[Authors]

Fabian Pascher* (IWES), Lennard Giesenberg (IWES), Alexander Spieß (IWES), Dr. Wolfram Münchgesang (IWES), Prof. Dr. Mathias Seitz (HoMe | IWES)

[Affiliations]

Fraunhofer-Institut für Windenergiesysteme (IWES) - Am Haupttor 4310 06237 Leuna

Hochschule Merseburg (HoMe) - Eberhard-Leibnitz-Str. 2 06217 Merseburg

[Background]

To understand the behavior of water electrolyzers under stress caused by fluctuating loads, a model-based analysis was performed. Thereby significant changes in electrochemical cell parameters were observed. This knowledge allows the development of reliable electrochemical models that can be used to design, optimize, and control the water splitting reaction in PEM electrolyzers.

[Methods]

A model-based approach has been used to deduce a measurement plan for accelerated stress tests and an analytic sequence of the electrolyzer states. Main parts of the analytic sequence were:

1. Exhaustive statistical analysis of the measured operating parameters

2. Deduction of a stochastic static model

3. Nonlinear regression and probabilistic analysis

4. Sensitivity study

[Results]

The evaluation of electrical impedance and galvanostatic data lead to consistent electrochemical and kinetic parameter guesses and offered a statistical understanding of parameter correlations and structural errors inside the models. The main achievements were:

1. Model discrimination based on measured variables

2. Comparable states (new system, aged system)

3. Parameter changes due to accelerated stress test

[Deduction]

The analytic approach gives insight in parameters that are not recorded by measurement directly. It allows the deduction of parameter variations over time & current, enables refining the measurement, and the models used for the analysis.