Comparative study of direct and multi-stage seawater electrolysis processes for hydrogen production

The transition to a carbon-neutral energy system increasingly relies on green hydrogen produced via water electrolysis. Seawater, as an abundant feedstock, offers a compelling alternative to freshwater but introduces challenges—chiefly chloride-induced side reactions, membrane fouling, and electrode corrosion. This thesis presents a comprehensive MATLAB/Simulink model to compare two process configurations for seawater-to-hydrogen: (1) a direct electrolysis route comprising only coarse and fine filtration before electrolysis, and (2) a multi-stage treatment which integrates filtration, pretreatment, reverse osmosis (RO), and electrodeionization (EDI) prior to electrolysis. Both configurations are evaluated in terms of energy consumption, system efficiency, water demand, and estimated operating expenditures. A combined sensitivity and uncertainty analysis quantifies the influence of key parameters such as salinity, current density, membrane type and performance, and Faradaic efficiency on overall performance and cost. Results show that the direct route minimizes pretreatment energy, but experiences accelerated electrolyzer degradation under high-salinity conditions. Conversely, the multi-stage train extends electrolyzer lifetime at the expense of higher pretreatment energy and capital cost. Importantly, the energy and cost required for seawater desalination are negligible compared to those of the electrolysis step itself. A trade-off analysis identifies critical salinity thresholds and operating regimes in which each configuration is preferable. These findings provide practical guidelines for designing and optimizing scalable, low-carbon seawater-to-hydrogen plants that balance simplicity, energy use, and equipment longevity. Future work should investigate the relationship between feedwater purity and electrolyzer degradation rates, assess product-gas purity under varying pretreatment schemes, and refine hydrogen-production cost models under dynamic operating conditions.