Please note that the ITEA Office will be closed from 25 December 2024 to 1 January 2025 inclusive.
Release of FMIFlux v0.13
- Project
- 22013 OpenSCALING
- Type
- Enhancement
- Description
FMIFlux allows the integration of FMUs into the SciML ecosystem of Julia. Consequently, the generation of PenODEs or Physics-informed neural networks with FMUs containing the physical equations becomes possible.
- Contact
- Lars Mikelsons, University of Augsburg
- lars.mikelsons@uni-a.de
- Research area(s)
- SciML
- Technical features
FMIFlux makes FMUs available in the Julia programming language as differentiable objects. The required derivatives are either computed directional derivatives from the FMI standard or sampled usinh finite differences. Moreover, this differentation is integrated with the Julia AD tools.
- Integration constraints
Same as for the Julia programming language, thus platform independent.
- Targeted customer(s)
See star gazers.
- Conditions for reuse
FMIFlux is released under the MIT licence.
- Confidentiality
- Public
- Publication date
- 11-09-2024
- Involved partners
- University of Augsburg (DEU)