Novel network reduction method for cellular-based network models with enhanced modeling accuracy for multi-energy-system approaches

[Display omitted] •Extended electric grids with fine time resolution require high computational effort.•Cellular approaches model real grids with decreased complexity for fast calculations.•Aggregated models require network reduction to preserve electrical behavior.•The presented network reduction method enables high modeling accuracy for cell models.•This method offers additional advantages for certain application purposes. Sustainable electricity supply can be achieved by expanding renewable energy sources (RES), which due to their volatile nature present challenges for the power grids. Modern electricity grids must be able to coordinate and balance these unpredictable generation patterns. One option for grid-friendly RES integration is multi-energy systems (MES) which enable energy use across energy carriers and, thus, relieve electrical grids. Therefore, long-term simulations of MES are indispensable and require appropriate models. Since MES have high systemic complexity due to high temporal resolution, spatial coverage, and hierarchical depth, modeling requires massive computational effort, impeding the investigations. Models with reduced complexity (spatial resolution reduction) and, thus, reduced computational effort can be created using a cellular approach. In particular, for electrical grids, reduction of complexity requires network reduction methods that create equal grid models regarding electrical behavior. Since most (numerical) network reduction methods (e.g., REI, WARD method) fail to replicate all required parameters, this work presents a novel network reduction method enhancing modeling accuracy, primarily regarding reactive power. The introduced method is based on a detailed parameter analysis to identify parameters responsible for deviations between original grid and reduced cell model. The validation of this method uses test networks at different voltage levels to reveal influencing variables that enhance modeling accuracy. This allows to derive trends for modelling accuracy of individual electrical parameters. The introduced method facilitates developing cell models for time-series-based calculations with maximum modeling accuracy and reasonable calculation effort. Additionally, this paper presents advantageous application purposes of this method.