Semantic and Physical Modeling and Simulation of Multi-Domain Energy Systems: Gas Turbines and Electrical Power Networks

Abstract: The ITEA3 OpenCPS (Open Cyber-Physical System Model-Driven CertifiedDevelopment) project focuses on interoperability between the Modelica/UnifiedModeling Language (UML)/Functional Mock-up Interface (FMI) standards, improved(co-)simulation execution speed, and verified code generation. The project aims to developa modeling and simulation framework for cyber-physical and multi-domain systems. Oneof the main use cases for the framework, is the multi-domain equation-based modelingand simulation of detailed gas turbine power plants (including the explicit equation-basedmodeling of turbomachinery dynamics) and the electrical power grid.In this work, UML class diagrams based on the Common Information Model (CIM)standard are used to describe the semantics of the electrical power grid. An extension basedon the standard ISO 15926 has been proposed to derive the multi-domain semanticsrequired by the models that integrate the electrical power grid with the detailed gas turbinedynamics.Furthermore, the multi-domain physical modeling and simulation Modelica language hasbeen employed to create the equation-based models of the use case of this project. Acomparative analysis between the Single-Domain and Multi-Domain model responses hasbeen performed both in time and frequency. The results show some interesting differencesbetween the turbine dynamics representation of the commonly used GGOV1 standardmodel and the less simplified model of a gas turbine.Finally, the models from each domain can be exchanged between two differentstakeholders by means of Functional Mock-Up Units (FMUs), defined by the FMIstandard. Promising test results were obtained with different simulation tools that supportthe standard, which demonstrates the feasibility of exchanging unambiguous multi-domainmodels with a detailed gas turbine representation. This shows the potential of the FMIstandard for manufacturers to exchange equation-based multi-domain models, while at thesame time protecting their intellectual property.