Heat transfer on the inner wall during gas filling into high-pressure tanks is a crucial phenomenon for understanding hydrogen storage in fuel cell vehicles. This process includes both natural and forced convection, which makes it difficult to predict heat transfer accurately. In previous studies, nitrogen gas was supplied to a model tank to investigate wall heat transfer characteristics and organize them using simple model equations. However, a more accurate prediction method is needed. In this study, we developed a polynomial machine learning model based on experimental data to predict heat transfer coefficients during gas filling, and examined its physical validity.
Kataoka et al. (Wed,) studied this question.