RP3 · AI design tools

AI-assisted Power Electronics Design

Establishing an AI-assisted, physics-informed design toolchain that unifies device modelling, circuit optimisation and magnetics design — dramatically shortening converter design cycles.

  • Physics-informed modelling
  • AI optimisation
  • Digital twin
  • Design automation
Challenge

Conventional power-electronics design relies on experience and repeated iteration — slow and costly. The high-frequency, high-density nature of wide-bandgap devices makes the design space even more complex, demanding intelligent tooling.

Research Focus

01

Physics-informed modelling

Embedding device physics and electromagnetic models into AI models for accurate yet fast behavioural modelling.

02

AI-driven optimisation

Machine-learning exploration of topology and parameter space toward the best efficiency–density–cost trade-off.

03

Converter digital twin

Digital twins mirroring physical converters for loss estimation, ageing tracking and on-line calibration.

04

Design-flow automation

An automated toolchain from specification to prototype, compressing design cycles from months to weeks.

Expected Outcomes

  • AI design-toolchain prototype
  • WBG converter design knowledge base
  • Validated design-cycle reduction cases