Section: New Results

Symbolic-Numeric Analysis

In [35] we address the reconstruction of convex polytopes, in any dimension $n$, from the knowledge of a finite set of directional moments of the shape. Starting with the formula relating the projection of the vertices to the directional moments, we employ established numerical algorithms for generalized eigenvalues and interval interpolation to recover the coordinates of the vertices. We perform the reconstruction of a diamond cut using our novel method.

This is joint work with Annie Cuyt, Wen-Shin Lee and Oliver Salazar Celis from University of Antwerp (Belgium).

The aim of the work [30] is to prove the usefulness of a bulbous bow for a fishing vessel, in terms of drag reduction, using an automated shape optimization procedure including hydrodynamic simulations. A bulbous bow is an appendage that is known to reduce the drag, thanks to its influence on the bow wave system. However, the definition of the geometrical parameters of the bulb, such as its length and thickness, is not intuitive, as both parameters are coupled with regards to their influence on the final drag. Therefore, we propose to use an automated shape optimization procedure, based on a high-fidelity flow solver, a surrogate model-based optimizer and a CAD-based geometrical model, to derive the characteristics of the bow geometry allowing to maximize the achievable drag reduction. The numerical tools are first presented, and then applied to the optimization of a bow shape for a real fishing vessel, in order to determine the optimal length and thickness of the bow for drag reduction purpose.

This is a joint work with Louis Blanchard and Régis Duvigneau (Inria, EPI OPALE), Elisa Berrini (MyCFD), Yann Roux (K-Epsilon) Eric Jean (Jean & Frasca Design).