A scalable simulation method for cyber-physical power systems
Happ, Sonja; Monti, Antonello (Thesis advisor); Benigni, Andrea (Thesis advisor)
1. - Aachen : E.ON Energy Research Center, RWTH Aachen University (2020, 2021)
Book, Dissertation / PhD Thesis
In: E.ON Energy Research Center 85
Page(s)/Article-Nr.: 1 Online-Ressource (xi, 151 Seiten) : Illustrationen, Diagramme
Dissertation, RWTH Aachen University, 2020
Electrical power systems undergo a paradigm change from unidirectional to bidirectional power flow. This change is caused by an increasing number of distributed energy resources installed especially in distribution systems. System operators are forced to rethink the traditional ways of system operation since these are not designed for a bidirectional power flow. Research has identified information and communication technology as a key enabler for the integration of distributed energy resources in the system operation through communication and data processing. The coupling of an electrical power system with a communication and data processing system results in a cyber-physical power system (CPPS). The overall behavior of a CPPS emerges from the interactions of its components through physical coupling and communication. Modeling and simulation of CPPS have to consider these interactions as the source and the means to obtain the overall system solution. The potentially large number of components and their heterogeneity constitute severe challenges that require solutions for both the modeling and the computation of the simulation itself. This dissertation identifies the following four major requirements of CPPS simulation based on a review of existing approaches and their shortcomings. The requirement of scalability (1) is motivated by the expected number of system components. The system solution has to emerge from the interactions of its components (2). Large-scale simulations require a scalable and flexible management of input and result data (3) that is not available in existing solutions. Interoperability (4) is required to enable the coupling with other simulation tools or data processing systems for CPPSs. As the main contribution of this dissertation, the simulation method Distributed Agent-based Simulation of Complex Power Systems (DistAIX) is proposed. It enables the simulative assessment of CPPSs at an unprecedented scale without neglecting interactions through simplification or aggregation of components. It utilizes parallel computing to achieve scalability and to overcome the limited processing power and memory resources of single computing nodes. Agent-based modeling and simulation addresses the requirement of emergence of the system solution from interactions. System components are modeled as agents that include an electrical model and a specific behavior. They are distributed to processes for the computation of the simulation. DistAIX is embedded in a framework facilitating a flexible scenario creation and a database-supported result storage. Results can be post-processed in a dedicated application that abstracts from the agent-based modeling approach and the database storage. A generic communication interface is integrated to exchange data with external systems during a simulation. The DistAIX method is introduced and assessed comprehensively throughout this dissertation with respect to all four identified CPPS simulation requirements.