Share

IIR document

Performance evaluation and multi-objective optimization of low-GWP refrigerants in vapor compression refrigeration systems using machine learning. 

Author(s) : ASHOUR A. M., KADHIM S. A., NAYYEF D. R. , ABDUL-ZAHRA A. S. , MOTA BABILONI A., RASHID F. L., BOUABIDI A.

Type of article: IJR article

Summary

The growing demand for electrified cooling systems and the global phase-down of high global warming potential (GWP) refrigerants necessitate the development of energy-efficient and environmentally sustainable vapor- compression refrigeration (VCR) systems. This study presents an integrated experimental, predictive, and multi-objective optimization framework to evaluate the performance of a VCR system using low-GWP re frigerants under extended operating conditions. Six refrigerants (R404A, R134a, R454C, R290, R1270, and R1234yf) are assessed over a wide range of evaporator temperatures ( 10 to 10 ◦ (35 to 55 ◦ C) and condenser temperatures C), representative of practical applications. Machine learning models are developed to predict key performance indicators, including coefficient of performance (COP), cooling capacity (Q e ), and compressor power consumption (W c ), with high accuracy (R² > 0.99). These models are integrated with the Non-dominated Sorting Genetic Algorithm (NSGA-II) to analyze trade-offs among competing objectives. Pareto analysis com bined with a score-based decision strategy is used to rank refrigerants and determine optimal operating condi tions. The novelty of this study lies in integrating experimentally grounded thermodynamic analysis with data- driven modeling to construct a physically consistent dataset over extended operating conditions, and embedding machine learning within a multi-objective optimization framework coupled with a decision-oriented ranking strategy. Results show that R290 and R1270 achieve superior efficiency and cooling performance, while R1234yf provides lower discharge temperatures. The multi-objective analysis identifies R134a as the most balanced refrigerant under the selected weighting scheme. The proposed framework offers a practical decision-support tool for sustainable refrigeration system design.

Available documents

Format PDF

Pages: 16

Available

  • Public price

    20 €

  • Member price*

    Free

* Best rate depending on membership category (see the detailed benefits of individual and corporate memberships).

Details

  • Original title: Performance evaluation and multi-objective optimization of low-GWP refrigerants in vapor compression refrigeration systems using machine learning. 
  • Record ID : 30034934
  • Languages: English
  • Subject: Technology
  • Source: International Journal of Refrigeration - Revue Internationale du Froid - vol. 187
  • Publication date: 2026/07
  • DOI: http://dx.doi.org/https://doi.org/10.1016/j.ijrefrig.2026.106962

Links


See other articles in this issue (35)
See the source

Indexing