A Hybrid πVAE-XGBoost Framework for High-Fidelity Simulation of Bifacial Solar Thermal Collectors
Imane FAKIR, Mehdi NEJJAR, Mohammad Hachim ERAISSOUNI, Adam KESSAB, Ahmed KHALLAAYOUN
Abstract. The following paper describes a new hybrid system which combines a deep learning-based pvae model with an XGBoost regressor to produce high-fidelity bifacial solar thermal collector high-fidelity simulation. Through modelling the proposed approach overcomes the above-mentioned challenges by capturing complex, nonlinear interactions between critical system variables, including temperature and irradiance and flow rates, as-sociated with the modeling of a pair-exposure collectors. The high-dimensional input space is effectively reduced to a strong latent representation using the pVAE and further predictions are refined by including the latent features along with the original measurements to the XGBoost component to make reliable predictions of heat transfer and outlet temperature. The validation based on a dataset of 147,396 sequential measurements, as well as comparisons of traditional MATLAB Simulink-CARNOT simulations, shows that the predictive performance is almost perfect, with the values of R2, Spearman, and Pearson correlation coefficients being over 99%. This mixed approach does not only increase the degree of simulation accuracy but also provides explainable information to optimize the design of solar thermal systems, which opens its way to the use in the further renewable energy systems.
Keywords
Bifacial Solar Thermal Collectors, Hybrid Simulation, Deep Learning, πVAE, XGBoost, Renewable Thermal Energy, Nonlinear Modeling
Published online 4/25/2026, 8 pages
Copyright © 2026 by the author(s)
Published under license by Materials Research Forum LLC., Millersville PA, USA
Citation: Imane FAKIR, Mehdi NEJJAR, Mohammad Hachim ERAISSOUNI, Adam KESSAB, Ahmed KHALLAAYOUN, A Hybrid πVAE-XGBoost Framework for High-Fidelity Simulation of Bifacial Solar Thermal Collectors, Materials Research Proceedings, Vol. 64, pp 171-178, 2026
DOI: https://doi.org/10.21741/9781644904091-21
The article was published as article 21 of the book Energy Futures
Content from this work may be used under the terms of the Creative Commons Attribution 3.0 license. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
References
[1] M. Benramdane, S. Abboudi, and M. Ghernaout, “Contribution to the simulation and parametric analysis of the operation of a solar concentration thermal installation,” International Journal of Heat and Technology, vol. 37, no. 2, pp. 446–456, Jun. 2019. https://doi.org/10.18280/ijht.370210.
[2] M. Ernst, G. E. J. Conechado, and C.-A. Asselineau, “Accelerating the simulation of annual bifacial illumination of real photovoltaic systems with ray tracing,” iScience, vol. 25, no. 1, p. 103698, Dec. 2021. https://doi.org/10.1016/j.isci.2021.103698.
[3] J. D. Osorio, M. De Florio, R. Hovsapian, C. Chryssostomidis, and G. E. Karniadakis, “Physics-informed machine learning for solar-thermal power systems,” Energy Conversion and Management, vol. 327, p. 119542, 2025. https://doi.org/10.1016/j.enconman.2025.119542.
[4] N. Nikolić and N. Lukić, “Theoretical and experimental investigation of the thermal performance of a double exposure flat-plate solar collector,” Solar Energy, vol. 119, pp. 100–113, Sep. 2015. https://doi.org/10.1016/j.solener.2015.06.038.
[5] N. Nikolic, N. Lukic, A. Nesovic, and D. Nikolic, “Optimal size and position of the planar back reflector moveable only in the direction normal to the bifacial solar collector plane,” Thermal Science, vol. 28, no. 6A, pp. 4483–4497, 2024. https://doi.org/10.2298/TSCI240123101N.
[6] T. Summ, M. Ehrenwirth, C. Trinkl, and W. Zörner, “Agri Solar Thermal Systems: A brief study on the energetic potential of bifacial solar thermal systems,” in Proc. EuroSun 2022 – ISES and IEA SHC Int. Conf. Solar Energy for Buildings and Industry, Kassel, Germany, 2022, pp. 1–8. https://doi.org/10.18086/eurosun.2022.12.16.
[7] T. Summ, B. Sturm, C. Scherer, C. Trinkl, and W. Zörner, “Analysis of thermomechanical stresses and strains for a novel design of insulating glass flat-plate collector,” in Proc. Solar World Congress 2023, New Delhi, India, 2023, pp. 1–11. https://doi.org/10.18086/swc.2023.04.14.
[8] S. Mishra, S. Flaxman, T. Berah, H. Zhu, M. Pakkanen, and S. Bhatt, “πVAE: A stochastic process prior for Bayesian deep learning with MCMC,” Statistics and Computing, vol. 32, no. 6, p. 96, Oct. 2022. https://doi.org/10.1007/s11222-022-10151-w.
[9] T. Chen and C. Guestrin, “XGBoost: A scalable tree boosting system,” arXiv preprint, arXiv:1603.02754, accessed Mar. 25, 2025. [Online]. Available: https://arxiv.org/abs/1603.02754
