Static simulation Fusion 360 calc vs hand calculation and hot spots

[Anonymous]

Hello.. The enclosed shows a simple ear with a horizontal load applied to the hole as a bearing load. The fusion 360 result show a max stress of about 20KSI in normal YY which about agrees with the hand calc. BUT this appears as a hot spot. As the stress does not imediatlly linerally decrease only the X axis. As you would see in the hand calcs as the simple moment stresses in Nomall YY decrease linearlly as you get close to the centroidial axis of he cross section. 

Even with a  convergence rate of 2.8% on the results the 20KSI appears as a hot spot.  OI have to move doen to roughly about 12ksi or so before the results start to change in  a linear fasion as would be expected.  

It seems difficult or even not possible to eliminate ALL hot spots and trust the max stress as a real max stress. This is typical of alot of these FEA's even after applying liberal fillets,which does decrease the hot spots. Using custom settings on the mesh sometimes helps but often can make the hots spots worse. 

?????? I do not see alot of comments about hot spots and verifying results of the FEA? Any references woould be appreciated.

[John_Holtz]

Hi @Anonymous

I agree that there is not a lot of discussion about "hot spots" in the stresses. I will see if anyone has any good references. Some ways to avoid them are as follows:

1. Some hot spots occur because the constraints "pinch" the model. Instead of using one constraint that fully fixes a face, a combination of constraints may reduce the hot spot. (More about this below.)
2. The mesh quality has a lot to do with hot spots. But without some ability to generate a hand-built mesh, or provide some type of hand-control to the mesh, hot spots are going to be unavoidable.
3. Recognize that they will occur, in which case the "solution" is to do a hand calculation to extrapolate the result from nearby nodes to the point of the "hot spot".

In your particular model, are your hand calculations just a simple "M*c/I", or is there a more sophisticated calculation specifically for ears and short cantilevers? The M*c/I calculation that gives a linear stress distribution through the neutral axis would produce a stress in the ZZ direction (not the YY direction as you indicated). See the attached image. The maximum stress ZZ is 54 ksi at the hot spot, which is considerably larger than the expected 20 ksi.

[Anonymous]

Hello and TKs for the feedback.  Yes it is ZZ in this case as i had another model with the ears welded to a plate where it was YY and mixed up the axis in my message.  Yes it was the simple MC/I calc. There is a ref that I bought that I think was mentioned in another post here.  It does talk about hot spots and what "custom mesh" settings can minimize them.  If you know of another i would like to see it.  there is not alot written about this. using hand calcs to extrapolate to the hot spot areas is what i usually do. The 2D FEA in Mechanical can also be helpfull in confirming results. But given that "real" stress concentrations can go as high as 3-5 times there is always some risk in dismissing (for lack of a better word) hot spots. Moving the cursor on the legend can also help define hot spots. IF you can move the cursor for a good dist without much increase in FEA higeher stress colors that area is probally or even most probally a hot spot.

I appreciate someone saying that "Hot spots" can rarely be fully eliminated.  Instead of just saying do this or that and the hot spots go away, which very seldom happens. Even after doing this or that. Fillets help alot.

Tim S

[John_Holtz · Accepted solution]

Hi Tim,

Here are a couple of sources:

1. Building Better Products with Finite Element Analysis. A good book about FEA - well written, good illustrations (even though they are all black and white). I have attached the few pages that discuss singularities.
2. A blog post that discusses singularities, and includes a video. Perhaps you saw this one already since the you and video discuss using fillets. QUICK TIP: Singularities and Simulation
3. Here is another blog post that discusses it. Singularities in Finite Element Models: Dealing with Red Spots

Hopefully, other people will add additional resources. 

[Anonymous]

Thanks will check this out

Tim S