Since many operating conditions need to be ran to get a full understanding of the drag and lift of a vehicle, companies in the automotive and aerospace sector generate hundreds of thousands of hours of CFD calculations.
While most engineers usually only analyse a few operating condition, Shape will make it possible to predict hundreds of results in a few seconds, spanning a range of flow velocities or angles.
This can be done predicted for a new design with simulations on only a few operating conditions or no new simulation at all.
With Shape, you can predict not only integrated scalar quantities such as lift and drag, but also pressures and velocities on surfaces, slices, or directly in 3D.
The Geometric Neural Networks used in Shape, are particularly good at reproducing strong non-linearities such as those induced by chock-waves or turbulent flows.
With Shape, you will be able to gather data from different design exploration campains and naturally explore and compare concepts.
Your ideas and your history are not bounded by parameters
Shape makes it possible to exploit your wind tunnel data in the best possible way.
Using transfer learning techniques, you will be able to complement your CFD with physical measurement.