Dynamic model-based state-of-charge estimation for an organic redox flow battery: Case study on the transferability of concepts for vanadium redox flow batteries to organic redox flow batteries

The state of charge (SoC) and charge capacity are key variables to ensure the efficient operation of organic redox flow batteries. There are many methods available for lab scale investigations, enabling the determination of these parameters. However, most of these cannot be implemented in full systems due to complexity (e.g., NMR spectroscopy), while other require installation of bypasses and additional instrumentation (e.g. UV–Vis spectroscopy). Non-invasive, model-based observers are therefore highly needed to estimate especially SoC. Here, we demonstrate how dynamic model-based SoC estimation methods can be applied to organic redox flow battery technology. The model developed has been adapted from vanadium technology. Some modifications have been implemented, especially with regards to the crossover. The model is parametrized and verified with experimental data on a stack level. Finally, one promising observer using redox potential measurements of both electrolytes was adopted based on the literature on vanadium redox flow batteries. Two extended Kalman filter approaches using the same observer structure were also developed, differing only in terms of the model used to estimate the redox potential. The validation with experimental data shows that the adopted literature-based approach is highly sensitive to model errors, while the extended Kalman filter approaches seem more robust.