Case Studies

Engine-Out NOx

Technologies

Test campaign DOE
Algorithm design
AI
Hybrid Modelling

Client’s Need

Electric Vehicles batteries can suddenly go into thermal runaway (TR) – bursting into flames or even exploding – with great risk for passengers, roads and infrastructures. The possibility of predicting when TR will occur allows to schedule battery replacement and avoid damages.
Aim: Avoiding sudden battery thermal runaway by providing suitable advanced warning.

Approach

Use AI to predict TR and provide advanced warnings.
Modelway designed, coordinated and carried out an extensive data collection campaign to build a proprietary dataset. The acquired data was then exploited to design and develop a custom AI algorithm. Our DVS® solution exploits hybrid modelling, merging data-driven and physical model-based approaches, to capture system complexities over extended operative conditions.

Main Outcomes

The prediction of thermal runaway start time and causes allows car makers to schedule maintenance, avoid damages to people and infrastructures and ensure car’s reliability.