Smart and cost-effective optimisation of photovoltaic cleaning schedules

Periodic cleaning schedules for large photovoltaic installations often result in unnecessary costs when panels are already clean or energy losses due to unexpected dust accumulation between cleanings. Manual inspections to address this issue are impractical given the vast number of panels. While recent research on cleaning schedule optimisation relies heavily on AI-based methods, these approaches require extensive datasets and still fail to guarantee optimal efficiency. This paper proposes a novel optimisation framework to enhance PV cleaning schedules without relying on large datasets or incurring additional costs. The approach combines three key strategies: adaptation of emerging Maximum Power Point Tracking (MPPT) techniques, metaheuristic optimisation, and an intelligent cleaning score mechanism (ICSM). This method bridges the gap between MPPT and optimal cleaning scheduling by framing it as a two-stage optimisation problem, where the first stage is performed in real-time and the second is conducted offline. The solution can be seamlessly incorporated into any existing MPPT method without additional costs, offering an economical and efficient way to improve PV system performance. Experimental tests confirm the effectiveness of the proposed method in reducing cleaning frequency while maintaining high energy performance, achieving 98.4 % accuracy with the Grey Wolf Optimizer and outperforming three other algorithms.<p></p>