I’m testing other user approach of bolting. I have certainly pushed hard the test as it has shown to be very stable in terms of convergence.
What I have notice is that as soon as I activate the pretension section (Very small value of 1N) the stud slips at the pretension section. The section forces shows that there is shear resistance, but the slip is very evident. Pictures are scale 1:1.
¿Is there any parameter that could be tweaked to controls that slip between both pretension sections of the stud ?. I guess it should be some kind of tangential stiffness combined with friction.
So it’s a single step with preload and actual load at the same time ? In such a case, it should be done in two steps - one with preload and the second one with pre-tension frozen (*BOUNDARY, FIXED) and actual load applied.
I did it with an amplitude card. First, I apply the pretension and only when it’s complete I start loading. It’s an imposed displacement to speed the process.
At least that’s the case in Abaqus but it seems to be the same in CalculiX. Give it a try, the 2 step approach should work and normal deformation of the bolt should be possible if you use *BOUNDARY, FIXED in the second step.
. NO, NO,My preload is applied in Step1. I thought it was the same if it is managed with the right amplitude. In fact the bolt remains like tied as both surfaces keep in the same plane but they slip. That is not a BC. Not mine at least.
From SOF:
normal force (+ = tension): 1.276357E+05
shear force (size) : 8.285414E+04
I will try with two steps and I hope it is the same.
Yeah, the meme that I shared (found on LinkedIn) is about another common mistake done by Abaqus users but here the effect could be similar with both loads in a single step.
If the *CLOAD or *BOUNDARY are acting in the wrong sequence or at the wrong time or interfering with *PRE-TENSION SECTION so that it requires 2 steps to separate them, I think that would be a bug. They’re supposed to follow their assigned *AMPLITUDEs and it looks like they do.
The problem is the *PRE-TENSION section’s MPCs are linear so they don’t rotate with the large rotation of the bolt. It’s constraining each face of the bolt to the same X,Y displacement, and leaving them free in Z, which is only correct in the initial configuration.
I made a single stud large enough to measure well shear and tension at different sections for verification purposes but I push it more and more just for pleasure. Then that separation clearly showed up.
For single shear loading condition, the center section of the stud rotates as the load increases.
Not sure what you mean. The MPCs from *PRE-TENSION SECTION transfer shear between the two halves of the bolt. As long as the rotation is small, it would work.
At least in Abaqus, the pre-tension and operational load should be separated into 2 steps:
You can maintain the initial adjustment of the pre-tension section by using a boundary condition fixing the degrees of freedom at their current values at the start of the step once an initial pre-tension is applied in the fastener. This technique enables the load across the pre-tension section to change according to the externally applied loads to maintain equilibrium. If the initial adjustment of a section is not maintained, the force in the fastener remains constant.
I would always advise doing the same in CalculiX but maybe they can somehow work together too in this solver, that would require testing. So far, the results shown by Disla don’t indicate this.
i’m not in detail, but as i understand pretension section of bolt can not be mixed with shear/bending actions. Resistance force are due to clamping face and friction. Probably, use complete bolt model with continuous mesh, common contact surface and initial condition type stress for small pretension (tightening) can solve the problem.
2 steps is for when you want both a specified initial force and the displacement/adjustment maintained afterwards to allow the force to subsequently change with external loads. You’re adding new constraints which requires a separate step. But in this example, it’s just a constant force, so that isn’t necessary. I’m sure Abaqus can do this simple case in one step too.
Yes, that’s right. This is how pretension is normally used.
Isn’t there both a pre-tension force and a separate prescribed displacement load (causing shearing of the joint) supposed to follow the established pretension here ? That’s how I understand this example but I haven’t seen the files so I can’t say for sure.
Result with two steps and fixed pretension displacements on the second step is the same.
I can share the prepomax file with you if you are interested. Maybe I made somethig wrong?¿?
The surfaces are connected rigidly in the tangential directions, so it doesn’t need a stiffness. Comments in the source indicate two MPCs “perpendicular to the normal direction” and the solution shows that X and Y displacement across the gap are enforced to be equal all time steps. That’s the tangential constraint that transmits shear force.
But that only works for small rotations because it doesn’t update the MPCs as the normal changes direction, so this large rotation example goes completely wrong.
If pretension is set up as force, the tensile force between the two sections will remain always below the set point no matter how the system evolves.
If the model BC (for example-imposed displacement as mine) induce an upper tensile force in the bolt, the pretension sections will open not to exceed the tensile force stablished in the pretension properties.
Pictures show the difference in reaction forces at the base of the bolt depending on if the setup is as pretension force or pretension by initial displacement (shrinkage of the bolt length).
I guess the correct way to impose a pretension force on a bolt in such a way that do not open and capture the further increase in internal forces in the bolt as the system evolve, is to impose an equivalent pretension “displacement” to obtain the desired initial tensile force. Then, the bolt doesn’t open , becomes stiffer, and captures the real internal stresses as the system evolves.
This is the standard way in Abaqus, as mentioned by @Calc_em 's cat. We may need to evaluate if a flag for “follower force” is required in the cases of large displacement.
1-Initial Pretension by means of force
2-Pretension redefined to displacement + freeze
3-Operational load.
4-New pretension by means of displacement but at a diferent section of the bolt.
. NO, NO,My preload is applied in Step1. I thought it was the same if it is managed with the right amplitude. In fact the bolt remains like tied as both surfaces keep in the same plane but they slip. That is not a BC. Not mine at least.
