Conference proceedings

ARDUINO BASED AUTOMATIC SOLAR PANEL DUST DISPOSITION 
ESTIMATION AND CLOUD BASED REPORTING 
Hebatullah Malik, Maha Alsabban, S. M. Qaisar†
Electrical and Computer Engineering Department, Effat University, Jeddah, 21478, Saudi Arabia [e-mail]†
Abstract.
Recent technological encroachments have evolved the usage of Internet of Things (IoT) and it is becoming an 
elementary part of our life. Regarding the photovoltaic (PV) renewable energy based systems there is a lot of potential to integrate the 
IoT based solutions in order to enhance the generation capabilities and to diminish the energy losses in the electricity production. 
Energy losses can occur due to several causes in the case of PVs. This work deals with the photovoltaic energy losses caused by the 
dust deposition. The idea is to intelligently employ the IoT framework in this regards. The system is based on a modular design 
approach. Each module digitizes the status of a concerned PV panel by using an embedded front-end controller. The readings are 
conveyed to a specifically developed automatic maintenance decision algorithm. 
 Conclusion 
In this paper a method is proposed for automatically identifying the dust disposition on photovoltaic panels. The 
luminosity sensing is embedded in the system to enhance the decision support as a function of different weather and 
daytime conditions. It suppresses the generation of false maintenance alarms. Furthermore, the intended photovoltaic 
module status is successfully logged on the cloud via the Mathworks based “thingspeakwrite” function. The authenticity 
of the cloud based log is confirmed with the “thingspeakread” function and data analysis. This data logging opens further 
post analysis and decision support possibilities. Additionally, based on the instantaneous luminance and open circuit 
voltage observations the features system automatically notifies the maintenance responsible via Email in a real-time 
manner. It allows them to automatically know about the concerned PV panel’s identity. In this fashion an efficient PV 
panel’s maintenance can be realized. It makes the most of the PVs over the system lifecycle. A system prototype is 
successfully implemented and tested. Experimental results have confirmed proper system functionality. The devised 
system is easily scalable. A future work is to add more wireless sensors in the front-end electronics module. It will allow 
monitoring other potential parameters of intended PV panels in order to further enhance system performance in terms of 
automatic fault identification and mitigation. Embedding the event-driven sensing based solutions can further augment 
the system effectiveness [13]-[15]. Investigation based on this axis is another prospect. Another aspect is to employ the 
cloud based logged data to forecast the system behavior by using the modern machine learning algorithms.