One-dimensional metal-hydride tank model and simulation in Matlab–Simulink. (8th March 2018)

• Record Type: Journal Article
• Authors: Abdin, Z.; Webb, C.J.; Gray, E.MacA.
• Is Part Of: International journal of hydrogen energy. Volume 43:Number 10(2018)
• Page Start: 5048
• Page End: 5067
• Publication Date: 2018-03-08
• Subjects: Metal-hydride -- Hydride tank -- Model -- Kinetics -- Effective thermal conductivity; Hydrogen as fuel -- Periodicals; Hydrogène (Combustible) -- Périodiques; Hydrogen as fuel; Periodicals; 665.81
• DOI: 10.1016/j.ijhydene.2018.01.100
• Languages: English
• ISSNs: 0360-3199

Abstract

Abstract: A model has been developed for a metal-hydride tank for hydrogen storage, based on linked modular mathematical models in Simulink ® . The objective of the work was a tank-level model suitable for incorporation into a whole-of-system model, implying modest computing demands and reduced complexity. Finite-element analysis was not used. Because the apparent kinetics of a practicable metal-hydride tank is dominated by heat flow originating in the enthalpy of hydrogen absorption/desorption, particular attention was paid to modelling the effective thermal conductivity, based on a detailed description of the thermal resistance between hydride particles. The model was tested against our own experimental data for a pair of tanks with 6.4 kg total hydrogen capacity, and compared with a published 2D model and published experimental data. In all cases, the new model performed very well. The incorporation of a physical model of the effective thermal conductivity means that the tank model can also be used as a research tool to investigate ways of improving tank performance by altering the physical characteristics of the metal-hydride itself to achieve an optimal thermal design and enhanced reaction kinetics. Highlights: One-dimensional model with low computing demand and no finite-element analysis used. Physically based with detailed model of effective thermal conductivity. Tested against real metal-hydride tanks, not laboratory-scale units. Works at least as well as more complex FEA models in 2 and 3 dimensions.