Master's thesis Frank Tilger


Development of a market-based Multi-Agent System for a distributed Smart Grid

A Diagramm Copyright: EBC

In this thesis a market-based multi-agent approach is implemented to coordinate the energy supply and demand of a distributed energy system ("microgrid"), which comprises 100 residential buildings. The energy management approach is focused on the coordination of electro-thermal heating devices, such as combined heat and power (CHP) units and heat pumps (HP), which are installed in each building of the microgrid. The goal of the system is to enhance the integration of renewable energy sources and to reduce the required electricity from the conventional grid in order to increase the autonomy of the microgrid. This is achieved by establishing a local electricity market on which the houses are able to trade electricity with each other. The variable market price serves as an incentive for the participants to shift the operation of their heating devices. This basic model is extended by a price anticipation model, in which the single houses react to a short price prediction in order to increase their profit and to further improve the autonomy of the microgrid.

In order to evaluate the impact of the energy management, the two coordinated scenarios (basic model and price anticipation model) are compared to a reference scenario, which represents the energy system of today, i.e. no energy management and fixed prices for private customers. Furthermore, the effect of different shares of renewable energy sources is analyzed.

The results show that energy management can help to reduce the electricity import of a microgrid by around 60%, while the export can be decreased by up to 80%. Besides, the integration of the price anticipation model can further diminish the electricity export by more than 3%, so that in total the dependence on the conventional grid is reduced significantly. From an ecological perspective, a CO2 reduction of 13% can be achieved, which is mainly attributable to the enhanced integration of renewable energy sources. Furthermore, the annual energy costs of the single houses can be reduced considerably compared to the reference scenario. Accordingly, for the HP houses annual savings between 12% and 28% are feasible, while the CHP houses exhibit savings between 8% and 17%. All in all, the coordinated scenarios feature economical, ecological and energetic advantages in comparison to the reference scenario.