Multilinear modeling and simulation of a multi-stack PEM electrolyzer with degradation for control concept comparison

Abstract

Hybrid energy systems, e.g., with wind energy and hydrogen production, have a high model complexity due to their multi-physics nature, which poses major control challenges for the optimization of plant operation. This work aims at addressing this issues by introducing a highly efficient modeling and simulation framework. A proton exchange membrane (PEM) electrolyzer stack, including degradation and controller, has been modeled using the multilinear class. This class enables the automatic append of individual models, which is used to stack a 100 multi-stack PEM electrolyzer model. Moreover, the multilinear class models can be represented as tensors, which allows for efficient decomposition methods and formats. This is used to considerably enhance the simulation performance of the system, making the simulation of a one year multi-stack electrolyzer operation possible, with a reasonable computational cost. In the simulation, two different high-level control modes are compared regarding overall degradation gain and electrolyzer efficiency. The developed modeling and simulation framework has proven its suitability for big-scale complex models, enabling efficient simulations for controller analysis.