Welcome to the forums. Please post in English or French.

You are not logged in.

#1 2022-05-09 07:17:04

jacob
Member
From: Dolní Benešov
Registered: 2022-03-07
Posts: 55

The best way how to do bolt preload?

Hello,
I would like to ask if somebody is familiar with a solution I propose. I know that there are many methods how to model pretension.

For example:
You can prescribe temperature and cool the bolts down.

Or prescribe initial deformation in the middle elements, but then you have negative strain in the middle which is physically nonsense.

Similar it is with initial stress.

Then you can use some initial clearance, but every time you want to change the preload you have to remesh.

However, all these method require some trial and error to achieve desired pretension.


For example ansys or nastran (at least I think) models bolt pretension this way:
It cuts the bolt in the middle, then applies desired force on both sliced sides with constraints in other two directions than is the force direction, and then it solves the problem. In the next step, it "locks" the bolt (adds the equation in third direction) and solves with additional loads.
This method is also applicable for beams.

I did something similar in CA and I think it works.


Attachments:
Bolt_Pretension_1.rar, Size: 332.95 KiB, Downloads: 48

Offline

#2 2022-05-09 07:27:37

jacob
Member
From: Dolní Benešov
Registered: 2022-03-07
Posts: 55

Re: The best way how to do bolt preload?

I think this similar method is even better.
The nut is actually pushed through the bolt like it is done in reality, and then it is locked. No slice in the middle.


Attachments:
Bolt_with_Nut_Preload_1.rar, Size: 346.09 KiB, Downloads: 54

Offline

#3 2022-05-10 21:00:10

Maggoo
Member
From: Rio de Janeiro, Brazil
Registered: 2011-12-21
Posts: 73

Re: The best way how to do bolt preload?

Dear colleagues,

I made a proposal on this subject some time ago. I think it is more complex than yours, but I believe that it is more close to what is done in commercial packages.

Best regards,

https:  //www.linkedin.com/feed/update/urn:li:activity:6875554910864248832?utm_source=linkedin_share&utm_medium=member_desktop_web

Last edited by Maggoo (2022-05-11 17:30:25)


---------------------------------------------------
Helio Carlos Bortolon,
Mech Engr, M.Eng. and Maggoo
---------------------------------------------------

Offline

#4 2022-05-11 05:59:50

mf
Member
Registered: 2019-06-18
Posts: 325

Re: The best way how to do bolt preload?

Hello,

there are many ways to simulate bolts, that is correct. Let me just briefly describe what I do nowadays:
1) the bolt shank is modeled by a POU_D_E
2) the bolt pretension is applied at t=0 with a CREA_CHAMP and setting a value in the 'N'-field of the POU_D_E
3) the bolt head is modeled in 3D with CONTACT to the underhead area in part A
4) the nodes of the POU_D_E are glued to the bolt head and the thread area in part B
5) contact of bolt underhead may be done with/without friction.

A similiar case (without the 3D bolt head) can be found in this forum ('TEST51...') https:__//www.code-aster.org/forum2/viewtopic.php?id=25524 (remove the __ no links allowed).

What this model can do:
1) it shows you the relaxation of the bolt tension in the first timestep
2) the model can be moved freely in space, it can also be subjected to temperature variations. Thus, you will also see the increase/decrease of bolt tension with change of T.
3) evolution of bolt tension can be followed easily (just on value per bolt)
4) if bolt head is modeled in 3D, stresses in the bolt head are realistic.
5) bolts can do lateral movement to a certain degree, stresses won't rise to unrealistic values with e.g. thermal elongation.
6) stresses in the underhead area of part are realistic.

What it cannot do:
1) it does not give you realistic stresses in the connection to the bolt head (only one node!)
2) no distribution of stresses in the bolt shank (only one value of force/stress)
3) currently no modeling of threads (only cylindrical surface instead of thread), thus no realistic stresses there. But, threads may also be modeled in a realistic way, but I have not done that until now (for me, thread stresses are not really important, due to the size of the threaded surfaces, they are most certainly, always lower than in e.g. underhead areas).

Ciao,

Mario.

Last edited by mf (2022-05-11 06:06:43)

Offline

#5 2022-05-11 08:14:28

jacob
Member
From: Dolní Benešov
Registered: 2022-03-07
Posts: 55

Re: The best way how to do bolt preload?

Hello,

Maggoo, the link says: This post cannot be displayed.

I have seen many methods how to do preload but the main problem with many methods is, that some of the initial stress is lost, so you have to input some value higher which you initially do not know.

With "my method" you don't have to iterate, and it behaves quite nicely. You can even see how the bolt is tightened with time (0 to 1 s ).


Attachments:
ezgif-4-01835036f3.gif, Size: 880.14 KiB, Downloads: 71

Offline

#6 2022-05-11 10:07:50

mf
Member
Registered: 2019-06-18
Posts: 325

Re: The best way how to do bolt preload?

Hello jacob,

you are absolutely correct with the iterating. It is also necessary with the method I described above (the relaxed bolt stress is then equal to the stress you apply in real life). That is just the way it is. It has an advantage though: if your difference between your applied and relaxed bolt stresses is VERY high, then you'd better take a look at your model, contacting surfaces etc. So it actually results in useful information,

Mario.

Offline

#7 2022-05-11 12:34:02

jacob
Member
From: Dolní Benešov
Registered: 2022-03-07
Posts: 55

Re: The best way how to do bolt preload?

Well, I don't think that there must be a problem if the difference between your applied and relaxed bolt stress is very high (even 50%). It just means that the flanges are not that stiff.

Last edited by jacob (2022-05-11 12:34:30)

Offline

#8 2022-05-11 17:33:07

Maggoo
Member
From: Rio de Janeiro, Brazil
Registered: 2011-12-21
Posts: 73

Re: The best way how to do bolt preload?

jacob wrote:

Maggoo, the link says: This post cannot be displayed.

Sorry for that. The protocols may be a drag.

Here is the file.


Attachments:
boltplate.zip, Size: 301.37 KiB, Downloads: 54

---------------------------------------------------
Helio Carlos Bortolon,
Mech Engr, M.Eng. and Maggoo
---------------------------------------------------

Offline