Method for "correcting" stress due to extra (non-modelled) thickness?

[JBlack68]

To all

I have an FEM with stress value in an area of interest. Stress is a bit too "high" and is deemed to be the result of the modelling technique used (RBE2,etc).
One does not want to remodel the area as it will most likely create issue. It is known that in that particular area an extra thickness exists due to steel washer.

The question is: Is there a way of "correcting" the predicted stress by taking into account that extra thickness? (using a method than most stress guy will "accept")

I thought of using the ratio (t1/t2)^2 (where t2 is the full "real" thickness)

What I am not sure is the effect of have a "bi-metallic" section (alu + steel)

Any thoughts/ suggestions?

Thanks
Regards

 

[gravityandinertia]

Could you draw a picture of some sort. If it's a washer, I personally probably wouldn't calculate it as extra capacity because it would only develop stress through friction, which may be limited. However, I may not be understanding your question correctly.

 

[JBlack68]

agree with the friction point but the washer is clamped between the bolt head and the "plate". So assuming its effective it will add some bending capability

 

[corus]

If you're using beam elements then you've calculated the nominal bending stress for which appropriate design stresses will apply. You can't realistically take into account a washer, nor for that matter the bolt head thickness, to reduce the calculated stress. In the same way you wouldn't consider the bolt hole as a stress raiser on the nominal bending stress. If anything these features would be considered only as a peak stress for fatigue damage.

 

[SWComposites]

The washer adds NO bending capability. No experienced stress person will accept that.

Stresses at any element connected to an RBE2 are probably wrong and should be ignored.

 

[JBlack68]

thanks for that SWComposites.
One option is to "ignore" the stress hot spot with the argument that it is due to the RBE2 and therefore unrealistic. select stress in next element, etc, etc. Never too keen on "wordy" argument.

Are we saying that the fact that there is an extra thickness (due to the washer) as not effect on the local "capability"?

 

[SWComposites]

The washer might affect the clampup pressure distribution. But you cannot add the washer thickness to the sheet thickness to reduce the sheet stress.

If you have modelled a bolted joint with RBEs at a single node or as a wagon wheel around a moodeled hole, then all stresses in all plate/shell elements attached to the rigid element are meaningless (a polite way of saying they are rubbish).

 

[rb1957]

a washer should not be included in effective area. It can have the practical use of distributing the load from the bolt. the FEM is modelling this diffuse load as a point, and so these "point load" effects are often (in my experience) written off as "model artifice". The model believes the load is applied at a point, but the real world knows it is distributed over the head of the screw (or the washer). There are several checks to see if the load is a real problem ... will it cause the head of the bolt to rip through the skin ? will it cause the skin to bend excessively ?

you need to be very careful with counting pad-ups as being fully effective. riveted on doublers are only slightly effective (like on a pressure shell a dblr the same thickness as the skin will only pick up 1/3rd of the load (not 1/2, unless it is very large). machined padups, integral with the skin are more effective but not 100% (unless very large).

another day in paradise, or is paradise one day closer ?

 

[aerostress82]

I understand your pain. You are working with automotive guys whereas you've always been used to working in aerospace stress where the main difference is that automotive guys are only relying on FEA stress rather than hand calculations (maybe this is your problem). The washer should be modeled as the same thickness as the plastic part if it is being modeled on a plastic part. If it is being modeled on a metal part, then it should be 3mm+thickness until a plate thickness of 3mm (if your part is metal and 2.5 mm thick, your washer elements will be 5.5mm). Your washer will have a perfect mesh where you can place 4 bar elements from the center node to the washer element nodes. Your 4 bars will look like a plus sign and your center node will be connected to your fastener element (most probably CBUSH fastener element). This will avoid the peak stresses that will occur due to the previous RBE2 modeling and will let you use any fatigue post processor (designlife, ncode, fesafe, nastran embedded fatigue & etc). Try to research this kind of modeling. I think this maybe what you were looking for. Also please let me know where you are using this modeling (aircraft, automotive?)

Spaceship!!
Aerospace Engineer, M.Sc. / Aircraft Stress Engineer

 

[JBlack68]

Thanks to all for suggestions, comments, opinions and advice. All greatly appreciated.
The structure being investigated is an aerospace one submitted to very high 'g' level
I dug a bit more into the results and the FEM is predicted very high shear value in the CBUSHs (much higher than the axial load) which must be the source of the high stress.

aerostress82: it looks like you have suggested an (company) modelling "rule" which will most likely impossible to find/confirm
the fixing is modelled with 1 CBUSH (K,1,2,3,5,6 used) and 1 RBE2 on each side (attached to each part!). On the part of interest the RBE2 extend to the head diameter.

 

[aerostress82]

In my above definition, change the word "washer" with "head". I defined it as clear as it can get - and if you follow what I wrote there one by one drawing on a piece of paper, you'll get to what you need.

Your problem is that your model is grossly rigid around the fastener area due to your RBE2 elements there. Which part of the A/C are you modeling? Interiors? Galleys? Primary/Secondary Structure Brackets? Primary/Secondary Structure itself? Wing, Fuselage, Landing Gear?

Every structure has assumptions of its own - so it really matters which part it is that you are modeling to find the most suitable FEA modeling method. Also if following hand calculations (for strength checks, buckling, crippling, fatigue joint & etc) are also existent, let me know that too - because it really changes a lot of things about your FEA. Also, how many years have you been working "professionally" on FEA? Your case is somewhat easy to resolve, but I guess you are new with FEA/stress.

Spaceship!!
Aerospace Engineer, M.Sc. / Aircraft Stress Engineer