A multi-objective framework for accurate parameter identification in Metal Hydride hydrogen tank

Accurate modeling of Metal Hydride (MH) hydrogen storage systems requires reliable estimates of several physico-chemical and flowed parameters, many of which cannot be measured directly. In this work, the Non-dominated Sorting Genetic Algorithm III (NSGA-III) is employed to perform multi-objective parameter identification on a coupled thermo-reactive model that integrates mass and energy conservation with detailed absorption kinetics. A zero-dimensional (0-D) modeling approach is adopted to describe the dynamic behavior of the storage tank. Three objective functions targeting equilibrium pressure, MH temperature, and gravimetric hydrogen capacity are simultaneously minimized to reduce error between simulations and experimental measurements. This identification yields a well-defined Pareto front and an optimal compromise solution achieving Mean Absolute Percentage Error (MAPE) of 3.33%, 0.14%, 5.32% respectively. Sensitivity analysis reveals a pronounced shift in the influence level of parameters governing the behavior of the MH system. Operational parameters (water and hydrogen mass flow m ̇ w , m ̇ H2 respectively) and structural parameters (metal alloy density at saturation ρ s s , and porosity ɛ ɛ ) progressively dominate system performance, whereas intrinsic thermochemical properties (enthalpy Δ H , kinetic constant of the reaction C a , activation energy E a ) lose relative significance. Overall, the proposed optimization framework enables accurate parameter identification and offers a robust tool for the design, monitoring, and long-term performance assessment of MH hydrogen storage tanks.