Share

IIR document

Optimizing thermophysical properties of non-Newtonian nano-refrigerants for refrigeration systems using machine learning approaches.

Author(s) : AKBARI M., BAGHERZADEH S. A., DEHKORDI M. H. R., NAGHSH A., AZIMY N., AZIMY H.

Type of article: IJR article

Summary

The thermophysical properties of non-Newtonian superparamagnetic nano-refrigerants play a pivotal role in enhancing the efficiency of advanced cooling and next-generation refrigeration systems. The scope of this study includes the prediction and optimization of thermal conductivity and viscosity of cobalt ferrite (CoFe₂O₄)-based nanofluids over a range of temperatures and concentrations, targeting practical use in refrigeration system design. This study introduces a machine learning framework designed to predict and optimize the thermophysical properties of cobalt ferrite (CoFe₂O₄)-based superparamagnetic nano-refrigerants, with a focus on applications in refrigeration and heat transfer systems. By leveraging the capabilities of Radial Basis Function Networks (RBFNs), particularly a Generalized Regression Neural Network (GRNN), the model captures the relationships between critical input parameters such as temperature and nano-refrigerant concentration and key thermophysical outputs, including thermal conductivity and viscosity. To further enhance the framework, Particle Swarm Optimization (PSO) is integrated for inverse modeling, enabling the identification of optimal decision variables that balance thermal performance and magnetocaloric nano-refrigerant flow behavior. Unlike previous studies that rely primarily on empirical correlations or conventional simulations, this work introduces a hybrid machine learning–optimization framework that enables both predictive modeling and inverse design of thermally responsive nanofluids. The results reveal that increasing temperature significantly improves thermal conductivity, with an increase of 21.13 % at 50 ◦C, while viscosity rises by up to 61.38 % depending on nanoparticle loading at 50 ◦C. The model identifies an optimal mass fraction of ~0.3092 % at 50 ◦C, which achieves the best trade-off between heat transfer enhancement and acceptable flow behavior. Importantly, the proposed framework incorporates physical consistency checks and experimental constraints, ensuring robust and reliable predictions. These results validate the effectiveness of the integrated framework in addressing the challenges of thermophysical modeling for advanced cooling fluids. This work not only advances the understanding of non-Newtonian superparamagnetic nano-refrigerants but also demonstrates the transformative potential of combining neural networks with optimization techniques for data-driven modeling in refrigeration and heat transfer applications. The findings offer valuable insights for the development of next-generation cooling systems with improved efficiency and performance.

Available documents

Format PDF

Pages: 12 p.

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: Optimizing thermophysical properties of non-Newtonian nano-refrigerants for refrigeration systems using machine learning approaches.
  • Record ID : 30034369
  • Languages: English
  • Subject: Technology
  • Source: International Journal of Refrigeration - Revue Internationale du Froid - vol. 179
  • Publication date: 2025/11

Links


See other articles in this issue (29)
See the source

Indexing