Online observer of the voltage components of the PEM electrolyzer based on the time-varying linearization of the semi-empirical model

This paper presents a method to estimate the components of the PEM electrolyzer voltage with the dynamic input current. The total voltage can be divided into reversible voltage and irreversible voltage that the former always can be described as a function of temperature and pressure. For the irreversible voltage, a dynamic semi-empirical model is linearized by a dynamic expansion point varying with the input power. A Luenberger observer is established from the linearized model to estimate the anode overvoltage, cathode overvoltage and ohmic overvoltage effectively and rapidly, for which the stability is proved by LMI approach. Considering the deviation caused by local linearization, a switching scheme of the expansion point is also analyzed to provide an appropriate solution for fluctuating current when accessed to renewable energy power. Simulations have been carried out to testify and verify the method with varying input current, different switching expansion points, and uncertain parameters. The results show that the online Luenberger observer can achieve fine accuracy. And for the varying input current, updating the expansion point with a certain scale can further improve the observation accuracy, which enhances the performance of the observer to adapt the hydrogen production with renewable energy to evaluate the behavior of the PEM electrolyzer. (c) 2023 The Author(s). Published by Elsevier Ltd. This is an open access article under theCCBYlicense (http://creativecommons.org/licenses/by/4.0/).