Predicting the Impact of Dust on PV Panel Productivity Using AI Algorithms

Renewable energy production became one of the main stream of energy production globally. Solar photovoltaic (PV) occupies a significant role in producing clean electrical energy. The more dust buildup on PV panels, the less efficiency of the energy produced. In this paper, an AI-based model has been proposed to assess and mitigate the impact of the dust on PV panels. Three common types of the panels, namely, monocrystalline, polycrystalline, and thin-film panels, have been utilized in the experiment. The experiment was conducted in an arid and semi-arid region in Jordan and focused on comparing the performance of the cleaned and uncleaned modules. Our results show that thin-film panels suffered from the highest losses, while monocrystalline modules have the lowest. Daily losses increase remarkably when neglecting dust accumulation. The relative power loss is the target in our AI-proposed model. In order to enhance the model accuracy, an accepted level of irradiance should be determined and outlier should be handled. The planned cleaning strategy guided by the AI model resulted in enhancing energy production and lowering overall maintenance cost. The promising results of the proposed AI model can provide a robust framework for optimizing PV maintenance schedules.