A Yolo Driven Automated PV Health Monitoring Cleaning System with Combined IOT Sensor Fusion

Abstract

The PV solar energy development has already emerged as both framework and keystone to sustainable development in 21st century which is being offered as a boon for a clean pathway of renewable sustainable source amid mounting climate change concerns. With rapid advancement in solar panel efficiency and declining product cost and its widespread adoption in developing nations, its pivotal role in shaping the global economy with transition to low carbon economy changed drastically. However, with the present scenario of rapid climate change and rising global energy demand concerns are being raised regarding the efficiency of solar panels and susceptibility to major factors of environment including dust accumulation, humidity, temperature fluctuation and reduced light intensity. In view of the prevailing condition and present circumstance with alignment to global current trend the following research paper presents and exhibits a novel solution of real time intelligent automated system of PV panel health monitoring and automated cleaning utilizing multimodal data fusion algorithms with embedded IOT control architecture. The proposed solution exhibits panel dirt detection and cleaning system as smart solution towards panel efficiency optimal maintenance through automated monitoring. The following system sites and classifies panel cleanliness in 2 categories namely clean and dirty using lightweight (YOLOv5) model with a webcam feed as aid for capturing and feeding visual input. As complementary to this scenario sensor data with embedded Arduino architecture is being captured with increased reliability and both former and latter data is being fused via Large Language model (LLM) to improve prediction. By this current approach of fusing visual data with sensor input, the system accurately determines when a panel is dirty or not. Upon detection, it alerts the user and can automatically trigger a cleaning mechanism via a relay module. By minimizing manual inspection and mitigating energy loss concerns this method of approach enhances its contribution towards more sustainable solar power systems though timely and intelligent data driven maintenance strategy ultimately promoting sustainable integration at global energy landscape

Introduction

Solar panels, or photovoltaic (PV) modules, are vital to the global shift toward clean and sustainable energy. They convert sunlight into electricity using the photovoltaic effect. Their decentralized and eco-friendly nature makes them suitable for both urban and remote areas. However, real-world deployment faces challenges, particularly soiling (dust, dirt, bird droppings), which can reduce efficiency by up to 30%.

2. Problem Statement:

Traditional manual cleaning methods are labor-intensive, water-wasteful, and not viable for large installations. Current IoT-based monitoring (using sensors like humidity, temperature, current, and voltage) lack visual feedback, leading to false negatives (i.e., panels appear functional despite partial obstruction).

3. Proposed Solution: Hybrid AI-IoT Framework:

A hybrid embedded system is proposed that integrates:

• IoT sensing (via Arduino and ESP8266) for real-time performance/environmental data.

• Computer vision using a lightweight YOLOv5 model deployed on Raspberry Pi 4 to detect visual anomalies like dust and cracks.

• Autonomous cleaning through a relay-controlled water pump, activated only when both sensor data and visual analysis confirm inefficiency.

Key Features:

• Sensor data and camera images are fused through serial communication between Arduino and Raspberry Pi.

• A lightweight LLM (Language Model) on Raspberry Pi interprets combined data and decides on cleaning action.

• The system is cost-effective, energy-efficient, and requires minimal human intervention.

4. System Design and Components:

• Sensors: Light, humidity, temperature, voltage, current.

• Microcontrollers: Arduino for sensor input, Raspberry Pi for AI processing and decision-making.

• Communication: ESP8266, TCP/IP stack, cloud platforms like Blynk/Ubidots.

• Visual Processing: YOLOv5 model trained on solar panel imagery to detect dust, cracks, or bird droppings.

• Actuation: Submersible water pump triggered by relay on confirmed cleaning need.

5. Literature Review Highlights:

• Various research works use IoT for real-time monitoring with platforms like NodeMCU, Raspberry Pi, and cloud interfaces.

• YOLO-based vision systems are effective for detecting soiling, hot spots, and physical anomalies using thermal and visual imaging.

• Robotic cleaning systems integrated with IoT enable precision cleaning and are scalable for industrial solar farms.

• Systems have been implemented using stepper motors, servo motors, and relay-controlled actuators for automated dust removal.

6. Contributions of the Research:

• End-to-end smart solar maintenance system combining real-time monitoring, image-based diagnostics, and autonomous cleaning.

• Context-aware cleaning based on both sensor anomalies and image detection—improving decision accuracy and reducing unnecessary cleanings.

• Use of open-source hardware and software, making the system low-cost and scalable.

• Implementation of long-term data logging and visualization using tools like Jupyter Notebook, R Studio, and Excel for predictive maintenance.

Conclusion

Through the seamless integration of edge AI using YOLOv8, automated cleaning control, and IoT-based sensor fusion, the proposed technology offers a complete, intelligent, and affordable solution for autonomous solar panel maintenance. The system efficiently tracks electrical and environmental data in real time, such as temperature, humidity, light intensity, voltage, current, and power, using sensors like DHT11, LDR, and INA219 in addition to a Raspberry Pi and Pi Camera. To evaluate the health of the panel surface, the recorded data is combined with computer vision-based brightness analysis and picture classification using a blank YOLOv8 model. Exact decision-making is made possible by the fusion logic, which only initiates cleaning in confirmed dusty or performance-degrading circumstances. Accurate detection, prompt cleaning activation, and dependable power restoration after cleaning are all demonstrated by experimental data recorded and displayed using ThingSpeak, confirming the efficacy of this intelligent monitoring framework over a range of time periods, including morning, afternoon, and evening. This system\'s installation not only lessens the need for manual inspections and energy losses brought on by dust buildup, but it also shows how microcontroller-based automation and edge computing driven by AI may be combined to create sustainable energy systems. The system is ideal for off-grid applications and scalable smart solar infrastructure in real-world deployments due to its modularity, low power consumption, and remote visualization capabilities.

References

[1] K. Ranabhat, L. Patrikeev, A. A. Revina and K. Andrianov, “An introduction to solar cell technology,” vol. 14, pp. 481 - 491, 15 December 2016. [2] A. O. Maka and J. M. Alabid, “Solar energy technology and its roles in sustainable development,” Clean Energy, vol. 6, no. 3, p. 476–483, 11 June 2022 [3] K. Vyas, “Solar Farming with Agricultural land,” ACTA SCIENTIFIC AGRICULTURE, vol. 3, pp. 23-25, 10 October 2019. [4] D. A. Widodo, P. Purwanto and H. , “Potential of Solar Energy in Residential Rooftop Surface Area in Semarang City, Indonesia,” Advances in Science, Technology and Engineering Systems, vol. 5, pp. 397-404, 2020. [5] W. Javed, B. Guo, Y. Wubulikasimu and B. W. , “Photovoltaic performance degradation due to soiling and characterization of the accumulated dust,” in 2016 IEEE International Conference on Power and Renewable Energy, ICPRE 2016, Shanghai, 2017. [6] L. A. Lamont and L. E. Chaar, “Automatic Versus Manual Solar Panel Cleaning for Remote Locations,” Journal of Technology Innovations in Renewable Energy, vol. 2, pp. 98-105, 2013. [7] K. Srinivasan , D. Vimaladevi and D. Chakravarthi , “Solar Energy Monitoring System by IOT,” International Journal of Advanced Networking and Application (IJANA), pp. 46-51, January 2019. [8] M. Gopal, T. C. Prakash, N. V. Ramakrishna and B. Prabhanjan, “IoT Based Solar Power Monitoring System,” IOP Conf. Series: Materials Science and Engineering, vol. 981, pp. 1-7, 19 December 2020. [9] B. S. Thrinath, B. L. Prasanna, J. Powl.S, N. S. Reddy and N. K. , “Real Time Monitoring and Analysis of String Faults in Solar Plant Using IoT,” in 3rd International Conference on Sustainable Green Energy Technologies (ICSGET 2025), 2024. [10] J. Huang, K. Zeng, Z. Zhang and W. Zhong, “Solar panel defect detection design based on YOLO v5 algorithm,” Heliyon, vol. 11, pp. 1-12, 2023. [11] R. H. P. Putra, D. Wahyudin and T. Sucita, “Designing Energy and Power Monitoring System on Solar Power Plant Using Raspberry Pi,” IOP Conf. Series: Materials Science and Engineering, no. IOP Publishing Ltd, pp. 1-7, 2018. [12] C. M. Nkinyam, C. O. Ujah, K. C. Nnakwo, O. Ezeudu, D. V. V. Kallon and I. I.-E. C. , “Design and implementation of a waterless solar panel cleaning system,” Unconventional Resources, pp. 1-11, 2025. [13] D. Jadhav , V. Muddebhalkar and A. Korke, “Dust Cleaner System for PV Panel using IoT,” International Journal of Computer Applications (0975 – 8887), vol. 178, pp. 31-34, 10 May 2019. [14] V. S. Priya and . S. Brindha, “Optimizing Solar Power Generation and Integration in Automotive Systems Through IoT,” Indian Journal of Science and Technology, vol. 17, no. 40, pp. 4158-4167, 2024. [15] S. Ajaykumar, P. Sandeep, D. Kiran, B. T. Venakatasai and M. K. Venkateswaro, “DESIGN AND DEVELOPMENT OF AUTOMATIC SOLAR PANEL CLEANING SYSTEM BASED ON ARDUINO,” International Research Journal of Modernization in Engineering Technology and Science, vol. 6, no. 4, pp. 7241-7249, April 2024. [16] Y. Bikrat , K. Salmi, A. Benlghazi, A. Benali and D. Moussaid, “A Photovoltaic Wireless Monitoring System,” in International Symposium on Advanced Electrical and Communication Technologies (ISAECT), 2018. [17] S. H. Cheragee, N. Hassan, S. Ahammed and A. Z. M. , “A STUDY OF IOT BASED REAL-TIME SOLAR POWER REMOTE MONITORING SYSTEM,” International Journal of Ambient Systems and Applications (IJASA), vol. 9, pp. 27-36, June 2021. [18] V. Kavitha and V. Malathi, “A SMART SOLAR PV MONITORING SYSTEM USING IOT,” in First International Conference on Secure Reconfigurable Architectures & Intelligent Computing, 2019. [19] J. E. Shaji, B. Varghese and R. V. , “A Health Care Monitoring System with Wireless Body Area Network using IoT,” International Journal of Recent Trends in Engineering & Research (IJRTER), vol. 3, no. 11, pp. 112-117, November 2017. [20] N. A. Othman, M. R. Zainodin, N. Anuar and N. S. Damanhuri, “Remote Monitoring System Development via Raspberry- Pi for Small Scale Standalone PV Plant,” in 2017 7th IEEE International Conference on Control System, Computing and Engineering (ICCSCE), Penang, Malaysia, 2017. [21] H. Eteruddin, D. Setiawan and A. Atmam, “Web Based Raspberry Monitoring System Solar Energy Power Plant,” in IOP Conference Series Earth and Environmental Science, 2020. [22] H. Ozbay, A. Karafil, Y. Onal, M. Kesler and H. Parmaksiz, “The Monitoring of Monthly, Seasonal and Yearly Optimum Tilt Angles by Raspberry Pi Card for Bilecik City, Turkey,” Energy Procedia, vol. 113, pp. 313-318, 1 May 2017. [23] R. S. Sarban Singh and M. F. Abbod, “Research and Integration of IoT based Solar Photovoltaic Panel Health Monitoring System,” Indian Journal of Public Health Research & Development, vol. 9, pp. 1751-1757, 1 December 2018. [24] I. M. Shirbhate and S. S. Barve, “Solar panel monitoring and energy prediction for smart solar system,” International Journal of Advances in Applied Sciences (IJAAS), vol. 8, pp. 136-142, June 2019. [25] U. Naeem, K. Chadda, S. Vahaji, J. Ahmad, X. Li and E. Asadi, “Aerial Imaging-Based Soiling Detection System for Solar Photovoltaic Panel Cleanliness Inspection,” Sensors, vol. 25, 25 January 2025. [26] R. Khanam, T. Asghar and M. Hussain, “Comparative Performance Evaluation of YOLOv5, YOLOv8,and YOLOv11for Solar Panel Defect Detection,” Solar, vol. 5, 2025. [27] M. Z. A. Hamid, K. Daud, Z. H. C. Soh, M. K. Osman, I. S. Isa and M. S. Jadin, “Deep Learning-Driven Thermal Imaging Hotspot Detection in Solar Photovoltaic Arrays Using YOLOv10,” ESTEEM Academic Journal, vol. 20, pp. 77-90, 30 September 2024. [28] S. Singarapu, K. Swaraja and M. Kirola, “Smart Iot Based Solar Panel Cleaning System,” E3S Web of Conferences, vol. 430, 6 October 2023. [29] M. M S, D. Rajinikanth, A. B, C. K G, G. C M and S. N R, “IoT Based Solar Tracking System with Automatic Dust Monitoring System,” International Research Journal on Advanced Engineering Hub (IRJAEH), vol. 2, pp. 1838-1844, 25 June 2024. [30] M. Hamid, M. F. Simamora, M. D. E. Vania and P. Cholillah, “Design of Automatic Cleaning System on Solar Panel Using IoT-based Wiper,” Journal of Physics Conference Series, vol. 2733, 1 March 2024. [31] N. Maindad, A. Gadhave, S. Satpute and B. Nanda, “Automatic Solar Panel Cleaning System,” in International Conference on Communication and Information Processing Automatic Solar Panel Cleaning System, 2021. [32] G. Anilkumar, K. Naveen, M. T. V. S. H. K. Pavan Kumar, G. V. B. Kumar, K. Palanisamy, A. Patil and D. Sachan, “Design and development of wireless networking for solar PV panel cleaning robots,” in IOP Conference Series Materials Science and Engineering, 2020. [33] N. Khadka, A. Bista, B. Adhikari, A. Shrestha and D. Bista, “Smart solar photovoltaic panel cleaning system,” IOP Conference Series: Earth and Environmental Science, vol. 463, 7 April 2020. [34] F. H. M. Noh, M. F. Yaakub, I. N. A. M. Nordin, N. Sahari, N. A. Zambri, S. Sy Yi and M. S. M. Saibon, “Development of solar panel cleaning robot using Arduino,” Indonesian Journal of Electrical Engineering and Computer Science, vol. 19, pp. 1245-1250, 1 September 2020. [35] R. M. Ramli and W. A. Jabbar, “Design and implementation of solar-powered with IoT-Enabled portable irrigation system,” Internet of Things and Cyber-Physical Systems, pp. 212-225, 1 December 2022. [36] R. H. P. Putra, D. Wahyudin and T. Sucita, “Designing Energy and Power Monitoring System on Solar,” IOP Conf. Series: Materials Science and Engineering(ISMEE), vol. 384, pp. 1-7, 2018.

Copyright

Copyright © 2025 Amal Sankar Ram T S, Htet Aung Hlaing. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.