Time-series based distribution grid planning considering interaction of network participants with a multi-agent system

With the expansion of distributed generation and innovative loads being connected on a low to high voltage level, the supply task for the distribution grid becomes increasingly volatile, which requires an adequate consideration in the planning process. However, since the primary objective of the distribution grid was the supply of loads in the past, the conventional planning process was simplified to dimensioning cases and the individual voltage levels were planned separately. Therefore, innovative and interacting network participants can only be accounted for with a basic estimation and the interdependency of voltage levels is neglected. Facing these challenges, a simulation framework based on the concept of multi-agent systems is developed in this thesis to determine the time dependent behaviour of all network participants and enabling the modelling of innovative market incentive concepts. A single agent, with individual objective functions and environmental conditions, represents every grid user. The resulting time series constitute a profound basis for a demand oriented distribution grid planning process, considering the probability of occurrence of network-loading situations.