The SoDeCo project, carried out from 1 June 2020 to 31 May 2022 with a cost‑neutral extension until 31 December 2022, aimed to create a sensor network that detects soiling on photovoltaic modules in semi‑arid North African climates and to optimise cleaning strategies accordingly. The consortium comprised German and Moroccan researchers, with the Hochschule Anhalt coordinating the scientific and organisational work. Three staff members from the German partner handled procurement, coordination and administration, while the Moroccan partner supplied local expertise and managed on‑site activities. Regular virtual project meetings were held throughout the programme, with the official start on 21 January 2021 and the final internal meeting on 31 May 2022.
Technically, the first work package focused on the design and prototyping of the sensor network. The concept was iteratively refined to include sensors for temperature, relative humidity, air quality, particle distribution, module electrical power and module temperature. Technical specifications such as measurement accuracy, range, power consumption and interface compatibility were defined, and the best available sensor technologies were selected. Prototypes were fabricated using high‑quality electronic components and microcontrollers, and their performance was evaluated in laboratory conditions before deployment.
The second work package carried out field tests at multiple sites in Morocco. The sensor arrays were installed on photovoltaic modules and monitored over several months. The data collected allowed a reliable assessment of soiling levels and their impact on module performance. By comparing on‑site measurements with laboratory analyses, the team confirmed that the sensor network could detect soiling‑induced power losses with sufficient sensitivity to trigger maintenance actions.
Work package three investigated cleaning methods. Various mechanical and chemical cleaning techniques were applied to the modules, and their effectiveness was quantified using the sensor data. The results showed that targeted cleaning, guided by real‑time soiling information, could reduce cleaning frequency while maintaining high energy yield. This finding underpins the development of a sensor‑based cleaning metric in work package four, which predicts optimal cleaning intervals based on continuous environmental monitoring and module performance trends.
The fifth work package addressed the commercialisation of the technology. Market research and analysis of existing soiling‑detection products informed a revised exploitation concept. The new plan emphasises integration of the sensor network and cleaning metric into existing monitoring systems or deployment as a standalone solution, thereby enhancing the value proposition for photovoltaic operators in Morocco and beyond.
Overall, the project demonstrated that a compact, low‑power sensor network can reliably monitor soiling in harsh, semi‑arid environments and that data‑driven cleaning schedules can maximise energy output while minimising labour and water consumption. Future research directions identified by the consortium include optimisation of sensor technology, incorporation of artificial intelligence and machine learning for predictive maintenance, and assessment of long‑term stability and economic impact. The results provide a solid foundation for scaling the solution to larger solar farms and for further development of sustainable photovoltaic operation in arid regions.