These slides were presented at the NAFEMS World Congress 2025, held in Salzburg, Austria from May 19–22, 2025.

Abstract

In this presentation, we introduce a novel approach in MSC Nastran that integrates generative design principles directly within the topology optimization algorithm, streamlining the optimization process by producing designs that are much closer to the final production requirements and eliminating user-based decisions on where to add or remove material '“ decisions that require extensive expertise and knowledge of the manufacturing processes. A new, voxel based engineering approach has been developed and integrated via a submodelling approach. A defined design space inside a complex assembly is identified via one unique property ID and a submodel automatically is set up including boundary conditions and optimization parameters. The underlying optimization algorithm itself creates a new voxel mesh which is refined depending on optimization progress up to extremely high resolutions. Using an engineering optimization approach with a hard-kill BESO algorithm instead of the well-known mathematical density based approach enables a clear geometry definition in each iteration including a high quality tet-mesh recreation after submodel optimization finished. This near-net shaped design candidate is transferred back to the original assembly and automatically validated in full assembly context by MSC Nastran. Due to the engineering approach of the algorithm, design constraints are much clearer covered with e.g. exact representation of tapering angles for casting with 3° or more or less. With this approach, the optimization result is not just a blurry idea of main load paths, but an actual design candidate which can be used downstream in the development process without the need for a manual CAD recreation. This new method reduces computational time (CPU & GPU accelerated) and engineering cycles significantly, enabling users to solve larger and more complex models than was previously possible. This opens up the path to design iterations and rapid prototyping of both components and full assemblies for both linear and non-linear simulation analyses. We will demonstrate all developments using representative industry-relevant examples. Take Aways: A novel approach in MSC Nastran that integrates generative design principles directly within the topology optimization algorithm, streamlining the optimization process by producing designs that are much closer to the final production requirements How to accelerate design iterations and prototyping in topology optimization applications leading to significant productivity improvements Application of the novel generative design technology to solve larger and more complex industrial models for both linear and nonlinear simulations Shifting use of topology optimization from design idea generation to actual geometry creation tool as a copilot for optimal part design generation.