Generating Physically Sound Designs from Text and a Set of Physical Constraints
This addresses the problem of automated design generation for engineers and designers, though it appears incremental as it combines existing text-image models and physics simulators in a novel application.
The paper tackles the problem of generating physically sound designs from text descriptions and physical constraints by presenting TIDES, which jointly optimizes structural and visual properties using a text-image model and differentiable physics simulator. The result shows it can satisfy engineering requirements like compliance and density while incorporating text-specified features, as validated through 3-point bending tests on 3D-printed beam designs.
We present TIDES, a text informed design approach for generating physically sound designs based on a textual description and a set of physical constraints. TIDES jointly optimizes structural (topology) and visual properties. A pre-trained text-image model is used to measure the design's visual alignment with a text prompt and a differentiable physics simulator is used to measure its physical performance. We evaluate TIDES on a series of structural optimization problems operating under different load and support conditions, at different resolutions, and experimentally in the lab by performing the 3-point bending test on 2D beam designs that are extruded and 3D printed. We find that it can jointly optimize the two objectives and return designs that satisfy engineering design requirements (compliance and density) while utilizing features specified by text.