This paper was produced for the 2019 NAFEMS World Congress in Quebec Canada
Because products are getting more and more complex, Systems Engineering requires sophisticated multi-disciplinary models to accurately simulate the system’s behavior in detail.
These models are specified from a top-down decomposition of the system and built from the bottom-up assembly of elementary models.
Several people from different domains of expertise collaborate on this development process. Hence, for an efficient communication within the project, there is a need for a common formalism to describe the aim, the expected behavior and the interfaces of each of the individual models that will be further incorporated into a simulation architecture representing the overall system. This is the purpose of the MIC concept (Model Identity Card): a MIC consists of predefined sets of attributes characterizing a model itself, as well as each of its input/output data (ports and variables) and parameters. Looking at the MIC of a model there shall be no ambiguity about what it simulates, how to use it, and the level of confidence on the validity of its results. Likewise, it shall be easy to determine whether two models can be connected just by checking the ports definitions in their respective MICs. It helps to identify the inconsistency problems at the early phases of projects (even before simulation models exist).
The COCTEL project, funded by the French ADEME organization, offers the opportunity to experiment the MIC concept on a real-scale prototype by developing a collaborative simulation platform for an automotive OEM (Original Equipment Manufacturer) where all the models are described by a MIC. One of the challenges of this project is bringing together a group of diverse partners from academic, software and industry to combine models of different origins, using a common platform and interface.
Several benefits are expected from this approach:
• An improved communication between the teams: they have access to a shared database of models where content management is consistent (all the models descriptions follow the same “standard” template based on a common vocabulary)
• Easy search for models in the platform: the search engine takes advantage of the MIC content to find the MICs matching the criteria and consequently the target models.
• The provided tools (MIC completion check, Ports compatibility check) help anticipating potential issues when attempting to assemble two models.
• As the platform is deployed on the cloud, it offers cloud computing capacities, by enabling shared access to the backend HPC resources to perform the simulations.
The specified scenario to be used for the validation of the platform is the optimization of an automotive electric powertrain. It involves specific models developed by different academic partners to take the electromagnetic, thermal and mechanical aspects of the electric engine into account. The project will also provide the opportunity to evaluate the benefits of the FMI standard to enable the co-simulation of these models.
In conclusion, several topics of future work have been identified, e.g.:
• How to import all the legacy models of the OEM with their MIC into the platform?
• How the exploitation of the MIC content can help automating the assembly of actual models?