Real-time generative design of diverse optimized structures with controllable structural complexities and high quality

Abstract

With the boom of artificial intelligence (AI), generative design attracts engineers in various disciplines. This work focuses on achieving the real-time generative design of optimized structures with various diversity and controllable structural complexity. To this end, a modified Moving Morphable Component (MMC) method and novel strategies are adopted to generate high-quality datasets. The complexity level of optimized structures is categorized by the topological invariant. By improving the loss function, a WGAN is trained to produce optimized designs with the input of loading position and complexity level in real-time. It is found that high-quality, diverse designs with a clear load transmission path and crisp boundary, even without requiring further optimization, can be generated by the proposed model. This method holds great potential for future applications of machine learning-enhanced intelligent design.