Integration within Fluid Dynamic Solvers of an Advanced Geometric Parameterization Based on Mesh Morphing

Abstract

Numerical optimization procedures are one of the most powerful approaches with which to support design processes. Their implementation, nevertheless, involves several conceptual and practical complexities. One of the key points relates to the geometric parameterization technique to be adopted and its coupling with the numerical solver. This paper describes the setup of a procedure in which the shape parameterization, based on mesh morphing, is integrated into the analysis tool, accessing the grid nodes directly within the solver environment. Such a coupling offers several advantages in terms of robustness and computational time. Furthermore, the ability to morph the mesh “on the fly” during the computation, without heavy Input/Output operations, extends the solver’s capability to evaluate multidisciplinary phenomena. The procedure was preliminary tested on a simple typical shape optimization problem and then applied to a complex setup of an industrial case: the identification of the shape of a Volvo side-view mirror that minimizes the accumulation of water on the lens of a camera mounted beneath.

Keywords

• mesh morphing
• radial basis functions
• shape optimization