Treffer: Hybrid solar-thermoelectric energy harvesting with machine learning models.

[Display omitted] • Peak output of 2.48 W occurs with 17–18 °C temperature difference. • Output power peaks at noon and 6 pm due to thermal and ambient conditions. • A MATLAB neural network model attains prediction accuracy with an R2 of 0.969. • Gaussian process regression is the top performer with the lowest error (MSE = 0.0105). • The conducted system is feasible under real conditions with green power generation. This study addresses the challenge of recovering waste heat from solar thermal systems by integrating thermoelectric generators with an evacuated tube solar collector to produce green power. Efficient waste heat recovery is critical for enhancing overall energy utilization and achieving low-carbon energy solutions. An experimental setup combining six SP1848-27145 modules and aluminum fin heat sinks is mounted on the tank surface of a solar water heater and tested under real outdoor conditions in Tainan, Taiwan. Data are recorded over 24 h, capturing solar radiation, temperature differences, and power output. Two machine learning approaches are applied to model and predict the electrical performance: artificial neural networks developed in PolyAnalyst and MATLAB, and six machine learning models evaluated via MATLAB's regression learner. The MATLAB-based artificial neural network with two hidden layers (80 neurons each) achieves the highest predictive accuracy, with an R2 of 0.969. Among regression learner models, Gaussian process regression yields the best performance with an R2 of 0.960 and a mean squared error of 0.0105. The experimental system produces a peak power output of 2.48 W with a maximum temperature difference of 17–18°C. This study is the first to integrate evacuated tube solar collectors and thermoelectric generator systems with IoT-based real-time monitoring and predictive machine learning models. It goes beyond prior efforts by demonstrating accurate, data-driven forecasting of thermoelectric generator output under fluctuating environmental conditions, paving the way for intelligent optimization of hybrid solar-thermoelectric generator systems in off-grid and resource-limited settings. [ABSTRACT FROM AUTHOR]

Copyright of Applied Thermal Engineering is the property of Pergamon Press - An Imprint of Elsevier Science and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)