Parallel Simulation and Robust Optimization of Electro-Mechanical Energy Converters
Accelerate transient simulations for electric machine design using time-parallel methods to enable efficient steady-state analysis and robust optimization of electromobility systems
Image copyright: TU Darmstadt
Involved researchers
- Iryna Kulchytska-Ruchka (TU Darmstadt)
- Sebastian Schöps (TU Darmstadt)
- Stefan Ulbrich (TU Darmstadt)
- Michael Hinze (University of Koblenz-Landau)
- Stephanie Friedhoff (University of Wuppertal)
Partners from industry / Research lab
- Oliver Rain (Robert Bosch GmbH)
Problems description
Designing electric machines involves computationally intensive simulations, especially when steady-state behavior must be analyzed through transient simulations of electromagnetism. Traditional time-domain methods become prohibitively slow during early design stages due to long settling times. Efficient and scalable numerical techniques are essential to accelerate simulations and optimizations for electromobility.
Goals
Accelerate transient simulations for electric machine design using time-parallel methods to enable efficient steady-state analysis and robust optimization of electromobility systems
Methods
Time domain finite element method; Parallel-in-time methods, e.g. Parareal; Shooting and harmonic balance; Robust optimization
Outcome
Achieved up to 28 times speed-up in steady-state simulations using 80 processors, enabling faster and more robust designs of induction motors for electric vehicles.
Reference
Iryna Kulchytska-Ruchka, Sebastian Schöps, Michael Hinze, Stephanie Friedhoff, Stefan Ulbrich, and Oliver Rain. PASIROM: parallel simulation and robust optimization of electro-mechanical energy converters. In German Success Stories in Industrial Mathematics, volume 35 of Mathematics in Industry. Springer, 2021. doi:10.1007/978-3-030-81455-7. Eingereicht von: Sebastian Schöps [04.05.25]
