Bolted connections approach

Rino_aic

Hello Guys,
It has certainly been covered repeatedly before, however I have not found an effective tutorial nor is the guide exhaustive in fully describing the functionality of Mecway software. I'm definitely going to miss something.
This is a repetitive problem that requires a non-linear approach to the solution that allows the FEA solution to converge to those available in the literature. In fact, in many structural tests they ultimately focus on the analysis of flanged connections in which the response of Bolts is essential.
In the attached example I loaded an application with a simple connection of 2 bolted plates subject to a force that generates a bending moment. The bolts are necessarily subject to a tightening torque which is distributed over the many interfaces present. Therefore I ask those who have already faced this problem several times what is the best solution to manage:

interfaces:

1) Plate - plate
2) Plate - washer
3) Washer-Nut
4) Washer-bolt
5) Plate holes-Bolts
-----------------------------------------
6) How to manage the tightening torque
7) The scheme of constraints

Last but not least, I have seen that it is preferable to insert a thermal stress on the bolts instead pre-tension.

Thanks to all those who want to contribute or who want to report good tutorials and experiences.

Victor

I don't know an accurate way to model tightening torque. I would imagine applying a fixed rotation angle to the bolt and constraining the nut against rotation after it is tensioned but not sure how valid that is.

Other than torque, put contacts between all the contacting surfaces and if necessary, constrain the free parts with weak springs. It would be also be safer to use quasi-static and apply the loads in sequence in case there's some path dependence.

Not sure about thermal stress vs pre-tension section with length adjustment either. Pre-tension section is supposed to be easier. Both ways can transmit torque and bending.

Mishal

Hi Guys

What is the purpose of doing a FEA for such a connection? or is this purely academic?

I have only done FEA for connections which cannot be designed by hand as it is very time consuming. Personally I would just use a structural code such as AISC360 or equivalent.

Rino_aic

Although I am a perfect beginner in using Macway FEA, I can only confirm that the part relating to connections with moving elements (bolts, screws, etc. etc.) is normally neglected in most FEA / FEM softwares.
However, I wouldn't agree that it could just be an academic question. In fact, generally the result of an FEA analysis is transferred to the reports to be delivered to customers, who in turn, through us or directly to them, use the reports to request approvals from the technical government bodies in charge.
Therefore, in addition to not being formally correct or not elegant, it is absolutely unprofessional to find in the reports components subject to verification, such as the joints, which have stress or deformation values ​​that are not in percentage compatible with what can be calculated in the technical literature.
You all teach me that if I perform the single FEA analysis of a simple assembly consisting of a single bolt and a single nut, we constrain the nut and apply a torque to the head of the bolt, the result changes considerably or the the software goes in error, if we consider the data constrained on the side faces or on the bottom surface, if we consider a bonded contact between bolt thread and nut or a frictional contacts. The situation expands considerably if we consider intermediate elements such as washer and blocked elements such as flanges.
Are we by any chance stating that report delivered to our customers must be a mixed report between FEA and hand made calculation only for connecting elements as are screws or bolts? We can't believe or accept this.

MikeMcMullen

In practice here in the US, for civil structures, especially bridges, torque is not used directly. Calibrated crush washers, turn of the nut, or torque jobsite determined on a sampling of bolts and nuts with the same grip, washers, batch, using a hydraulic load sensor. Even then the torque is such as to generate a strain in the non-linear region. This fastidiousness about bolting even creeps into situations where preload is not desired (connections intended to slip in service), and plans normally need to make exceptions for them and make other provisions to keep vibration or thermal cycling from loosening the nuts.

Rino_aic

Thank you Victor and Mike.

JohnM

Some practical suggestions for modeling bolted joints:

1- Create a library of bolt models using hex elements (these make stretching easy).
2- Include bolt pretension and bolt/flange bonded contact, including "empty" flange face.
3- When bolt model is imported from library to your "main" model, the mesh, pretension
and bonded contact are brought in as well.
4- Rotate and stretch bolt to flange width and orientation
5- Create a node at center of bolt hole flange face using "click 3 nodes for center node"
python script (see Forum- Python API scripts).
6- Select nodes on bolt, drag by a center node to center node on flange, adjust axial
location if required.
7- Select flange surfaces and add to bonded contact face set.



I used this approach on your model with a dummy M12 bolt, and contact and friction between the flange faces. Bolts are bonded to flange faces. Bolt load is ramped 0 to 0.1, then the bending load is applied.



FYI, I think your beam elastic modulus had a units error, I fixed in model attached.

This looks like a lot of steps but it is actually very quick and easy!

disla

Very cool. Thanks for sharing

Hengre

Hi JohnM

thank you for your suggesions for modelling bolted joins. May I ask some simple questions to your points 1, 4 and 5

1- Create a library of bolt models using hex elements (these make stretching easy).
How did you generate the Hex mesh? Was it done in netgen or gmsh and then
importet to Mecway?
2- Include bolt pretension and bolt/flange bonded contact, including "empty" flange face.
3- When bolt model is imported from library to your "main" model, the mesh, pretension
and bonded contact are brought in as well.
4- Rotate and stretch bolt to flange width and orientation
For this I think you will use Mecway's Move/Copy/rotate.. functionallity?
5- Create a node at center of bolt hole flange face using "click 3 nodes for center node"
python script (see Forum- Python API scripts).
Is it right, that this node is only for positioning? This node is not connected to any other node?
6- Select nodes on bolt, drag by a center node to center node on flange, adjust axial
location if required.
7- Select flange surfaces and add to bonded contact face set.

JohnM

1- You can use many methods, but the simplest is Mecway:


Add the pretension and bonded contact, store in your Library, then grab when needed:


4- yes, use Move to displace sections of bolt to stretch to proper length. Keep your library bolt models located at the origin, that makes them easier to rotate/scale/stretch. Depending on your applications, you can pre-make in standard lengths, shapes, sizes and orientations.

5- yes, center node is simply for dragging bolt model to center:


Note: if you import and make multiple copies of the bolt model after import, you will need to delete and re-define the pretension for the bolt copies. (try, you'll see what I mean)

Rino_aic

Thank you JohnM,
I will deepen your valuable suggestions

Hengre

Thank you JohnM
for that nice presentation, I am happy to see that others find it also valuable. Thx