A stacking ensemble classifier-based machine learning model for classifying pollution sources on photovoltaic panels.

Area of Science:

• Renewable Energy Systems
• Machine Learning Applications
• Environmental Science

Background:

• Solar photovoltaic (PV) panels are crucial for clean energy generation.
• Surface pollution on PV panels significantly reduces their efficiency by affecting solar radiation, transmittance, and temperature.
• Maintaining PV panel cleanliness is essential for optimal energy output and system longevity.

Purpose of the Study:

• To develop and evaluate a robust machine learning model for identifying diverse pollution sources on solar panels.
• To enhance the accuracy and reliability of pollution detection for improved PV panel performance.
• To provide a data-driven approach for proactive maintenance of solar energy systems.

Main Methods:

• A stacking ensemble classifier was developed, integrating gradient boost, extra tree, and random forest algorithms.
• The extra tree classifier was utilized as a meta-learner for improved predictive performance.
• The model was trained using data on various pollution types and weather features, including irradiance and temperature.

Main Results:

• The proposed stacking ensemble model achieved a high accuracy score of 97.37% in classifying pollution sources.
• The model demonstrated superior performance compared to existing state-of-the-art machine learning models.
• Accurate identification of pollution sources leads to increased power generation efficiency and extended PV panel lifespan.

Conclusions:

• The developed machine learning model offers an effective solution for identifying solar panel pollution.
• This approach enables targeted maintenance, ensuring PV panels operate at peak efficiency.
• The findings contribute to enhancing the overall efficiency, reliability, and sustainability of solar energy systems.