I am modeling a bolted connection for structural loads. I have had some sucess when I model a bolted connection in inventor and import into FEMPRO. In some instances, e.g. sloted holes, I have not had sucess. So, I am using Beam elements to apply bolt loads. My question is... Can beam elements be designated as compression only links? This would allow me to tie together the perimeter of the bolt hole with beam elements without transferring tensile stresses. When designing a bearing only connection, there will be no tensile loads. Is there a technique?
Eng-Tips is the largest engineering community on the Internet
Intelligent Work Forums for Engineering Professionals
-
Congratulations waross on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!
Bolted Connections
- Thread starter jrepese
- Start date
What you're actually talking about isn't a "beam" element, but a "gap" element. Gap elements can be compression only and they can be with or without a gap. I'm assuming you are referring to a linear static analysis, so a gap element would be appropriate.
As for the Inventor export on slotted holes, you may want to check your facet resolution (in older AutoDesk products, the command was "facetres"). Circular holes are easily defined in the translation from CAD model to FEA pre-processor. Arcs (such as the end of slotted holes) are sometimes more difficult. Increasing the facet resolution (which, I always have to test this on a small model to figure it out, I think means decreasing the numerical value of the variable "facetres") sometimes clears up the translation issues.
Garland
Garland E. Borowski, PE
As for the Inventor export on slotted holes, you may want to check your facet resolution (in older AutoDesk products, the command was "facetres"). Circular holes are easily defined in the translation from CAD model to FEA pre-processor. Arcs (such as the end of slotted holes) are sometimes more difficult. Increasing the facet resolution (which, I always have to test this on a small model to figure it out, I think means decreasing the numerical value of the variable "facetres") sometimes clears up the translation issues.
Garland
Garland E. Borowski, PE
- Thread starter
- #3
- Thread starter
- #4
If I understand your situation correctly, I would recommend something very high (10,000 or so), but let me make sure I understand what you have done:
You imported a model from Inventor that contained slotted and round bolt holes with bolts through them. In the case of the slotted holes, some of the translation didn't work too well, so you've inserted beam elements between the bolts and the hole edge. In order to determine the load transmission into the side of the bolt hole (the bearing force), you are now using gap elements between the bolt and the hole wall.
If this is correct, than my recommendation stands. In reality, that contact has a virtually infinite stiffness and the materials will flex under the contact.
One concern, however, for the slotted holes is that you may need gaps before the bolt would actually contact the hole wall. If, for instance, the bolt is at one end of the slot and must first slide to the other end of the slot, then the gap element would need a "gap" of just shorter than the element length.
And, please, feel free to call me Garland...we will likely never meet, so the informality couldn't possibly offend me.
Garland E. Borowski, PE
You imported a model from Inventor that contained slotted and round bolt holes with bolts through them. In the case of the slotted holes, some of the translation didn't work too well, so you've inserted beam elements between the bolts and the hole edge. In order to determine the load transmission into the side of the bolt hole (the bearing force), you are now using gap elements between the bolt and the hole wall.
If this is correct, than my recommendation stands. In reality, that contact has a virtually infinite stiffness and the materials will flex under the contact.
One concern, however, for the slotted holes is that you may need gaps before the bolt would actually contact the hole wall. If, for instance, the bolt is at one end of the slot and must first slide to the other end of the slot, then the gap element would need a "gap" of just shorter than the element length.
And, please, feel free to call me Garland...we will likely never meet, so the informality couldn't possibly offend me.
Garland E. Borowski, PE
CoopEngineerVan
Mechanical
jrepese -
I'm probably using an older version of Algor, but I've recently also had a lot of trouble modelling bolts and bolted connections for linear stress applications.
If you just have bolt holes that you need to hold down your model, you can create joint elements by selecting the two inner surfaces of the bolt hole in and then selecting "create joint". I then apply the appropriate material to these joints and make one point fixed in all directions and the other point elastic (generally in the z direction) with a k factor that I calculate for the dimensions of the bolt, based on its Young's modulus.
Sometimes, however, this doesn't always work due to the nature of these pesky joints (a web of truss elements). My colleagues have theorized that perhaps the thickness is applying to each and every truss except the joint as a whole.
If you can model a portion of whatever it is being bolted ONTO, I've found through trial and error that modelling the bolts and caps themslves in Inventor gives much more accurate results. Having it pre-stressed (compression forces) could be tricky, though, perhaps you could apply the equivalent downward force to the top surface of the "nuts" to pin it down?
I'm probably using an older version of Algor, but I've recently also had a lot of trouble modelling bolts and bolted connections for linear stress applications.
If you just have bolt holes that you need to hold down your model, you can create joint elements by selecting the two inner surfaces of the bolt hole in and then selecting "create joint". I then apply the appropriate material to these joints and make one point fixed in all directions and the other point elastic (generally in the z direction) with a k factor that I calculate for the dimensions of the bolt, based on its Young's modulus.
Sometimes, however, this doesn't always work due to the nature of these pesky joints (a web of truss elements). My colleagues have theorized that perhaps the thickness is applying to each and every truss except the joint as a whole.
If you can model a portion of whatever it is being bolted ONTO, I've found through trial and error that modelling the bolts and caps themslves in Inventor gives much more accurate results. Having it pre-stressed (compression forces) could be tricky, though, perhaps you could apply the equivalent downward force to the top surface of the "nuts" to pin it down?
You may also try to pre-stress with a thermal variation from the material baseline. If you know the approximate stress that you need, correlate it to strain with Hooke's law, back out the temperature need to cause that strain via the thermal expansion coefficient.
Garland E. Borowski, PE
Garland E. Borowski, PE
- Status