Eur. Phys. J. Plus (2025) 140: 1217
https://doi.org/10.1140/epjp/s13360-025-07149-4

Regular Article

Machine learning-based exergy and environmental optimization of a photovoltaic–thermal-assisted CO₂ ejector heat pump

¹ Department of Chemical Engineering, College of Engineering, Imam Mohammad Ibn Saud Islamic University (IMSIU), 11432, Riyadh, Saudi Arabia
² Advanced Technical College, University of Warith Al-Anbiyaa, Karbala, Iraq
³ Department of Civil Engineering, College of Engineering, Cihan University-Erbil, Erbil, Iraq
⁴ Independent Researcher, Mechanical Engineering, Doha, Qatar
⁵ Faculty of Data Science and Information Technology, INTI International University, Persiaran Perdana BBN, 71800, Putra Nilai, Nilai, Malaysia
⁶ Physics Department, Faculty of Science, Islamic University of Madinah, P. O. Box: 170, Madinah 42351, Saudi Arabia
⁷ Deanship of Scientific Research, Imam Mohammad Ibn Saud Islamic University (IMSIU), Riyadh, Saudi Arabia
⁸ Al-Manara College for Medical Sciences, Amarah, Maysan, Iraq

^a This email address is being protected from spambots. You need JavaScript enabled to view it. , This email address is being protected from spambots. You need JavaScript enabled to view it.
^b This email address is being protected from spambots. You need JavaScript enabled to view it.

Received: 30 May 2025
Accepted: 2 December 2025
Published online: 17 December 2025

Abstract

Integrating solar energy with CO₂ heat pumps offers a promising solution to mitigate global warming. This study investigates the use of a PV/T collector as the evaporator heat source in an ejector-based transcritical CO₂ (EJRT-CO₂) heat pump. Key parameters—ambient temperature ($T_{\text{amb}}$), load input conditions ($T_{in,w}$ and ${\dot{m}}_{in,w}$), evaporator outlet temperature ($T_{\text{evp}}$), PV/T area ($A_{\text{PVT}}$), solar radiation ($G_t$), and inlet/outlet temperature rise ($\mathrm{\Delta }{T}_{\text{SC}}$)—were analyzed for their impact on system performance (${\text{COP}}_{\text{sys}}$), exergy efficiency (${\eta }_{\text{II}}$), sustainability index (SI), and emissions. Results showed the PV/T and ejector contributed 77 and 6%, respectively, to total exergy destruction. At an 8 kW heating load (${\dot{m}}_{in,w}$= 0.096 kg/s, $T_{in,w}$= 50 °C), optimal performance (${\eta }_{\text{II}}$ = 23.26%, SI = 1.30) occurred with $A_{\text{PVT}}$= 17 m², $T_{\text{evp}}$= 10 °C, T_amb = 2 °C, and $G_t$= 550 W/m². Higher ${\dot{m}}_{in,w}$ or $\mathrm{\Delta }{T}_{\text{SC}}$ reduced ${\text{COP}}_{\text{sys}}$, while increasing $T_{in,w}$, $T_{\text{evp}}$, $A_{\text{PVT}}$, or $G_t$ improved it. Peak efficiency was achieved at a minimal $\mathrm{\Delta }{T}_{\text{SC}}$ (5–6 °C). A neural network model predicted ${\eta }_{\text{II}}$, ${\text{COP}}_{\text{sys}}$, emissions, and thermal energy balance (${\dot{Q}}_{\text{ex}/\text{sh}}$) with R² > 0.999. Optimization for an 8 kW load at $T_{\text{amb}}$= − 10 °C yielded optimal performance (${\eta }_{\text{II}}$= 22.47%, ${\text{COP}}_{\text{sys}}$= 4.79, SI = 1.29) with $A_{\text{PVT}}$= 19 m², $T_{\text{evp}}$= 12.75 °C, and $G_t$= 614 W/m².