Prediction and analysis on the effects of different inlet diameters of a hydrogen gas tank during fast fill using ANN by the neurofit technique

The greater possibility of explosions and thermal instability connected with hydrogen fuel containers have drawn a lot of attention recently, leading to a lot of studies being conducted. A number of techniques have been used, including the use of different materials for the tank walls and gas pre-cooling. The current investigation is to examine the impact of employing various inlet sizes on the spatial temperature distribution. The FEA tool was the primary application software used in this investigation. In order to predict the temperature increase that takes place during the quick fill process of hydrogen gas, a model of the hydrogen gas tank was built within this software using computational fluid dynamics (CFD) simulation. Initially, the model was constructed, and the experiment was verified through a comparative analysis of the outcomes with previously published literature. Subsequently, the dimension of the tank inlet was altered, and the resulting increase in temperature was observed. Changes in temperature and variations in the hydrogen gas were monitored for their respective patterns. It has been demonstrated that the temperature variations that are seen during the quick fill procedure for gas tanks with larger diameters of 20 and 25 mm are higher. An artificial neural network (ANN) was adopted to predict the temperature rise in the tank during fast filling. The ANN technique accurately predicted the outcome. The overall mean square error is \({R}^{2}\) $R^2$ =0.9975, which is a minimal error. As a result, the numerical method and ANN outcomes are reliable and agree well.