From Experiment to Deployment: An Integrated Framework for Predictive Modelling and Visualisation of Thermoelectric HVAC Systems

Tajul Rosli Razak; Cagri Kutlu; Tianhong Zheng; Hasila Jarimi; Yuehong Su; Saffa Riffat; Preethi Jayakumar; Dhruv Shah

doi:10.65582/aifsc.2026.008

Authors

Tajul Rosli Razak Department of Architecture and Built Environment, The University of Nottingham, University Park, Nottingham, NG7 2RD, United Kingdom; Faculty of Computer and Mathematical Science, Universiti Teknologi MARA (UiTM), Shah Alam, 40450, Malaysia. https://orcid.org/0000-0002-6389-8108
Cagri Kutlu Department of Mechanical Engineering, Necmettin Erbakan University, Konya, Türkiye. https://orcid.org/0000-0002-8462-0420
Tianhong Zheng Department of Architecture and Built Environment, The University of Nottingham, University Park, Nottingham, NG7 2RD, United Kingdom. https://orcid.org/0000-0002-7761-8939
Hasila Jarimi Department of Architecture and Built Environment, The University of Nottingham, University Park, Nottingham, NG7 2RD, United Kingdom. https://orcid.org/0000-0003-0921-3283
Yuehong Su Department of Architecture and Built Environment, The University of Nottingham, University Park, Nottingham, NG7 2RD, United Kingdom. https://orcid.org/0000-0002-6616-7626
Saffa Riffat Department of Architecture and Built Environment, The University of Nottingham, University Park, Nottingham, NG7 2RD, United Kingdom. https://orcid.org/0000-0002-3911-0851
Preethi Jayakumar Anzen Innovations Limited, NW10 6HJ, United Kingdom.
Dhruv Shah Anzen Innovations Limited, NW10 6HJ, United Kingdom.

DOI:

https://doi.org/10.65582/aifsc.2026.008

Keywords:

Thermoelectric HVAC, Artificial Neural Networks, Performance Prediction, Explainable AI, Sustainable Building Systems

Abstract

Thermoelectric (TE) heating, ventilation, and air-conditioning (HVAC) systems represent a sustainable alternative to conventional vapor-compression technologies, but they remain limited by modest coefficients of performance (COPs) and a lack of robust optimization methods. This study presents a comprehensive experimental, computational, and deployment pipeline to address these challenges. A custom-built test rig was designed to generate high-quality data under controlled laboratory conditions, capturing all key thermodynamic variables relevant to TE operation. Multiple regression and machine learning models were systematically benchmarked, with Linear Regression emerging as the most accurate and parsimonious predictor of hot-side exit temperature. Beyond statistical metrics, predictions were validated against recomputed heat transfer rate and COP, ensuring thermodynamic consistency and physical interpretability. The best-performing model was embedded into an interactive Shiny-based graphical user interface (GUI) to bridge the gap between research outputs and practical usability. The GUI allows real-time adjustment of system inputs, dynamic prediction of outputs, and visualization of performance through bar, point, and time-series charts, supported by logged historical data. By combining experimental rigor, comparative regression analysis, thermodynamic validation, and deployment into an accessible decision-support tool, this work demonstrates a practical pathway for advancing TE HVAC systems from laboratory studies toward real-world applications in sustainable building energy management.

References

Adibhesami, M. A. and Hassanzadeh, A., 2025. Optimizing HVAC energy efficiency in low-energy buildings: a comparative analysis of reinforcement learning control strategies under Tehran climate conditions, Data-Centric Engineering, 6. DOI: doi.org/10.1017/dce.2025.10014.

Alghamdi, F. and Krarti, M., 2025. Review Analysis for the Energy Performance of Integrated Air-Conditioning Systems, Energies, 18(7). DOI: doi.org/10.3390/en18071611.

Bahk, J. H. et al., 2025. Performance evaluation and optimization of air-to-air thermoelectric heat pump system, Energy and Buildings, 344(May), p. 115994. DOI: doi.org/10.1016/j.enbuild.2025.115994.

Dizaji, H. S., Jafarmadar, S. and Khalilarya, S., 2019. Novel experiments on COP improvement of thermoelectric air coolers, Energy Conversion and Management, 187(January), pp. 328–338. DOI: doi.org/10.1016/j.enconman.2019.03.025.

Duan, M. et al., 2021. Evaluation on the applicability of thermoelectric air cooling systems for buildings with thermoelectric material optimization, Energy, 221, p. 119723. DOI: doi.org/10.1016/j.energy.2020.119723.

He, Z., Zuazua-Ros, A. and Martín-Gómez, C., 2024. Current-dependent temperature change model of a thermoelectric window frame, Applied Thermal Engineering, 247(March), pp. 1–14. DOI: doi.org/10.1016/j.applthermaleng.2024.123081.

Ibañez-Puy, M. et al., 2017. Thermoelectric cooling heating unit performance under real conditions, Applied Energy, 200, pp. 303–314. DOI: doi.org/10.1016/j.apenergy.2017.05.020.

Li, M., Dai, L. and Hu, Y., 2022. Machine Learning for Harnessing Thermal Energy: From Materials Discovery to System Optimization, ACS Energy Letters, 7(10), pp. 3204–3226. DOI: doi.org/10.1021/acsenergylett.2c01836.

Ma, X. et al. 2023. Performance modelling on a thermoelectric air conditioning system using high power heat sinks and promoting waste heat utilization, Energy, 268(November 2022), p. 126612. DOI: doi.org/10.1016/j.energy.2023.126612.

Manikandan, S. and Kaushik, S. C., 2017. Transient Thermal Behavior of Annular Thermoelectric Cooling System, Journal of Electronic Materials, 46(5), pp. 2560–2569. DOI: doi.org/10.1007/s11664-016-5055-7.

Myat, A., Rahman, M. M. and Akbar, M., 2025. Nanofluid-Enhanced HVAC&R Systems (2015–2025): Experimental, Numerical, and AI-Driven Insights with a Strategic Roadmap, Sustainability (Switzerland), 17(16). DOI: doi.org/10.3390/su17167371.

Nassif, N., 2014. Modeling and optimization of HVAC systems using artificial neural network and genetic algorithm, Building Simulation, 7(3), pp. 237–245. DOI: doi.org/10.1007/s12273-013-0138-3.

Ni, H. P. et al., 2025. Enhancing HVAC energy efficiency modeling in semiconductor manufacturing facilities using tree-structured parzen estimator-optimized deep learning, Building and Environment, 271(January), p. 112589. DOI: doi.org/10.1016/j.buildenv.2025.112589.

Zaki, A. M. et al., 2025. Environmental and energy performance analyses of HVAC systems in office buildings using boosted ensembled regression trees: Machine learning strategy for energy saving of air conditioning and lighting facilities, Process Safety and Environmental Protection, 198(November 2024), p. 107214. DOI: doi.org/10.1016/j.psep.2025.107214.

Zhou, G., 2026. Thermoelectric coolers : Systematic Analysis on Principle , Use and Limitations. Atlantis Press International BV. DOI: doi.org/10.2991/978-94-6463-986-5.