Altair physicsAI - AI for Design Evaluation

In this session, we give an overview of how Altair physicsAI has delivered powerful results in a variety of domains including structural, crash and CFD applications. physicsAI can train using past simulation data, which usually get discarded, and even learn from existing physicsAI models, reducing the solver time spent on data generation. Once trained, physicsAI models can be easily shared as a single file which can then be deployed to make predictions on new CAD or mesh models with great accuracy and about 10-100x quicker than traditional FEM solvers. Working in tandem with our in-house tools such as Hyper Study, RADIOSS as well as other commercial solvers, a seamless solution for an optimized design can be developed and delivered.

Presented as part of Altair ATCx AI for Engineers 2024 conference.

The presentation is in English by default.

Duration 19 minutes.