Residual Physics-Informed Neural Networks for Robust Photovoltaic Cell Temperature Modeling Across Seasonal and Data-Scarce Conditions

Erhan Baran

doi:10.54287/gujsa.1963984

Residual Physics-Informed Neural Networks for Robust Photovoltaic Cell Temperature Modeling Across Seasonal and Data-Scarce Conditions

Abstract

Accurate photovoltaic (PV) cell temperature prediction is essential for improving power forecasting, efficiency estimation, and long-term reliability analysis of photovoltaic systems. However, conventional empirical thermal models often lack adaptability under varying environmental conditions, whereas purely data-driven methods may exhibit poor generalization and physically inconsistent predictions. This study proposes a Residual Physics-Informed Neural Network (Residual-PINN) that integrates a simplified lumped thermal energy balance model with residual learning to enhance both predictive accuracy and physical consistency. The physics-based model first estimates the dominant thermal behavior, while the neural network learns only the residual temperature component associated with nonlinear environmental effects. The proposed framework was comprehensively evaluated through random-split, cross-season, extreme-irradiance, physics-consistency and low-data experiments. Experimental results demonstrate that the proposed Residual-PINN achieves the best overall performance with a MAE of 1.82°C, an RMSE of 2.61°C, and an R² value of 0.95, outperforming all benchmark models. Compared with the conventional multilayer perceptron, the proposed approach reduces RMSE by approximately 20% while decreasing the nighttime prediction error to 0.36°C and limiting the monotonicity violation ratio to 0.9%. Furthermore, the model maintains superior robustness under seasonal distribution shifts and low-data conditions, achieving an RMSE of 2.95°C when trained using only 20% of the available data. These results demonstrate that integrating simplified thermal physics with residual learning provides a robust, interpretable, and data-efficient framework for photovoltaic temperature prediction.

Keywords

References

Baran, E. (2026). Deep learning-based detection and forecasting of performance losses in solar PV systems using multi-sensor data. Applied Sciences, 16(4), 1709. https://doi.org/10.3390/app16041709
Baran, E., & Korkusuz Polat, T. (2022). Classification of industry 4.0 for total quality management: A review. Sustainability, 14(6), 3329. https://doi.org/10.3390/su14063329
Faiman, D. (2008). Assessing the outdoor operating temperature of photovoltaic modules. Progress in Photovoltaics: Research and Applications, 16(4), 307-315. https://doi.org/10.1002/pip.813
Farea, A., Yli-Harja, O., & Emmert-Streib, F. (2024). Understanding Physics-Informed Neural Networks: Techniques, Applications, Trends, and Challenges. AI, 5(3), 1534-1557. https://doi.org/10.3390/ai5030074
Foncubierta Blázquez, J. L., Mena Baladés, J. D., Sánchez Orihuela, I., Jiménez Come, M. J., & González Siles, G. (2025). Development, implementation, and experimental validation of a novel thermal-optical-electrical model for photovoltaic glazing. Applied Sciences, 15(22), 12041. https://doi.org/10.3390/app152212041
Goodfriend, W., Pieters, E. B., Tsvetelina, M., Solomon, A., Ezema, F., & Rau, U. (2024). Development and improvement of a transient temperature model of PV modules: Concept of trailing data. Progress in Photovoltaics: Research and Applications, 32(6), 399-405. https://doi.org/10.1002/pip.3785
Karniadakis, G. E., Kevrekidis, I. G., Lu, L., Perdikaris, P., Wang, S., & Yang, L. (2021). Physics-Informed Machine Learning. Nature Reviews Physics, 3(6), 422-440. https://doi.org/10.1038/s42254-021-00314-5
Keddouda, A., Ihaddadene, R., Boukhari, A., Atia, A., Arıcı, M., Lebbihiat, N., & Ihaddadene, N. (2024). Photovoltaic module temperature prediction using various machine learning algorithms: Performance evaluation. Applied Energy, 363, 123064. https://doi.org/10.1016/j.apenergy.2024.123064

Li, P., Luo, Y., Xia, X., Shi, W., Zheng, J., Liao, Z., Gao, X., & Chang, R. (2024). Advancing photovoltaic panel temperature forecasting: a comparative study of numerical simulation and machine learning in two types of PV power plant. Renewable Energy, 237, 121602. https://doi.org/10.1016/j.renene.2024.121602
Li, G., Yu, W., Chen, L., Lu, Y., Wu, Q., Guo, M., Liu, W., Wang, S., & Han, F. (2025). An energy balance boundary model for the thermal energy budget of aeolian sand surfaces on the Tibetan Plateau. Thermal Science and Engineering Progress, 68, 104388. https://doi.org/10.1016/j.tsep.2025.104388
Mellit, A., & Kalogirou, S. A. (2008). Artificial intelligence techniques for photovoltaic applications: a review. Progress in Energy and Combustion Science, 34(5), 574-632. https://doi.org/10.1016/j.pecs.2008.01.001
Nicola, M., & Berwind, M. (2024). Improving module temperature prediction models for floating photovoltaic systems: Analytical insights from operational data. Energies, 17(17), 4289. https://doi.org/10.3390/en17174289
Qian, Y., Liu, R., Wang, H., Zhang, H., Zhou, X., & Zhao, L. (2026). Applying occupant-centered control systems to balance energy consumption and thermal comfort: a case study in China’s severe cold region. Energy and Buildings, 351, 116729. https://doi.org/10.1016/j.enbuild.2025.116729
Raissi, M., Perdikaris, P., & Karniadakis, G. E. (2019). Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. Journal of Computational Physics, 378, 686-707. https://doi.org/10.1016/j.jcp.2018.10.045
Santos, L. de O., de Carvalho, P. C. M., & Filho, C. de O. C. (2022). Photovoltaic cell operating temperature models: a review of correlations and parameters. IEEE Journal of Photovoltaics, 12(1), 179-190. https://doi.org/10.1109/JPHOTOV.2021.3113156
Santos, L. O., Souza, F. A. A., Carvalho Filho, C. O., Carvalho, P. C. M., AISkaif, T., & Pereira, R.I.S. (2024). Hybrid modeling for photovoltaic module operating temperature estimation. IEEE Journal of Photovoltaics, 14(3), 488-496. https://doi.org/10.1109/JPHOTOV.2024.3372328
Skoplaki E., & Palyvos, J. A. (2009). On the temperature dependence of photovoltaic module electrical performance: A review of efficiency/power correlations. Solar Energy, 83(5), 614-624. https://doi.org/10.1016/j.solener.2008.10.008
Voyant, C., Notton, G., Kalogirou, S., Nivet, M.-L., Paoli, C., Motte, F., & Fouilloy, A. (2017). Machine learning methods for solar radiation forecasting: A review. Renewable Energy, 105, 569-582. https://doi.org/10.1016/j.renene.2016.12.095
Wang, K., Wang, L., Meng, Q., Yang, C., Lin, Y., Zhu, J., Zhao, Z., Zhou, C., Zheng, C., & Gao, X. (2025). Accurate photovoltaic power prediction via temperature correction with physics-informed neural networks. Energy, 328, 136546. https://doi.org/10.1016/j.energy.2025.136546
Wang, D., Liang, Z., Zhang, Z., & Li, M. (2025). Efficient estimation of convective cooling of photovoltaic arrays: A physics-informed machine learning approach. Energy and AI, 20, 100499. https://doi.org/10.1016/j.egyai.2025.100499
Wu, Y., Sicard, B., & Gadsden, S. A. (2024). Physics-informed machine learning: a comprehensive review on applications in anomaly detection and condition monitoring. Expert Systems with Applications, 255, 124678. https://doi.org/10.1016/j.eswa.2024.124678
Zhou, K., Liu, Z., Qiao, Y., Xiang, T., & Loy, C. C. (2023). Domain generalization: a survey. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(4), 4396-4415. https://doi.org/10.1109/TPAMI.2022.3195549

Details

Primary Language

English

Subjects

Renewable Energy Resources , Industrial Engineering

Journal Section

Research Article

Authors

Erhan Baran ^*
0000-0002-5897-8680
Türkiye

Early Pub Date

September 24, 2026

Publication Date

September 30, 2026

Submission Date

June 4, 2026

Acceptance Date

July 20, 2026

Published in Issue

Year 2026 Volume: 13 Number: 3

DOI

https://doi.org/10.54287/gujsa.1963984

IZ

https://izlik.org/JA83JK74EH

Cite

RIS / Bibtex

APA

Baran, E. (2026). Residual Physics-Informed Neural Networks for Robust Photovoltaic Cell Temperature Modeling Across Seasonal and Data-Scarce Conditions. Gazi University Journal of Science Part A: Engineering and Innovation, 13(3), 1169-1199. https://doi.org/10.54287/gujsa.1963984

AMA

1.Baran E. Residual Physics-Informed Neural Networks for Robust Photovoltaic Cell Temperature Modeling Across Seasonal and Data-Scarce Conditions. GU J Sci, Part A. 2026;13(3):1169-1199. doi:10.54287/gujsa.1963984

Chicago

Baran, Erhan. 2026. “Residual Physics-Informed Neural Networks for Robust Photovoltaic Cell Temperature Modeling Across Seasonal and Data-Scarce Conditions”. Gazi University Journal of Science Part A: Engineering and Innovation 13 (3): 1169-99. https://doi.org/10.54287/gujsa.1963984.

EndNote

Baran E (September 1, 2026) Residual Physics-Informed Neural Networks for Robust Photovoltaic Cell Temperature Modeling Across Seasonal and Data-Scarce Conditions. Gazi University Journal of Science Part A: Engineering and Innovation 13 3 1169–1199.

IEEE

[1]E. Baran, “Residual Physics-Informed Neural Networks for Robust Photovoltaic Cell Temperature Modeling Across Seasonal and Data-Scarce Conditions”, GU J Sci, Part A, vol. 13, no. 3, pp. 1169–1199, Sept. 2026, doi: 10.54287/gujsa.1963984.

ISNAD

Baran, Erhan. “Residual Physics-Informed Neural Networks for Robust Photovoltaic Cell Temperature Modeling Across Seasonal and Data-Scarce Conditions”. Gazi University Journal of Science Part A: Engineering and Innovation 13/3 (September 1, 2026): 1169-1199. https://doi.org/10.54287/gujsa.1963984.

JAMA

1.Baran E. Residual Physics-Informed Neural Networks for Robust Photovoltaic Cell Temperature Modeling Across Seasonal and Data-Scarce Conditions. GU J Sci, Part A. 2026;13:1169–1199.

MLA

Baran, Erhan. “Residual Physics-Informed Neural Networks for Robust Photovoltaic Cell Temperature Modeling Across Seasonal and Data-Scarce Conditions”. Gazi University Journal of Science Part A: Engineering and Innovation, vol. 13, no. 3, Sept. 2026, pp. 1169-9, doi:10.54287/gujsa.1963984.

Vancouver

1.Erhan Baran. Residual Physics-Informed Neural Networks for Robust Photovoltaic Cell Temperature Modeling Across Seasonal and Data-Scarce Conditions. GU J Sci, Part A. 2026 Sep. 1;13(3):1169-9. doi:10.54287/gujsa.1963984