AI Surrogate Models of Power Converters with Altair romAI | Webinar Recording
Altair Employee
This session shows how to build surrogate AI models of PSIM power converters using Altair romAI. These AI reduced-order models (ROMs) can be used to run system simulations much faster, while still capturing system dynamics with high accuracy.
Power converters in battery charging applications are often computationally expensive to simulate, and the large timeframes required for analyzing battery charging systems make this a very time-consuming process. Altair romAI can produce faster than real-time results so you can focus on evaluating and improving your designs.
This is a hands-on demonstration of the whole process, from defining the high-fidelity model, creating the required datasets, training the AI model and validating its performance accuracy and speed.
All relevant files can be found here:
- Slides:
- Open loop resonant LLC ROM:
- Model files:
- HyperStudy archive (run again to generate all results files):
- OML script for datasets compilation:
- romAI ROM:
- Twin Activate validation files:
- Parallel resonant LLC battery charging:
- Model files:
- HyperStudy archive (run again to generate all results files):
- OML scripts for datasets compilation:
- romAI ROMs:
- Twin Activate validation files:
- Input/output currents & thermal losses ROMs: