Uncertainty Quantification for Modular Co-Simulation

Interdisciplinary research is becoming more and more important in a world full of complex systems of systems. One prominent example of such multidisciplinary systems are smart grids, defined here as ICT-based energy systems of the future. A promising method for analysis and development of these systems is co-simulation. It allows the reuse of existing and established models of different system components and couples them in a common simulation environment, transcending differences in programming language and modeling paradigm. So-called co-simulation frameworks provide modular simulator coupling and standardized interfacing to boost collaboration. A big caveat of co-simulation, however, is the so far insufficiently covered quantification of model and data uncertainty. Such a procedure would be required for comparison of simulation output and real-world behavior and is thus the basis for assessment of the significance of tests. Uncertainty quantification is already an intensively studied topic in monolithic simulation, but it has so far received little attention in the smart grid co-simulation domain.
The presented work proposes a structured concept of an uncertainty quantification system for flexible smart grid co-simulation. It builds up on a thorough analysis of system requirements and the state of the art. The design of the system itself is strongly based on the notion of modularity, considering the possibility of diverse uncertainty sources and representations (probabilistic as well as non-probabilistic). A prototypical implementation demonstrates the concept's functionalities and applicability. With its extensive outlook, the thesis lays the groundwork for future research and development in the field of uncertainty quantification for co-simulation.