Design and implementation of a LoRa-enabled IoT system for photovoltaic monitoring and machine learning-based power estimation
DOI:
https://doi.org/10.54939/1859-1043.j.mst.114.2026.12-21Keywords:
Photovoltaic (PV) systems; Machine learning; IoT; LoRa; Power estimation.Abstract
Photovoltaic (PV) systems are increasingly deployed as a sustainable energy solution; however, because they strongly depend on environmental conditions and time-related factors, their output power varies significantly in real-world operation. This study proposes a low-cost IoT-based PV monitoring system integrated with LoRa wireless communication to estimate power and support anomaly detection. Machine learning regression models were first evaluated using a public dataset, incorporating both current environmental features and time-related features such as hour of the day and lag variables to capture the influence of previous conditions and environmental inertia on the panels. The predicted power was then compared with measured values to quantify errors and detect anomalies. The results show that Random Forest provides more stable prediction performance than Linear Regression, highlighting the importance of nonlinear models in capturing complex environmental relationships. Based on these findings, the models are considered references for future power prediction using data collected from the proposed standalone PV monitoring system, in which an STM32F103C8T6 microcontroller is used at the LoRa sensor node to acquire measurements from the environmental and electrical sensors, while an ESP32 microcontroller acts as the LoRa gateway to forward the data to a cloud platform. These results support the feasibility of applying machine learning-based power prediction to the proposed IoT-Based intelligent PV monitoring system.
References
[1]. https://www.iea.org/commentaries/access-to-electricity-stagnates-leaving-globally-730-million-in-the-dark (2025).
[2]. Z. M. Alaas, "The Effects of Temperature on Photovoltaic and Different Mitigation Techniques: A Review", IEEE Access, vol. 12, pp. 180309-180327, (2024). DOI: 10.1109/ACCESS.2024.3504009
[3]. A. Abdessadak, H. Ghennioui, B. El Bhiri, N. Thirion-Moreau, M. Abraim, S. Merzouk, "Assessing the Effects of Dust on Solar Panel Performance: A Comprehensive Review and Future Directions", Eng. Proc., vol. 112, no. 9, (2025). https://doi.org/10.3390/engproc2025112009
[4]. H. S. Ahmed, A. J. Abid, A. A. Obed, "Mitigating Partial Shading Effects in Photovoltaic Arrays using a Switch Matrix: Design, Modeling, and Performance Evaluation", 2023 Sixth International Conference on Vocational Education and Electrical Engineering (ICVEE), pp. 202-208, (2023).
[5]. L. Lillo-Sánchez, G. López-Lara, Jonathan Vera, E. Pérez-Aparicio, Isidoro Lillo, "Degradation analysis of photovoltaic modules after operating for 22 years. A case study with comparisons", Solar Energy, vol. 222, pp. 84-94, (2021). DOI: 10.1016/j.solener.2021.04.026
[6]. D. Witczak, S. Szymoniak, "Review of Monitoring and Control Systems Based on Internet of Things", Appl. Sci., vol. 14, no. 8943, (2024). https://doi.org/10.3390/app14198943
[7]. J. M. Paredes-Parra, A. J. García-Sánchez, A. Mateo-Aroca, Á. Molina-García, "An Alternative Internet-of-Things Solution Based on LoRa for PV Power Plants: Data Monitoring and Management", Energies, vol. 12, no. 881, (2019). https://doi.org/10.3390/en12050881
[8]. B. H. Hameed, S. Kurnaz, "Secure low-cost photovoltaic monitoring system based on LoRaWAN network and artificial intelligence", Discov Computing, vol. 27, no. 36, (2024).
[9]. Ahzam Shadab, "Spatial Forecasting of Solar Radiation using ARIMA Model", Remote Sensing Applications Society and Environment, vol. 20, (2020). DOI: 10.1016/j.rsase.2020.100427
[10]. E. Chodakowska, J. Nazarko, Ł. Nazarko, H. S. Rabayah, R. M. Abendeh, R. Alawneh, "ARIMA Models in Solar Radiation Forecasting in Different Geographic Locations", Energies, vol. 16, no. 5029, (2023). https://doi.org/10.3390/en16135029
[11]. K. J. Iheanetu, "Solar Photovoltaic Power Forecasting: A Review", Sustainability, vol. 14, no. 17005, (2022).
[12]. Younes Ledmaoui, Adila Maghraoui, Mohamed Aroussi, Saadane Rachid, Ahmed Chebak, Abdellah Chehri, "Forecasting solar energy production: A comparative study of machine learning algorithms", Energy Reports, vol. 10, pp. 0-8, (2023).
[13]. S. Cantillo-Luna, R. Moreno-Chuquen, D. Celeita, G. Anders, "Deep and Machine Learning Models to Forecast Photovoltaic Power Generation", Energies, vol. 16, no. 4097, (2023).
[14]. Y. Essam, A. N. Ahmed, R. Ramli, K. W. Chau, M. S. Idris Ibrahim, M. Sherif, A. El-Shafie, "Investigating photovoltaic solar power output forecasting using machine learning algorithms", Engineering Applications of Computational Fluid Mechanics, vol. 16, no. 1, pp. 2002–2034, (2022).
[15]. Kossoko Didavi, Richard Agbokpanzo, Macaire Agbomahena, "LSTM and XGBoost models for 24-hour photovoltaic power forecasting from direct irradiation data", (2023).
[16]. A. Anik, "Solar Power Generation Data", (2019). https://www.kaggle.com/datasets/anikannal/solar-power-generation-data
[17]. D.-A. T. Le et al., "Patient Remote Monitoring System Using MQTT Protocol for ECG Signals," 2024 9th International Conference on Integrated Circuits, Design, and Verification (ICDV), Hanoi, Vietnam, 2024, pp. 114–118, doi: 10.1109/ICDV61346.2024.10617252.
[18]. V. N. Pham, V. T. Nguyen, T. C. Trinh, and T. H. Tran, “Multi-domain feature-based early detection of bearing faults using MLP classifier on NASA IMS dataset”, Journal of Military Science and Technology, vol. 106, no. 106, pp. 48–54, Oct. 2025, doi: 10.54939/1859-1043.j.mst.106.2025.48-54.
[19]. V. N. Pham, T. H. Nguyen, V. H. Pham, V. M. Pham, and T. H. Tran, “Designing an ECG signal measurement and analysis system using SVM algorithm for arrhythmia classification”, Journal of Military Science and Technology, vol. 105, no. 105, pp. 36–43, Aug. 2025, doi: 10.54939/1859-1043.j.mst.105.2025.36-43.
