The NTC model considers a fixed capacity for each border independently, whereas the FB model describes electric lines which may be freely used for one border or another, thus linking exchanges with each other (ex: in NTC, France ? Belgium exchanges cannot exceed 2000MW, and France ? Germany 1500MW. In FB, the overall capacity from France to Germany+Belgium is about 4000MW and can be used for both borders with power transfer distribution factor (PTDF), leading the maximum cross-border flows with Germany and Belgium to be linked). Flow-based is the preferred approach for day-ahead capacity allocation (cf. EU electricity target model).
In order to ensure the comparability between the NTC-based and FB models, the cross-border capacities (NTC limits and critical branches limits) have been defined to mimic the same underlying physical network . The coherence of both cross-border domains is a necessary condition to ensure that the differences in terms of performance exclusively come from the intrinsic characteristics of each approach and not from differences in the underlying physical network. Generation units were aggregated (from 620 realistic ones into 90 aggregate ones) in order to reduce the computation time. When changing from NTC to FB, importing countries are shown to be able to import even more energy.
Cheap generation is nuclear, hydro and intermittent renewable, whereas more expensive generation is mostly conventional thermal; as a result, increasing the cross-border capacity led to a decrease in CO2 emissions. Should the cheap generation be a high CO2 emitter, increasing cross-border capacity would lead to an increase in CO2 emissions. Total generation costs decreased between 0.25 and 6%, depending on the studied renewable energy development level and season; social welfare  usually increased by approximately 2M€ per day (by decreasing -resp. increasing- prices in importing - resp. exporting – areas). Besides, the frequency of negative price peaks coming from time steps with very high renewable levels (in our case in Germany) decreased when moving to FB, because of the better use of the available cross-border capacity to evacuate this energy .
As far as network congestions are concerned , the FB method leads to more time steps with no congestion at all; this effect is due to the fact that, in FB, capacity which is not used for a given border can be used for another, whereas this switch is impossible in NTC. The number of time steps with full congestion (i.e. four different prices) also increased; this increase also comes from the fact that all cross-border flows are linked with each other: in some cases, it may be more profitable to decrease the capacity of the market area A- market area C border (leading to slight prices differences between these two countries) in order to export along the market area A-market area B border (because market area B’s price may otherwise be much higher than the other two areas’ prices).
The cross-border model also impacted the redistribution of social welfare among countries and stakeholders. Consumer surpluses increased in importing areas, while producer surpluses increased in exporting areas. The congestion revenue was assumed to be uniformly spread among the four countries. The overall congestion revenue decreased, mostly because price spreads (across borders) decreased when moving to FB. Not all market areas benefited from the change to FB in terms of social welfare: the social welfare decreased in Germany and increased in all other countries (this result is specific to our set of generation capacities). This decrease may lead to opposition from the regulator or policy makers in Germany against such a cross-border model change.
Overall, switching from NTC-based to Flow-Based market coupling increased overall economic efficiency (although some individual stakeholders’ surpluses decreased) and improved the renewable energy integration (because cross-border flows better “adapted” to varying intermittent generation). CO2 emissions decreased because, in this study setup, cheap generation did not emit much CO2. The effect on security of supply remains unclear, especially because the physical network average use is higher with the FB approach (but the TSO is assumed to have tendered some margin on all critical branches before the market coupling).
 The FlowBased domain was designed to exactly enclose the NTC domain.
 Social welfare is computed as the sum of consumer surplus, producer surplus and congestion revenue.
 The physical network capacity did not change in itself, but it was better used, leading to more exports from Germany during these time steps (and probably less from other countries at that time).
 Here, a border will be congested when the two neighbouring areas’ prices are not equal.