Allowable Stress using FEA

[13stone]

The short of it is: I need advice on how to determine the allowable stress value for running an FEA on Algor. Or rather, what is an acceptable factor of safety?

The long of it is: I am not an Algor user. I have a colleague who is, and he has designed a piece of equipment for my department's use. He performed a FEA with Algor and used my department's standard design factor of 3:1 based on yield. The designed piece is grossly oversized. We both understand that the 3:1 design factor was never intended for FEA, but we are coming up short on finding information or advice on determining an acceptable (and reasonable) factor of safety. Neither of our departments has any formal guidance on this.

With the understanding that FEA is supposed to make designs lighter and more economical, we are looking at this thing and scratching our heads.

The piece is simply a tension link made of plate steel; it will connect a large shackle (with sling) to a large clevis. The expected load is known. No impact loads expected.

Thanks for any advice you have to offer.

 

[GBor]

FEA is expected to make things lighter, but not to violate factors of safety. If the analysis is suggesting that your current design should be heavier than what physically has been working for some time, I would make two suggestions:

1) Analyze the current configuration and see if you think it actually has a safety factor of 3 according to the FEA.

2) If you don't have something already in existence to which to compare, continue to refine your mesh and make certain that the stresses that are showing up in the FEA and that are limiting you are real.

Some general guildlines that I use:

If you have greater than a 10% stress gradient in a single element, the mesh needs to be refined in that area.

If you enhance your mesh and the stress range changes by more than 10%, you need to continue to refine your mesh.

Garland E. Borowski, PE
Borowski Engineering & Analytical Services, Inc.
Lower Alabama SolidWorks Users Group

 

[doatis]

There is more to a safety factor than just the ratio of actual stress vs. allowable.

1 - Have you applied a load that represents a worst-case scenario, or is it a typical operating condition?
2 - Do your material properties represent average/below average/etc. values?
3 - Is your geometry modeled at minimum/worst-case thickness?
4 - Can you accept localized yielding - i.e. can a small fillet/chamfer/etc. yield without you considering it a failure?
5 - As stated previously, check to see if your FEA has actually converged on the answer

For example, when I used to be in aerospace we had specs (which it sounds like you don't have) that specified we had to have a M.S. >1, but that was for minimum thickness geometry, minimum material strength, and maximum predicted loading condition. We even considered that a little excessive, but we had previous experience to back up the methods.

It's not so much a study in FEA, but in probability.

Good luck

 

[GBor]

You might want to explain the difference between Margin of Safety (MS) and Factor of Safety...

 

[doatis]

I use the two interchangibly, since both tell you the same information. However, if you want to be specific:

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

I would also recommend consulting your Mechanics book, I highly recommend Shigley:

http://www.amazon.com/Mechanical-En...=pd_bbs_1/102-1518219-3980142?ie=UTF8&s=books

 

[GBor]

I wasn't trying to be rude, just that Factor of Safety is generally understood and consistent in application. Margin of Safety is sometimes interchangable and sometimes equal to factor of safety - 1, so a positive "margin of safety" is good.