3D Shape Synthesis for Conceptual Design and Optimization Using Variational Autoencoders
This addresses the need for efficient 3D shape synthesis in engineering design, though it is incremental as it builds on existing VAE methods with a new representation.
The paper tackles the problem of generating 3D shapes for conceptual design by proposing a data-driven method using variational autoencoders to learn from existing designs and produce new ones, demonstrating that it can generate candidate glider designs achieving prescribed performance goals even with limited data.
We propose a data-driven 3D shape design method that can learn a generative model from a corpus of existing designs, and use this model to produce a wide range of new designs. The approach learns an encoding of the samples in the training corpus using an unsupervised variational autoencoder-decoder architecture, without the need for an explicit parametric representation of the original designs. To facilitate the generation of smooth final surfaces, we develop a 3D shape representation based on a distance transformation of the original 3D data, rather than using the commonly utilized binary voxel representation. Once established, the generator maps the latent space representations to the high-dimensional distance transformation fields, which are then automatically surfaced to produce 3D representations amenable to physics simulations or other objective function evaluation modules. We demonstrate our approach for the computational design of gliders that are optimized to attain prescribed performance scores. Our results show that when combined with genetic optimization, the proposed approach can generate a rich set of candidate concept designs that achieve prescribed functional goals, even when the original dataset has only a few or no solutions that achieve these goals.