European-scale power flow studies are usually made with models representing the extra high voltage (EHV) grid, ranging from 220 to 400 kV, like the TYNDP [1] or PyPSA-Eur [2]. In those models, sub-transmission grid (63 to 150 kV) and distribution grid are fully aggregated to the EHV grid (“Simplified grid”). However, the current energy transition requires a sharp rise in renewable energy sources (RES), often connected to sub-transmission and below. Due to the expansion of RES, new congestions are being observed by system operators (SO), notably on sub-transmission. A comparative study of sub-transmission grids in Germany, Italy, and France is given in the first part for a more detailed context, with a focus on France. Hence, adapting the modeling of European grids to better reproduce RES behavior and their impact on power flows is a new challenge. However, for long term studies (5 to 10 years prospective) with many scenarios, a complete modeling of sub-transmission as made in [3] would induce a consequent rise in complexity (“fully disaggregated grid”). An intermediary proposition would be to apply a selective spatial disaggregation, e.g. modeling explicitly the sub-transmission on a geographically limited area among the main EHV grid model (creating a “partially disaggregated grid”). Only two types of zones would be targeted: -Sub-transmission grids at risk of congestion, -Sub-transmission grids directly connected to nodes that have the strongest influence on regularly congested EHV lines. In the first case, it would disclose sub-transmission congestions and the potential curtailment induced, which remains invisible in the “simplified grid” model. In the second case, it would verify if a more precise model of the congestion’s surrounding tempers or increases the congestion. This type of “partially disaggregated grid” models selectively captures new grid constraints while containing complexity increase. It can help RES producers to position new assets and guide the development of flexibility means. The article proposes a frame to calculate the benefits and costs of enhancing EHV grid models through a selective spatial grid disaggregation The indicators proposed are valid for EHV models across countries, with different voltage levels, considering their respective weights. They enable a comparison between several local disaggregation possibilities to identify those with the strongest impact. A combination of local disaggregations forms a “partially disaggregated grid”. These indicators are calculated on an example grid, illustrating impacts of a local disaggregation through a sub-transmission mesh with a congested line.