Meaning of forces extracted from "bolt" element 3

[rgsds]

Hi All,

Apologies if this has been answered before but I could not find it anywhere on the forum (or outside). Hope someone more experienced can help or send me in the right direction.

When modelling a bolted connection in FEA it is common to use a very stiff beam element, then connect to the parts it is supposed to clamp using a rigid connection.

This ensures that the stress read on the beam is the stress the bolt would "feel". I can then extract the forces and moments and use them to size my bolts and determine the correct preload using hand calcs.

My question is: Do I need to take into account the moments and the added tension these would add to the bolt, or is that value already included in the force vector? i.e. if I take my axial force component on the beam to be e.g. Fz, is the beam actually experiencing Fz+(axial force caused by moment) or is Fz already composed of a mixture of pure axial force and additional axial force caused by the eccentricity (moment) of the force applied?

Thank you to whomever bothers to read this! I am using FEMAP if that helps.

 

[NRP99]

IMO - Fz already composed of a mixture of pure axial force(pretension) and additional axial force caused by the eccentricity (moment) of the force applied.

No need to take into account the moments. Bolts are used for clamping and as shear connection. Please take into account shear forces for checking against bolt shear. The purpose of clamping will be defeated if direct bending of bolts is allowed.

 

[FEA way]

Prepare some simple test model of a bolt represented this way and short pieces of members joined by it. Load it in various ways (including preload) and extract forces to see how it works. That’s the best way to learn various modeling approaches in FEA.

There’s a recently published book titled "Practical Finite Element Analysis for Mechanical Engineers" by D. Madier. It features a whole chapter about the modeling of bolted connections so you may find it interesting.

 

[rgsds]

NRP99:

that's what i would think but havent been able to verify. the stress would have to show the resultant but i am not sure the force components would.

All my design cases are slip critical so both ends of the beam element are fully constrained (by an MPC) to their side of the joint. I am not worried about the bolts failing in shear, more worried about the reduction in preload which would cause the friction to decrease and cause slip. I am not sure if this is what you meant regarding checking against shear.

FEA way:

Ok thank you for the feedback. I am in the process of setting up a model where i can isolate these issues (i'm a bit of a noob in fea). I will check out this book.

 

[rb1957]

bolts carry shear and tension loads, and should (IMHO) be modelled to provide this (like using a CBUSH element).

bolts can be analyzed by hand, if the length of the bolt is creating significant bolt bending.

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

 

[NRP99]

All my design cases are slip critical so both ends of the beam element are fully constrained (by an MPC) to their side of the joint. I am not worried about the bolts failing in shear, more worried about the reduction in preload which would cause the friction to decrease and cause slip.

For slip-critical joint, large shear force is consumed in friction and hence bolt would see almost no shear. And slip would not occur. For these joints, bolt are preloaded to high degree. And hence very large shear force is required to cause bolts to shear and then slip. But then why you are worry about loss of preload and then slip? Unless there is very large force of which tensile component causing loss of preload to large extent which is of high degree in first place and then shear component accelerating slip.

Which software are you using? Can you post a picture of your model setup and force component/stress results at the beam? How are you applying loads? Its difficult to provide any direction by just guessing about your model/work.

If you are stuck with the FEA output quantities that are not matching the hand calculation or intuition, then perhaps you should check whether you are calling correct output/made a correct analysis setup/other sanity checks. Beam elements require special output . Start checking there.

 

[MegaStructures]

Relevant Literature:

See here for a practical explanation of bolt modeling in FEA:

https://appliedcax.com/support-and-...CAx-User-Guide-FEA-of-Bolted-Joints-Rev-0.pdf

See also the following study for a comparison of bolt modeling methods and a description of each method:

https://www.sciencedirect.com/science/article/pii/S0307904X0600062X

Now to your specific point:

What moment are you talking about?

Moment in Beam Elements: Your RBE's should not be transferring moment to the bolt shank (beam element), you should transfer only forces to the center node, so there should be no moment present in the bolt. Think about it, how could a moment be transferred from a plate to a bolt head? The head would have to be welded to the plate. I recall that the engineers at appliedCAX recommended adding some "accidental eccentricity" factor to the bolts of around 0.1*Ftension. I would have to dig the article back up to recall why and I can't seem to find it now.

Moment in Connection: If you are talking about the global moment in the connection (like in the case of a bolted end-plate connection), yes, that will be transferred to the bolt. Still, it will be transferred to bolts as a force, likely tension, which will decompress the steel plies and reduce preload on one side and do the opposite on the other. The FEA model will intrinsically distribute this force to the bolts; you don't need to perform further hand calculations.

Like others, I recommend creating a solid bolt model to validate your simplified modeling method. I'm not sure what type of connections you work with, but you might want to look into a program called IdeaStatica, which will help lessen your learning curve. They also have good technical resources you can use to sanity-check your modeling methods

https://www.ideastatica.com/blog/bolts-and-bolted-connections.

The component-based finite element method is robust, and that is really the type of model you're making, maybe without knowing it.


“The most successful people in life are the ones who ask questions. They’re always learning. They’re always growing. They’re always pushing.” Robert Kiyosaki