BiGPlan – Big Data for Distribution Grid Planning

Due to increased sector coupling loads in low voltage distribution are raising fastly, which results in a challenging grid reinforcement task. Precise grid modelling may at least avoid unnecessary grid expansion. Therefore, a realistic, spatially localized, time-series-based modeling of the low-voltage grid under the consideration of operational management is developed.

The energy, transport and heating transition are a challenge for electricity grids and their operators. According to current planning guidelines, the distribution grids would have to be massively expanded in the event of a high penetration of heat pumps and electric vehicles. This leads to considerable costs and, in the worst-case scenario, can significantly slow down the transition to a CO₂-neutral society. As current planning guidelines are based on unrealistically high load assumptions and therefore postulate an excessive need for grid expansion, grid expansion can be reduced by using realistic power time series for consumers and producers. Smart metering systems can be used to reduce grid loads in a targeted manner through control interventions. In order to reduce grid expansion, this must already be taken into account in grid planning. The aim of the project is to validate the modeling of representative grids with and without grid operation management through measurement campaigns in real laboratories. Finally, generalizable planning principles are to be derived from this. The first step is to characterize the grid clients as good as possible using big data methods, whereby »state of the art« methods are improved in order to map demographic change and commuting routes, for example. The collected data is then merged and calibrated with the internal, automatically processed information of the distribution system operators (DSOs). The load profile generator »synPro« from Fraunhofer ISE can be used to generate consumption, generation and flexibility time series for the grid clients characterized in this way. By feeding back measured values, synPro is further improved so that local specifics can be taken into account for short and medium-term forecasts. The improved time series are used for grid modeling, which is then validated. Finally, network planning principles are derived.