When modelling energy system components for use in energy management systems, single-value coefficients or linearised characteristic curves are often assumed for the operating range. Frequently this approach proves to be too imprecise in predicting the system status. However, when determining operating point-dependent efficiencies, there are often gaps in the available information which can only be filled with an investment of time and effort. To address this, a parameter identification approach to determining the operating point-dependent efficiencies of energy converter systems was implemented and evaluated. Data from a vanadium redox flow battery out of the “Smart Region Pellworm” research project was used as the research object. A genetic algorithm was used to parameterise a simplified electrochemical surrogate model. With the help of the parameterised surrogate model, efficiencies and power limits were sampled over the entire operating range to create an easily transferable battery model for the prediction of state of charge responses to known unit commitment plans. A model with a fixed efficiency for charging and discharging and a model whose efficiency map is based on an experiment protocol were used as benchmarks for the calculated efficiency map. The approach was evaluated in terms of transferability and model error.