Ansys is a global leader in numerical simulation and 3D design software with recognized expertise in multiphysics numerical modeling. This CIFRE thesis, in association with LIRIS, focuses on the analysis of numerical data carried by meshes.

This initiative covers simulations involving structural mechanics, fluid mechanics, electromagnetism, heat transfer, optics, etc., where the complexity and accuracy of the models require high-performance computing and generate very large volumes of data. These simulations are performed using Ansys solvers. Drawing on LIRIS's expertise in exploiting the geometry of data in relation to meshes, Ansys seeks to improve the storage and exploitation of high-precision simulation results, simplifying their remote graphical exploitation during and after computation, but also facilitating their long-term archiving.
Certain types of simulations and their results are particularly difficult to compress due to their complexity and the nature of the data involved.

These datasets can commonly contain hundreds of millions of values evolving over hundreds of time steps.
With this in mind, Ansys wants to develop compact representations that are lossless or have a predefined low level of loss and are easy to use. LIRIS brings its expertise in exploiting similarity in geometric data and wants to study its extension to simulation result fields. One of the challenges will be the possibility of on-the-fly compression, or at least with a controlled memory footprint.