FEA on Bolted Joint to determine the reaction loads

[kaffy]

Good Morning Fellow Engineers,

I am trying to determine the generated reaction loads at bolted joints when a horizontal force acts on center of gravity of machine (Seismic Loads).

My original Intent was to use the following approach
-Create a coordinate system at COG of machine
-Create a split surface (Bottom surface of top flange as shown in screenshot #2). Split surface is created to mimic the flat washer
-Apply a remote load at coordinate system and Connect the remote load to the faces shown in screenshot #2
-Run FEA and check reaction loads at the faces

*Since I am not interested in checking the base part, I will use a lot of stiffeners etc. to stiffen it up so the stress in part does not go beyond the yield stress
*I am planning to use remote load (direct transfer)

I have the following questions.
-Is there any alternate / Better approach to solve those kind of problems?
-Is using the remote Load (direct transfer) correct in this application or should I use Remote load (rigid connection)? If Rigid Connection, Why? Screenshot #3 has the explanation of both cases
-Any other free online tool available to analyze those kind of problems?

Thank you very much

 

[FEA way]

Since you are using SolidWorks Simulation, there are more suitable approaches that you can use for seismic loads. Response spectrum analysis would be the way to go in most cases but you would need input response spectrum data for base excitation. Another option is random vibration analysis. You could also use harmonic or modal time history analysis but you should use base excitation anyway.

 

[kaffy]

Thank you very much but Unfortunately I do not have access to either response spectrum analysis or random vibration analysis. I only have simulation professional and it doesn't allow for linear dynamic simulations. Although I used the word seismic but sometimes I have situations where machine is attached at angle or side and attached to base via bolts (similar arrangement). So I was thinking if I could actually use the mentioned approach. So from your perspective, Should I use the mentioned approach?

 

[FEA way]

It would be best to use linear dynamics for that. Since you’ve mentioned using another software that is available for free, I would suggest CalculiX. It’s a very good open-source solver based on Abaqus. It doesn’t support all the linear dynamics procedures (response spectrum and random vibration analyses are not available) but allows you to run harmonic or modal time history simulations with base motion excitation.

 

[rb1957]

Does the OP want to apply a direct force ? FEA seems to be talking about the more practical side of things, that the machine is probably vibrating or reacting dynamic (seismic) loads. But the OP replies "IDK" ?

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

 

[kaffy]

@FEAWay: Thank you very much for the valuable advices. I will look into the software mentioned
#Rb1957: Based on mentioned link, seismic force acts on the center of gravity of the equipment and cause overturning that generates moment forces at the restraint points. I have calculated the magnitude of seismic force and wanted to apply a Direct force (magnitude equal to seismic force) acting on center of gravity of machine.

https://acrefine.com/services/seism...nitude of the,forces at the restraint points.

 

[rb1957]

so you want to do a quasi-static analysis of a dynamic (siesmic) load by applying a static force at the CG ? Not saying this is wrong, just trying to understand what you're doing.

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

 

[SWComposites]

And if doing a quasi-static analysis and only want the attach bolt loads, why not just do a conservative hand analysis?

 

[kaffy]

@rb1957: Yes, that is precisely I am doing. My understanding is that results should be more conservative. Please correct me If I am wrong.
@SWComposites: I have already done a conservative hand analysis. I was looking for something to compare my numbers with. I am using the attached spreadsheet

 

[rb1957]

ok, you're applying a static load. What are your questions ? Your attachment didn't explain much in the way of questions.

How did you determine the static load ? Just guessed a number ? Calculated something for the seismic loading ?? Why do you think your load is conservative ?

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

 

[kaffy]

Total Load (vertical) = weight of machine + suspended load = F1
Seismic Load (horizontal) = Coefficient of seismic force * (Weight of machine + suspended load) (coefficient of seismic force is calculated based on site conditions and building codes) = F2

I understand that I should be using linear dynamics. I was trying to solve it by using the following approach
-Create a coordinate system at COG of machine
-Create a split surface (Bottom surface of top flange as shown in screenshot #2). Split surface is created to mimic the flat washer
-Apply a remote load at coordinate system and Connect the remote load to the faces shown in screenshot #2
-Run FEA and check reaction loads at the faces

My questions are:

Is the mentioned approach more conservative than linear dynamic analysis? or is it impossible to say and I should not be using the mentioned approach at all?
If using the mentioned approach make any sense then: Is using the remote Load (direct transfer) correct in this application or should I use Remote load (rigid connection)? If Rigid Connection, Why? Screenshot #3 has the explanation of both cases.

Previously attached spreadsheet has a tab called Bolt group (<=25) which has the hand calculation in order to solve the generated shear & axial load when an out of plane eccentric load acts. My end goal is to compare the numbers generated from spreadsheet to numbers generated from FEA study

 

[rb1957]

I really should let someone else reply ... someone who does seismic analysis ... but ...

1) you say "I understand that I should be using linear dynamics." but you also have said you're doing a static analysis. Are you saying ...
I understand that I should be using linear dynamics, but I am doing a static analysis. ?

2) what code are you using ? I don't recognise "remote Load (direct transfer)" or "Remote load (rigid connection)". Reading screen shot 3 (again) it looks like part of a manual describing "remote loadings". It seems to me to be like NASTRAN's RBE (rigid body elements).

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

 

[Stress_Eng]

Just thought you may like an alternative hand calc to compare / confirm results. The calc’s haven’t been checked but they do seem to confirm with yours.

 

[Stress_Eng]

I don’t know how realistic you are intending to model the attachment joints. I understand you want to compare a bolt group analysis to an alternative method, namely FEA. If you intend to just use RB type elements to fixed points, then that would be similar in methodology. If you are intending to have an I section base support, and seeing the unit has flanged attachments, it may be worth thinking about the differences in the compression and tension load paths at the joints, and the subsequent differences in stiffness. For example, for a joint in tension, stiffness will include bolt tension and flange bending / shear through thickness deflections (possible heel & toe). When in compression, the load path will be straight to the I section web. To what extent the joints are going to be simplified, will have an influence on joint reactions.

 

[Stress_Eng]

Just one other observation. In screen shot 1, reaction forces from the top cylindrical item and the supporting S shaped frames on each side will come out at the 2x 4 bolt group sets. You could do a stiffness weighted 3D bolt group analysis under x,y,z unit load cases, then apply those to the analysis of the larger unit.