Text-Guided Bio-Architectured Materials Library Building and Application to Structural Design

Advancements in deep learning techniques have not only enhanced intuitive data analysis in structural design but have also facilitated the transfer of deep-level information.The microstructural features of biomaterials, termed "structural genes," inspire engineering design with their uniqueness and complexity, containing a wealth of potential optimization resources.Structural genes are categorized into three main groups based on their functions: structure and mechanical properties, fluid dynamics and substance exchange, and energy management and interaction.The construction of a deep information database founded on these structural genes adds a new dimension to materials science research and revitalizes structural optimization methods.This study refines the text-to-image model to construct a biomaterials database, integrating it into the topology optimization design.This integration allows the design process to incorporate nature's optimization strategies, generating engineering structures that meet mechanical requirements and possess bio-inspired characteristics.Experimental validation presented in this paper showcases a novel paradigm for functional biomimetic design.