Machine-assisted quantification of droplet boiling upon multiple solid materials

The intricate nature of droplet wetting and boiling on heated nano-/micro-structured surfaces has significant implications for practical applications. However, the multitude of factors influencing these liquid-vapor-solid interactions complicates the understanding of boiling behaviors across a wide range of operating conditions. In this study, we examined droplet impingement boiling over a large range of Weber numbers (We) on various surfaces, from smooth to nano-micro hierarchical ones, and from stationary to vibrated ones. The results indicate that the nano-micro hierarchical surface possesses the greatest heat transfer ability due to its superior superhydrophilicity, enhanced by nano-structures, and the vapor buffer effect caused by micro-structures. Additionally, vibration significantly improves boiling performance by 55.4%. However, contrary to previous findings, a higher We does not always enhance heat transfer performance due to a pronounced water hammer effect. We discovered that when the net pressure reaches a threshold value (~ 90 kPa), a new phenomenon - jet flow - occurs, leading to a deterioration in heat transfer. Utilizing our database, we trained a machine learning framework to generate accurate droplet boiling data, providing a droplet boiling law that can extensively predict heat transfer performance. This paper proposes a fusion of physics and artificial intelligence to enhance our understanding and application of droplet boiling phenomena. Our findings are expected to contribute to the development of more efficient droplet-based boiling heat transfer devices and configurations across a wider range of industrial settings.