Machine Learning for Modeling Oscillating Heat Pipes: A Review

Abstract

Oscillating heat pipes are heat transfer devices with the potential of addressing some of the most pressing current thermal management problems, from the miniaturization of microchips to the development of hypersonic vehicles. Since their invention in the 1990s, numerous studies have attempted to develop predictive and inverse design models for oscillating heat pipe function. However, the field still lacks robust and flexible models that can be used to prescribe design specifications based on a target performance. The fundamental difficulty lies in the fact that, despite the simplicity of their design, the mechanisms behind the operation of oscillating heat pipes are complex and only partially understood. To circumvent this limitation, over the last several years, there has been increasing interest in the application of machine learning techniques to oscillating heat pipe modeling. Our survey of the literature has revealed that machine learning techniques have successfully been used to predict different aspects of the operation of these devices. However, many fundamental questions such as which machine learning models are better suited for this task or whether their results can extrapolate to different experimental setups remain unanswered. Moreover, the wealth of knowledge that the field has produced regarding the physical phenomena behind oscillating heat pipes is still to be leveraged by machine learning techniques. Herein, we discuss these applications in detail, emphasizing their advantages, limitations, as well as potential paths forward.