DEWA study assesses PV cleaning robot performance and safety

A research team from the United Arab Emirates’ Dubai Electricity and Water Authority (DEWA) has evaluated the cleaning performance of several PV cleaning robots in a desert environment. The team also outlined operation and maintenance recommendations for stakeholders based on the results.

Bhaduri added that the team evaluated the durability and operational reliability of the robots themselves under real desert conditions, identifying practical vulnerabilities associated with high temperatures, batteries, ingress protection, electronics, mechanical alignment and material degradation. “Based on these field observations and industry feedback, we further proposed a practical checklist for evaluating robotic PV cleaners before deployment,” she added.

Overall, the study evaluated five autonomous dry-cleaning robots. Robots A and B used bristle-based brushes and operated on fixed-tilt PV systems in a two-in-portrait configuration. Robot A weighed 48 kg, while Robot B weighed 45 kg. Robots C, D and E were deployed on single-axis trackers in a one-in-portrait configuration and weighed less than 40 kg. Robots C and E used microfiber cloth-based cleaning systems, while Robot D used bristles. Robots C and D were produced by the same manufacturer, allowing a direct comparison of the two cleaning materials. The systems were manufactured in China, India and Greece by undisclosed companies.

All robots were tested at the DEWA Cleaning Test Facility, located next to the Mohammed bin Rashid Al Maktoum Solar Park in Dubai. The test field comprised 164 crystalline-silicon PV modules rated between 445 W and 505 W, including monofacial PERC glass-backsheet modules and bifacial PERC glass-glass modules. Outdoor performance monitoring was carried out from July 22, 2024, to August 26, 2025, with occasional laboratory I–V, reflectance, electroluminescence (EL), dark lock-in thermography (DLIT) and microscopy measurements.

“Daily robotic cleaning reduced soiling rates from approximately 0.14–0.33% per day to below 0.02% per day,” Bhaduri said. “The evaluated robots achieved cleaning efficiencies of 69–99%, while EL and DLIT testing found no robot-induced cell cracks or electrical defects over two years.”

Bhaduri added, however, that repeated cleaning caused minor wear to the anti-reflective coating, with greater abrasion observed for bristle-based cleaning than for cloth-based cleaning. “We have further identified important safety and reliability concerns, including battery overheating, inadequate ingress protection, corrosion, frame misalignment and UV degradation of polymer components,” she added.

In conclusion, the group proposed a 12-point checklist for evaluating PV cleaning robots before deployment. Key recommendations include using IEC 62133-certified batteries with thermal management, ensuring at least IP65/IP66 protection for electronic components, using UV- and corrosion-resistant materials, conducting PV-specific abrasion testing, integrating SCADA-based fault and cleaning verification, and monitoring module misalignment and frame deformation.

The study “Performance and safety evaluation of PV cleaning robots in desert environment” was published in Solar Energy.