Marchés pair-à-pair de l’électricité dans les réseaux électriques

The deployment of distributed energy resources, combined with a more proactive demand side management and energy management systems, is inducing a new paradigm in power system operation and electricity markets. Within a consumercentric market framework, peer-to-peer approaches have gained substantial interest. Peer-to-peer markets rely on multi-bilateral negotiation among all agents to match supply and demand. These markets can yield a complete mapping of exchanges onto the grid, hence allowing to rethink market–grid interactions.This thesis treats three main challenges which needs to be overcome before considering real world implementations: (i) scalability to host a growing number of distributed users and resources, (ii) compatibility with grid constraints, and (iii) resilience to stochastic power injections. After a complexity analysis, scalability of peer-topeer markets and the proposed negotiation mechanism to solve them is enhanced by three improvements reducing algorithmic and structural complexities. Feasibility of the peer-to-peer electricity market is eventually obtained with the use of network charges. Two approaches are proposed to handle these network charges. The first, exogenous, requires the system operator to provide them a priori before negotiations start. In the second, the system operator updates network charges endogenously at each iteration to better account for the current grid status. Finally, power forecasts of stochastic agents are taken in a more comprehensive way by the developpement of peer-to-peer market on both energy and capacities, used to restore power balance in case of misdipatch due to forecast errors.