Discuss your needsLet’s talk
EN
DTSC / BRAIN / Applied research
BRAIN / Applied research

Energy - PINN Transformers

Combine physical laws and learning to anticipate the behaviour of transformers.

BRAINEnergy - PINN TransformersEnergy · Physical modellingBRAINNo. 07 07Energy - PINN TransformersBRAIN
Past researchPINNsHybrid modelsPredictive maintenance
Past researchEnergy - PINN Transformers project artwork

The approach

The project models the thermal behaviour of transformers in the Belgian electricity transmission network using physics-informed neural networks (PINNs).

Thermal equations are combined with historical measurements, load and ambient conditions to estimate internal temperatures and identify overheating risks.

This hybrid approach links physical modelling and AI prediction, in service of predictive maintenance.

The contributions

Source: BRAIN portfolio.

Discuss this project