Renewable energy sources are usually heterogeneous and owned by various parties. These RES need to be controlled taking this diversity and decentralization into account, while, in addition, acting fast on changes in the energy system. This project focuses on the development of fast, market-based control approaches in multi-agent systems, where the aim is to design control approaches for grids with numerous and heterogeneous RES, owned by different parties.
The objectives of the project is to design such market mechanisms as well as agent strategies and agent decision modules that can act on a time scale of at most seconds or in fractions of these. This means that the solutions to be designed need to: (1) contain a type of market mechanism for which the allocation or adjustment of generation per party is computed efficiently, i.e. in at most seconds; (2) have this market mechanism regulate the overall power injected into the grid in an effective way, yielding stable behaviour of the grid as well as optimization of the electricity production; (3) can be computed efficiently by individual agents, especially regarding the agent strategies for acting in these markets and controlling their local generation; and (4) have a low communication complexity within the agent system.
Tasks
- Modeling smart grids with numerous, heterogeneous RES owned by different parties, for control purposes.
- Developing fast and efficient control strategies using market-based agent systems that can act on a time scale of at most seconds.
- Developing agent strategies and agent decision modules that can act in these markets in fractions of these time scales.
Expected Results
- Fast decentralised control solutions for renewable generation, based on market-based agent systems, with efficient computation and communication complexity and with effective allocation of local generation and/or adjustments of the injected power into the grid.
- The control solutions consist of models, software prototypes and extensive evaluation of these by simulation and analysis.
Early Stage Researcher:
H.A.M.F.Abdelghany@tudelft.nl - -
Supervisors:
Han La Poutré
Simon Tindemans
Host institution:
