Performance optimization of adsorption hydrogen storage system via computation fluid dynamics and machine learning

Abstract

Adsorption hydrogen storage paves an alternative way for reliable hydrogen storage. Here, impacts of dimensionless fin geometry configurations of the fin-tube metal hydride bed on the hydrogen storage performance are systematically investigated. Results show employing fins can strengthen the heat and mass transfer characteristics in the adsorbent bed and shorten the adsorption duration, especially at large fin numbers. Larger dimensionless fin thickness, height and width can lower the average bed temperature, which contributes to the adsorption kinetics and hydrogen fraction. At a small adsorption duration, larger dimensionless fin thickness, height and width improve the hydrogen storage amount due to obviously augmented hydrogen fraction. At a larger adsorption duration, there exist optimal dimensionless fin thickness, height and width leading to the maximum hydrogen storage amount originating from the compromise between the hydrogen fraction and adsorbent filled. Based on machine learning, the relationships between dimensionless fin geometry parameters and hydrogen storage amount have been obtained, and the optimal adsorption duration depended dimensionless fin configurations are obtained based on genetic algorithm. At a moderate adsorption duration, the optimized finned bed can significantly enhance the hydrogen storage amount. At the adsorption duration of 400 s, the hydrogen storage amount is augmented by 12.8%.