The operation of PEM electrolyzers under intermittent renewable sources introduces dynamic load regimes that challenge design limits and traditional safety protocols. Although it is recognized that fluctuations in pressure, temperature, and current accelerate degradation mechanisms, the quantitative characterization of this operational "stress" still lacks systematic empirical evidence. This work presents a comparative statistical analysis of the operational variables of a 1.25 MW PEM electrolyzer (NLR dataset), contrasting steady-state and intermittent operation regimes. Through distribution analysis, temporal variability, and cross-correlation techniques, the deviation of system behavior from the nominal operational context is quantified. The results provide empirical data for calibrating Probabilistic Safety Assessment (PSA) models, enabling a more precise definition of safety margins and failure rates in green hydrogen production environments subject to high energy variability.