Session 5B - Open standards (2) FMI/DCP

Title:	A Cloud-native Implementation of the Simulation as a Service-Concept Based on FMI
Authors:	Moritz Stüber and Georg Frey
Abstract:	Providing modelling and simulation capabilities as a service promises to increase their value by improving accessibility for non-expert users and software agents as well as by leveraging cloud-computing technology to scale simulation performance beyond the capabilities of a single computer. In order to reach this potential, implementations must align their design with the architectural styles of cloud computing applications and the web in general. We present an open-source, cloud-native Simulation as a Service (SIMaaS)-implementation that gives access to models and allows simulating them on the web. The implementation uses Functional Mockup Units (FMUs) for co-simulation as an executable form of a model and relies on FMPy for simulation. It is realized as a microservice in the form of a REST-based HTTP-API. Functionality and performance are demonstrated by using the service to create ensemble forecasts for PV systems and to search for an optimal parameter set using a genetic algorithm. Conceptual limitations and the resulting opportunities for further work are summarized.
Keywords:	simulation as a service, cloud-native simulation, service-oriented software architecture, FMI 2.0
Paper:	full paper
Bibtex:	@InProceedings{modelica.org:Stuber:2021, title = "{A Cloud-native Implementation of the Simulation as a Service-Concept Based on FMI}", author = {Moritz St\"uber and Georg Frey}, pages = {393--402}, doi = {10.3384/ecp21181393}, booktitle = {Proceedings of the 14th International Modelica Conference}, location = {Link\"oping, Sweden}, editor = {Martin Sj\"olund and Lena Buffoni and Adrian Pop and Lennart Ochel}, isbn = {978-91-7929-027-6}, issn = {1650-3740}, month = sep, series = {Link\"oping Electronic Conference Proceedings}, number = {181}, publisher = {Modelica Association and Link\"oping University Electronic Press}, year = {2021} }

Title:	Python Framework for Wind Turbines Enabling Test Automation of MoWiT
Authors:	Johannes Fricke, Marcus Wiens, Niklas Requate and Mareike Leimeister
Abstract:	The development and simulation of engineering systems, especially wind turbines, is becoming increasingly complex and elaborate. At the Fraunhofer Institute for Wind Energy Systems (IWES), the in-house developed tool MoWiT (Modelica library for Wind Turbines) is being developed for load simulation. MoWiT is based on Modelica and is constantly evolving. It is, thus, also becoming more and more enhanced. This results in an increased need for automation for the complex simulation setups and a need for quality assurance of simulation code used. Test automation is used to always ensure the quality of the code. The automation of various simulations and the test automation for the load simulation code are provided by PyWiT (Python Framework for Wind Turbines), which will be presented here in more detail.
Keywords:	Modelica, MoWiT, Python, Wind Turbines, Test Automation
Paper:	full paper
Bibtex:	@InProceedings{modelica.org:Fricke:2021, title = "{Python Framework for Wind Turbines Enabling Test Automation of MoWiT}", author = {Johannes Fricke and Marcus Wiens and Niklas Requate and Mareike Leimeister}, pages = {403--409}, doi = {10.3384/ecp21181403}, booktitle = {Proceedings of the 14th International Modelica Conference}, location = {Link\"oping, Sweden}, editor = {Martin Sj\"olund and Lena Buffoni and Adrian Pop and Lennart Ochel}, isbn = {978-91-7929-027-6}, issn = {1650-3740}, month = sep, series = {Link\"oping Electronic Conference Proceedings}, number = {181}, publisher = {Modelica Association and Link\"oping University Electronic Press}, year = {2021} }

Title:	A Graph-Based Meta-Data Model for DevOps in Simulation-Driven Development and Generation of DCP Configurations
Authors:	Stefan H. Reiterer and Clemens Schiffer
Abstract:	In order to improve the quality of model based development and to reduce testing effort DevOps practices gain more and more importance. However, most system engineers are not DevOps specialists and there are a lot of manual steps involved when writing build pipelines and configurations of simulations. For this purpose an abstract graph-based metadata model is proposed which allows to auto generate scenario descriptions for the DCP standard and code for the build server where the simulation is set up and executed. A simple use case is described as an example of how this could be applied in practice. Furthermore, a Python implementation of a DCP master and a simple FMI to DCP wrapper are presented in this as well.
Keywords:	Continuous Integration, DevOps, MBSE, NoSQL Graph Data Bases, DCP, SysML, UML, SSP
Paper:	full paper
Bibtex:	@InProceedings{modelica.org:Reiterer:2021, title = "{A Graph-Based Meta-Data Model for DevOps in Simulation-Driven Development and Generation of DCP Configurations}", author = {Stefan H. Reiterer and Clemens Schiffer}, pages = {411--417}, doi = {10.3384/ecp21181411}, booktitle = {Proceedings of the 14th International Modelica Conference}, location = {Link\"oping, Sweden}, editor = {Martin Sj\"olund and Lena Buffoni and Adrian Pop and Lennart Ochel}, isbn = {978-91-7929-027-6}, issn = {1650-3740}, month = sep, series = {Link\"oping Electronic Conference Proceedings}, number = {181}, publisher = {Modelica Association and Link\"oping University Electronic Press}, year = {2021} }

Title:	Portable runtime environments for Python-based FMUs: Adding Docker support to UniFMU
Authors:	Thomas Schranz, Christian Møldrup Legaard, Daniella Tola and Gerald Schweiger
Abstract:	Co-simulation is a means to combine and leverage the strengths of different modeling tools, environments and formalisms and has been applied successfully in various domains. The Functional Mock-Up Interface (FMI) is the most commonly used standard for co-simulation. In this paper we extend UniFMU, a tool that allows users to build Functional Mock-Up Units (FMUs) in virtually any programming language, to support execution within Docker. As a result the generated FMUs can be distributed in an environment containing all runtime dependencies. To describe the process of creating dockerized FMUs using UniFMU, we show how to model and co-simulate a robotic arm and a controller using two Python-based FMUs.
Keywords:	FMI, Co-Sim, Python, Tool-Coupling, Docker
Paper:	full paper
Bibtex:	@InProceedings{modelica.org:Schranz:2021, title = "{Portable runtime environments for Python-based FMUs: Adding Docker support to UniFMU}", author = {Thomas Schranz and Christian M{\o}ldrup Legaard and Daniella Tola and Gerald Schweiger}, pages = {419--424}, doi = {10.3384/ecp21181419}, booktitle = {Proceedings of the 14th International Modelica Conference}, location = {Link\"oping, Sweden}, editor = {Martin Sj\"olund and Lena Buffoni and Adrian Pop and Lennart Ochel}, isbn = {978-91-7929-027-6}, issn = {1650-3740}, month = sep, series = {Link\"oping Electronic Conference Proceedings}, number = {181}, publisher = {Modelica Association and Link\"oping University Electronic Press}, year = {2021} }