Dynamic modeling and fault diagnostics of large-scale alkaline electrolyzers

The thesis develops an advanced dynamic model of large-scale alkaline electrolyzers, aimed at creating a digital twin for operational analysis and fault diagnostics. The model integrates, through co-simulation, the MATLAB/Simulink environment for electrochemical and control modeling with MATLAB/Simscape for the physical representation of the Balance of Plant, including gas–liquid separators, heat exchangers, pumps, and valves. This architecture enables the capture of multi-domain interactions among thermal, hydraulic, and electrical variables, overcoming the limitations of traditional lumped models. The digital twin is used to analyze system behavior under realistic operating conditions, particularly under variable renewable energy profiles, highlighting impacts on efficiency, stability, and operational limits. In parallel, a model-based diagnostic approach is developed, relying on residuals, thresholds, and fault signatures to detect and isolate anomalies even in the absence of real plant data. The results demonstrate the effectiveness of the model as a tool to support the design, control, and maintenance of green hydrogen production systems.