National energy system modeling for decarbonization pathways considering distribution grids

Abstract

Long-term planning tools compute investment in energy technologies to reach techno-economic objectives. These tools are typically used to define decarbonization pathways with the corresponding capacities for several types of technologies. However, for most renewable-based resources that are expected to be connected in distribution grids, the typical temporal and geographical resolutions of long-term models are deemed insufficient. They cannot capture the need for dynamic flexibility and/or grid reinforcement in distribution and thus the impact on the final costs or carbon emissions. This paper assesses the integration of distribution grids in long-term energy system planning. A coupling is proposed between the long-term model POLES, which computes the technology capacities to install each year up to 2050, with the dispatch/investment model Backbone, which optimizes the operation and the repartition of investments at finer resolutions. The coupling being computationally demanding, several simplifications of Backbone are investigated to find a trade-off with accuracy. Results show that simplifying the investment options and distribution grid models can achieve a 90 % precision on investment and energy generation compared to baseline optimization, with a computational time divided by 100. Finally, comparative simulations are run with the models coupling up to 2050 for the French power system with and without considering medium voltage distribution grids in addition to the transmission one. In the latter case, minor changes occur in the capacity investments. However, operational constraints increase the power system decarbonization cost by 20 % due to distribution constraints, underlining the necessity of considering finer temporal and geographical resolutions.