Analysis of future residual loads in Germany for 2030-2050 and possible accuracy gains of their modelling

The future German energy system will focus heavily on renewable energies in order to comply with the ambitious climate protection targets. A key challenge of this energy transition is ensuring a reliable coverage of residual load through renewable energy sources. This thesis addresses two research questions. Firstly, it is examined how the residual load will evolve in the future and derives implications for the design of the future energy system. Secondly, it will be investigated if significant accuracy gains can be achieved in the modelling of the future residual load. To answer this, a simulation model was developed to estimate the future electricity generation from PV and wind power, taking into account three different scenarios. This generation is offset against the electricity load to calculate the residual load at an hourly resolution. The model was implemented in Python, with supplementary analyses conducted in Excel. The spatial granularity distinguishes between an aggregated national level and a federal state specific level. The renewable capacity data were sourced from the Marktstammdatenregister. The different scenarios are compared with respect to their results as well as modelling accuracy. The analysis reveals a significant increase in negative residual load in the future, while variations in positive residual load remain comparatively smaller. This is due to a significant temporal mismatch between high renewable electricity generation and high electricity demand. These findings highlight the urgent need for advanced flexibility options, such as energy storage, demand-side management, and sector coupling, to ensure a stable renewable-based energy supply. Furthermore, the results indicate that higher spatial granularity and the consideration of future technological advancements can improve the modelling accuracy. Future research should focus on optimizing the design and implementation of flexibility options. The proposed model provides a solid foundation for further refinements and scenario-based evaluations.