A comprehensive review on the application of nanofluid in heat pipe based on the machine learning: Theory, application and prediction. (October 2021)

• Record Type: Journal Article
• Authors: Wang, Xianling; Luo, Liang; Xiang, Jinwei; Zheng, Senlin; Shittu, Samson; Wang, Zhangyuan; Zhao, Xudong
• Is Part Of: Renewable & sustainable energy reviews. Volume 150(2021)
• Publication Date: 2021-10
• Subjects: Heat pipe -- Nanofluid -- Machine learning -- Theory -- Application -- Prediction; Renewable energy sources -- Periodicals; Power resources -- Periodicals; Énergies renouvelables -- Périodiques; Ressources énergétiques -- Périodiques; 333.794
• DOI: 10.1016/j.rser.2021.111434
• Languages: English
• ISSNs: 1364-0321

Abstract

Abstract: This paper introduces three paramount factors i.e. viscosity, thermal conductivity and stability that affect the application of mono and hybrid nanofluids in heat pipes. The applications of nanofluids in various types of heat pipes are reviewed and the mechanism of heat transfer enhancement or inhibition is summarized. The applications of machine learning in nanofluids (thermal conductivity and dynamic viscosity) and heat pipes charged with nanofluids are presented. The main challenges include: (1) difference and uncertainty on thermal conductivity and viscosity, as well as undesirability on stability property of nanofluid; (2) lack of comprehension of time-dependent property of heat pipes; (3) limitation of predictive models based on machine learning; and (4) lack of an appropriate standard for selecting the appropriate machine learning algorithm. To tackle the above imminent challenges, further opportunities are revealed including: (1) exploring the mechanism at nanoscale and establishing unified standards, as well as exploring the effect of surfactant and smaller particle size; (2) focusing on the nanoparticle deposition layer; (3) establishing the large, exclusive databases and expanding the input variables; and (4) defining specific standard by horizontal comparison and using more advanced algorithms. This review-based study provides the guidelines for the development of heat pipes charged with nanofluids and establishes the foundation for the application of machine learning technology in heat pipes and nanofluids. Highlights: There are some differences and uncertainties on thermal conductivity and viscosity of nanofluids. There is a lack of comprehension of the time-dependent property of heat pipes. There is a limitation of predictive models based on machine learning techniques. The mechanism of thermo-physical properties for nanofluids at nanoscale should be explored. The exclusive large databases should be established and the input variables should be expanded.