Multivariable Internet of Things Monitoring for Off-Grid Photovoltaic Operational Diagnostics
Abstract
Off-grid photovoltaic (PV) systems require low-cost monitoring that captures environmental conditions and energy flows for operational diagnosis. This study evaluates an Internet of Things monitoring framework using 2,500 synchronized observations recorded at 10-minute intervals in May 2025 at a laboratory of Universitas Muhammadiyah Palembang, Indonesia. The objective was to assess whether component-resolved multivariable monitoring could characterize environmental–electrical coupling, identify interpretable operating states, and support operational screening before labeled fault data are available. The monitored system comprised a 1,200-Wp PV array, a 12-V 400-Ah battery bank, two 60-A MPPT controllers, a 3,000-W/VA inverter, and a 350-W water-pump load. Descriptive statistics, Pearson correlation, HC3-robust multiple linear regression, K-means clustering, and power-balance residual analysis were applied. Mean panel temperature, estimated solar irradiance, PV power, charging power, discharging power, and load power were 27.75 °C, 219.77 W/m², 190.14 W, 93.99 W, 98.74 W, and 198.56 W, respectively. PV power correlated strongly with estimated irradiance (r = 0.999) and charging power (r = 0.963), and negatively with discharging power (r = -0.533). The thermal and daytime PV-power models explained 97.7% and 99.5% of within-sample variance, respectively. Residual distributions varied systematically among operating states, supporting their use as internal-consistency and disturbance-sensitive screening indicators rather than definitive fault labels. Because irradiance was estimated from BH1750 illuminance using a published empirical relation without on-site pyranometer calibration, the strong irradiance–PV power correlation indicates internal association rather than validation of absolute irradiance accuracy. The framework therefore provides interpretable operational diagnostics before labeled fault data are available for off-grid PV systems
Keywords
References
A. Mellit, M. Benghanem, S. Kalogirou, and A. Massi Pavan, “An embedded system for remote monitoring and fault diagnosis of photovoltaic arrays using machine learning and the internet of things,” Renew. Energy, vol. 208, pp. 399–408, 2023, doi: 10.1016/j.renene.2023.03.096.
M. I. Shajahan et al., “Edge-integrated IoT framework for real-time fault diagnosis and performance degradation analysis in photovoltaic modules,” Renew. Energy, vol. 258, 2026, doi: 10.1016/j.renene.2025.124928.
M. N. Akhter et al., “A hybrid deep learning method for an hour ahead power output forecasting of three different photovoltaic systems,” Appl. Energy, vol. 307, 2022, doi: 10.1016/j.apenergy.2021.118185.
A. Keddouda et al., “Photovoltaic module temperature prediction using various machine learning algorithms: Performance evaluation,” Appl. Energy, vol. 363, 2024, doi: 10.1016/j.apenergy.2024.123064.
A. Keddouda et al., “Solar photovoltaic power prediction using artificial neural network and multiple regression considering ambient and operating conditions,” Energy Convers. Manag., vol. 288, p. 117186, 2023.
V. Abad-Alcaraz, M. Castilla, J. A. Carballo, J. Bonilla, and J. D. Álvarez, “Multimodal deep learning for solar radiation forecasting,” Appl. Energy, vol. 393, 2025, doi: 10.1016/j.apenergy.2025.126061.
M. Leal et al., “Optical performance analysis of a hybrid parabolic trough collector with photovoltaic slats,” Appl. Opt., vol. 63, no. 34, pp. 8878–8889, 2024, doi: 10.1364/AO.537784.
F. Bayrak, “Prediction of photovoltaic panel cell temperatures: Application of empirical and machine learning models,” Energy, vol. 323, 2025, doi: 10.1016/j.energy.2025.135764.
A. Anisi, M. Sheikholeslami, Z. Khalili, and F. M. Boora, “Optimizing solar panel performance with new algorithm incorporating duct with helical tape and hybrid nanofluid,” J. Taiwan Inst. Chem. Eng., vol. 167, 2025, doi: 10.1016/j.jtice.2024.105908.
A. Carpentieri, D. Folini, J. Leinonen, and A. Meyer, “Extending intraday solar forecast horizons with deep generative models,” Appl. Energy, vol. 377, 2025, doi: 10.1016/j.apenergy.2024.124186.
J. Zou, Z. Wang, Z. Zhu, and Z. Tan, “Explainable and physics-constrained PV power prediction via a hybrid framework Integrating secondary decomposition and improved Transformer-LSTM,” Energy, vol. 347, 2026, doi: 10.1016/j.energy.2026.140314.
S. Shan et al., “SolarFM: Spatio-temporal solar irradiance forecasting based on time series foundation model,” Appl. Energy, vol. 410, 2026, doi: 10.1016/j.apenergy.2026.127532.
W. Zhang, H. Li, M. Wang, W. Lv, J. Huang, and L. Yang, “Enhancing typical Meteorological Year generation for diverse energy systems: A hybrid Sandia-machine learning approach,” Renew. Energy, vol. 225, 2024, doi: 10.1016/j.renene.2024.120369.
Y. Gao et al., “Unlocking predictive insights and interpretability in deep reinforcement learning for Building-Integrated Photovoltaic and Battery (BIPVB) systems,” Appl. Energy, vol. 384, 2025, doi: 10.1016/j.apenergy.2025.125387.
M. Dey, S. P. Rana, C. V Simmons, and S. Dudley, “Solar farm voltage anomaly detection using high-resolution ?PMU data-driven unsupervised machine learning,” Appl. Energy, vol. 303, 2021, doi: 10.1016/j.apenergy.2021.117656.
M. Kheir Abadi, V. Davoodi, M. Deymi-Dashtebayaz, and A. Ebrahimi-Moghadam, “Determining the best scenario for providing electrical, cooling, and hot water consuming of a building with utilizing a novel wind/solar-based hybrid system,” Energy, vol. 273, 2023, doi: 10.1016/j.energy.2023.127239.
S. Kataria, M. Kumar, and A. Sikander, “A novel reconfiguration strategy for PV system under mismatch irradiance with experimental validation and machine learning-driven fault detection and diagnosis,” Renew. Energy, vol. 261, 2026, doi: 10.1016/j.renene.2026.125323.
J. Van Gompel, D. Spina, and C. Develder, “Satellite based fault diagnosis of photovoltaic systems using recurrent neural networks,” Appl. Energy, vol. 305, no. June 2021, p. 117874, 2022, doi: 10.1016/j.apenergy.2021.117874.
P. Arévalo, D. Benavides, M. Tostado-Véliz, J. A. Aguado, and F. Jurado, “Smart monitoring method for photovoltaic systems and failure control based on power smoothing techniques,” Renew. Energy, vol. 205, pp. 366–383, 2023, doi: 10.1016/j.renene.2023.01.059.
A. Cano, P. Arévalo, D. Benavides, and F. Jurado, “Comparative analysis of HESS (battery/supercapacitor) for power smoothing of PV/HKT, simulation and experimental analysis,” J. Power Sources, vol. 549, 2022, doi: 10.1016/j.jpowsour.2022.232137.
X. Chen et al., “Open data sets for assessing photovoltaic system reliability,” Appl. Energy, vol. 395, 2025, doi: 10.1016/j.apenergy.2025.126132.
R. Prusty, N. Sivakumar, A. Telukdarie, M. G. Nkadimeng, L. Malebo, and T. H. Katsumbe, “Optimizing solar energy generation through machine learning: A framework for CUF analysis,” Renew. Energy, vol. 256, 2026, doi: 10.1016/j.renene.2025.124251.
Q. Sun, F. Yan, W. Sun, and Y. Zhou, “DWT-Former: Fusing wavelet-based multi-scale features and transformer-based temporal representations for photovoltaic power forecasting,” Energy, vol. 341, 2025, doi: 10.1016/j.energy.2025.139283.
J. A. Nava-pintor et al., “Development and Evaluation of Solar Radiation Sensor Using Cost-Effective Light Sensors and Machine Learning Techniques,” pp. 1–31, 2025.
A. Beyaz and V. Gül, “Determination of Low-Cost Arduino Based Light Intensity Sensors Effectiveness for Agricultural Applications,” Brazilian Arch. Biol. Technol., vol. 65, pp. 1–14, 2022, doi: 10.1590/1678-4324-2022220172.
DOI: https://doi.org/10.52088/ijesty.v6i4.1892
Refbacks
- There are currently no refbacks.
Copyright (c) 2026 Feby Ardianto, Eko Ariyanto, Bengawan Alfaresi, Muhammad Syaukani, Wira Hidayat Bin Mohd Saad































