In the real-time energy management of offshore islanded microgrids, determining the optimal operating points of storage systems, particularly in hydrogen-based storage, poses significant challenges. These arise from the stochastic nature of offshore Renewable Energy Sources (RES), variable power demand, and the volatility of hydrogen systems. To address these, a novel objective function has been developed that integrates electrolyzer system efficiency into the Energy Management Strategy (EMS) of a DC microgrid. Unlike most existing literature, which considers electrolyzer efficiency as a constant, this work treats efficiency as a dynamic variable that depends on operating current, temperature, and pressure. The behavior of the electrolyzer efficiency with respect to these parameters is modeled, verified, and subsequently incorporated into the EMS. A MINLP based EMS is developed to implement the proposed formulation, and its effectiveness is validated by demonstrating optimal microgrid performance under the above scenario. Notably, the impact of optimal temperature and pressure control is evidenced by electrolyzer efficiency improvements of 1.7% and 3.1% in the efficiency-focused and multi-objective scenarios, respectively. The proposed EMS is evaluated against a rule-based and heuristic (PSO) methods. In the cost-based case, the developed method shows 36.5% and 0.04% cost reductions relative to the rule-based approaches, and a 0.39% reduction relative to PSO. In the efficiency-based scheme, it attains 0.10% efficiency gain over PSO for the optimal temperature-pressure method, and in the multi-objective case, it delivers 7.40% efficiency gain versus PSO. • Novel optimization with electrolyzer efficiency for offshore islanded DC microgrids. • Electrolyzer efficiency modeled dynamically via current, temperature, and pressure. • Multi-objective framework balances cost and efficiency in flexible microgrid EMS. • Sensitivity to electrolyzer operating conditions is analyzed. • EMS validated against rule-based and metaheuristic approaches.
Indrajith et al. (2026) studied this question.