Data-driven cleaning optimisation strategy for multi-technology PV systems in the higher education sector in arid climate: A case study perspective in MENA region

The accumulation of dust and other contaminants on photovoltaic (PV) panels is a multifactorial process that significantly affects system performance. While cleaning is vital to maintaining energy output and efficiency, its methods, frequency, and procedures also influence environmental impact, resource use, and operational costs.
This study investigates the effects of cleaning protocols on PV energy generation at the Applied Science University (ASU) campus in Amman, Jordan, addressing challenges faced by higher education institutions (HEIs) in the Middle East and North Africa (MENA) region. A controlled intervention was implemented on eight PV arrays with different technologies and installation configurations over a 19-week period. Machine learning techniques were applied for data imputation, and Analysis of Covariance (ANCOVA) was used to assess the significance of cleaning interventions on energy performance.
The findings demonstrate that uniform cleaning schedules are suboptimal, as different PV technologies and orientations exhibit varying responses to maintenance interventions. The study underscores the importance of customised cleaning strategies that account for technological type and system configuration to maximise power generation and efficiency. These results provide valuable insights for developing sustainable PV maintenance frameworks for HEIs and other institutions operating in arid climates across the MENA region.