produce at different times, such as one day, one week, or one month ahead. Unlike most studies that predict only one output, this research uses a multi-label approach, forecasting both DC (PV array) and AC (grid power) outputs at once. Real data were collected from a solar plant in Cairo, Egypt, and tested with several AI models. The results show that decision tree, deep learning, and random forest models achieved the highest accuracy.

Featured Image

Why is it important?

Accurate forecasting of solar power generation is essential for maintaining grid stability and improving energy management. By predicting both PV and AC power over multiple time horizons, this study provides reliable tools for optimizing inverter performance and integrating solar energy into the power grid. The proposed multi-label machine learning models help operators plan for energy fluctuations, enhance power dispatch efficiency, and support the wider adoption of renewable energy systems.

Perspectives

Future work may focus on extending the proposed multi-label forecasting framework to larger PV plants and different climatic regions to validate model generalization. Incorporating additional meteorological parameters such as humidity and solar angle could further enhance prediction accuracy. Moreover, integrating the developed models with real-time control and energy management systems would enable adaptive grid operation and improved renewable energy utilization.

Dr. Amal A. Hassan
Electronics Research Institute

Read the Original

This page is a summary of: Multi-label machine learning for power forecasting of a grid-connected photovoltaic solar plant over multiple time horizons, Scientific Reports, September 2025, Springer Science + Business Media,
DOI: 10.1038/s41598-025-20251-y.
You can read the full text:

Read

Contributors

The following have contributed to this page