The development of efficient electric vehicle (EV) charging infrastructure requires a modeling of customer behavior at an appropriate level of detail. Since only limited information about real customers is available, most simulation approaches employ a stochastic approach by combining known or estimated customer features with random variations. A typical example is to model EV charging customers by an arrival and a targeted departure time, plus the requested amount of energy or increased state of charge (SoC), where values are drawn from normal (Gaussian) distributions with mean and variance values derived from user studies of obviously limited sample size. In this work, we compare this basic approach with a more detailed customer model employing a multi-agent simulation (MAS) framework in order to investigate how a customer behavior that responds to external factors (like weather) or historical data (like satisfaction in past charging sessions) impacts the essential key performance indicators of the charging system. Our findings show that small changes in the way customers are modeled can lead to quantitative and qualitative differences in the simulated performance of EV charging systems.

The interdisciplinary modelling language Modelica is increasingly used in the design and evaluation of energy systems. Heat supply represents a considerable share of the global energy supply. Especially in European cities, district heating grids are often used and implemented for heat coverage. The increasing integration of renewable energies and the extension of existing grids require engineers to be able to analyze and evaluate the behavior of such grids, not only statically in certain operating conditions, but also dynamically to enable the representation of complex system interaction. This paper shows and describes a new approach as to how Modelica models can be used to evaluate the dynamic behavior of district heating grids. It furthermore introduces a consistent framework to parameterize these models with GIS-data via the COM interface. The advantages of the shown approach compared to previously used static methods are shown with specific case studies.

No abstract