Shape Optimization with a Flattening-Based Morphing Method

Abstract

In shape optimization problems, generating variously shaped designs is an important task. In this study, a new design method called the flattening-based morphing method, which can create various designs efficiently based on baseline objects, is proposed. In the flattening-based method, anchor points are defined for each baseline object to set correspondence among the baseline objects, and each baseline object is mapped to 2D parametric space in a way that places all corresponding anchor points of the baseline objects at the same location. Then, remeshing is carried out to make the baseline objects’ mesh topologically identical in the parametric space. After these remeshed baseline objects are parameterized back to the physical space, the morphed object is created by computing the positions of its vertices as a weighted sum of the baseline meshes’ vertices. When the flattening-based morphing method is applied to find the optimal shape of a blended-wing body aircraft using an artificial neural network (ANN), the aerodynamic performance enhanced optimal model with an appropriate loading capacity is successfully achieved using three baseline models. The simulation results of the baseline models and optimization results are also provided in the current study.

Keywords

• flattening-based morphing method
• artificial neural network
• optimization
• blended-wing body
• lift-to-drag ratio