Pushing Generative Design to the limits

[dban1357]

Hi Everyone,

I'm a 17 year old designer making a mountain bike frame using bonding adhesives, tubes, and generatively designed metal lugs intended for DMLS fabrication. I’m familiar with the general workflow in Fusions Generative Design, including setting up preserve/obstacle geometry, load cases, and manufacturing constraints. However, I’m seeking guidance on pushing the workflow further to ensure the results are functionally validated, manufacturable, and structurally sound under real-world dynamic conditions.

Here are some more questions I’d appreciate insight on:

1. Obstacle Geometry and Load Application:
   Is it possible to apply forces directly to obstacle geometry to simulate load paths from moving parts (e.g., crank arms, suspension links), or must all load-bearing surfaces be preserve geometry?

2. Constraint & Load Strategy for Dynamic Systems:
   For applications like a bike frame, where loads come from a combination of seated pedaling, standing climbs, lateral cornering, and landing forces, what are best practices for modeling complex, multi-directional load cases that reflect dynamic usage? Is there a recommended way to combine or layer load scenarios?

3. Intermediate Preserved Geometry:
   Does Autodesk support or recommend using intermediate preserved geometry (non-functional solids placed between key load paths) as a method of influencing structural flow or introducing design intent, especially in the absence of a visual goal like Voronoi?

4. Validation Workflow:
   After a generative study, what’s the best method to validate performance in Fusion? Are there guidelines on mesh fidelity, converting generative bodies to solids, or preparing them for high-resolution FEA within Fusion or external tools?

5. Controlling Design Space Efficiency:
   How can I better guide the growth of the design space to avoid overly organic or unmanufacturable shapes? I’ve experimented with obstacle and starting geometry, but I’d like to better bias the outcome toward mechanical efficiency and simplicity for DMLS.

I’m currently running multiple studies with different constraints and setups to evaluate various strategies. Any guidance on advanced generative workflows, post-processing recommendations, or load case design frameworks would be much appreciated.

Thank you for your time and support, I can PM you my f3d/f3z generative design set up file upon request.

[CGBenner]

@dylan.banera 

Welcome to the Community.  I moved your post here to get more of the expert users to see this.  Good luck!