Generative Design - Wrong Safety Factor

[Anonymous]

Our team has experienced the following problem: the Generative Design outcomes come with a certain Safety Factor. Testing such design outcomes in the simulation environment, under the same loads and constraints, gives a different (always lower) Safety Factor. Why? What FoS should we trust? Do you really want us to spend thousands of dollars into testing a prototype that may have a lower FoS than expected? Is our team missing something obvious or the service you are providing is just flawed and unreliable? How do we solve this problem? Thanks.

[TheCADWhisperer]

What is the definition of Safety Factor?

[Anonymous]

Excuse me? For us there is a difference if a subsea riser clamp designed for the offshore industry can withstand a total weight of 36 thousand Kg or 24 thousand Kg before having a Safety Factor lower than 3. 

We had generative design outcomes with a Safety Factor between 3 and 6 and after testing them in the Simulation environment they would have a FoS < 2. Why is that? We have a legitimate question, can you give a proper answer?

If you are genuinely asking for what a Safety Factor is, then please consult the Wikipedia article:

https://en.wikipedia.org/wiki/Factor_of_safety

[bmfrank]

Without seeing the results I don't know the answer for this specific case, but there are some things to check.   Sometimes when the geometry conversion takes place the method that is used to join your preserve geometry with the geometry generated can result in features that may introduce stress concentrations.  These stress concentrations are not present in the synthesized geometry solution, but the method for creating editable geometry might introduce them.  These may yield a lower safety factor than what is computed during the geometry synthesis.  In this can you can likely edit the geometry (adding fillets or chamfers, modifying blends) to remove the likelihood of creating stress concentrations.  Alternatively, you could download a mesh/stl version of the outcome and use this as the basis for design.  Feel free to DM me or post additional information here if needed.  

[TrippyLighting]

---

@Anonymous wrote:

We have a legitimate question, can you give a proper answer?

---

Then provide relevant data in form of a model and workflow description. Screencasts are a very helpful tool.

What other software and methods are you using to verify results ?

[Anonymous]

Thank you. Please allow our team a couple of days to prepare the data. I appreciate the support.

[Anonymous]

Your reply was very useful, thank you. Our team will test the recommendations you have suggested.