Akademik Veri Yönetim Sistemi
Sadhana - Academy Proceedings in Engineering Sciences, cilt.50, sa.4, 2025 (SCI-Expanded, Scopus)
The rising energy demand for residential heating and cooling contributes significantly to environmental pollution through carbon emissions from conventional energy sources. The present study explores vapour injection heat pump (VIHP) systems to address these challenges, evaluating their performance under varying operating conditions and examining the potential of machine learning (ML) models for rapid and accurate predictions without relying on complex engineering equations. Parametric analyses were conducted for five refrigerants (R123, R134a, R152a, R32 and R1234yf) across a range of injection mass ratios (0.02–0.24) and outdoor temperatures (−9 to 38°C). The results highlight how heating and cooling capacities, compressor power and system performance vary under different conditions. For instance, performance improvements were most pronounced for R123 with increasing IMR, while R32 showed the largest sensitivity to outdoor temperature changes. Environmental impacts were quantified, with R123 producing the highest carbon emissions and R32 the lowest. Additionally, five ML models were evaluated, revealing that AdaBoost Regressor was the least accurate, ExtraTrees regressor model provided robust predictions. These findings support the design and optimization of VIHP systems, offering a cost-effective approach to improving energy efficiency and reducing environmental impacts before undertaking costly experimental studies.