vibration modes ?

[DavidJ]

Has anyone had any success using Simulate to check natural frequencies of a part or assembly?

I get as far as specifying all part to be rigidly joined (it is a welded mixing shaft assembly), and choosing vibrational analysis - after that I can't work out how to proceed. Generally the help/examples are a bit thin on explanation.

Do I have to set this up as a dynamic simulation? - I'd have expected not, just for vibration.

Any tips most welcome, I'm getting quite frustrated.

 

[Alexander]

Hi David,
Did you manage to produce a vibrational analysis for the above, I know its a few years ago but I have a similar need for a project.
I have DST Motion but not the FEA which I am considering purchasing if it's up to the job?

Cheers

 

[Lew_Merrick]

To be clear, Dynamic Simulation Technologies should stay away from things they do not understand! Material Properties and Analysis rank quite high on that list. A compatriot of mine (Dr. Alex Deak) wrote a rather scathing review not long after they took over Visual NASTRAN (I no longer remember where he posted it). One of Alex's more "publicizable" comments was, "These people cannot accurately relate room temperature to absolute zero accurately!"

Now, to be precise, Alex hates virtually all "preprogrammed analysis systems." I had him use FEMdesigner on a project we worked together. Alex's comment was that it (FEMdesigner) was "almost acceptable." [This, as an "evaluation" from Alex is quite high praise.] FYI, Alex was a very senior "systems dynamic analyst" for NASA from shortly after the Moon landing until the early 2000-teens. Alex is one of four people to whom I pat close attention in this realm of things.

 

[Alexander]

Thanks for the advice Lew.
I'm desperately trying to find a way to simulate vibrations for a personal project I have been working on for a few years now.
The frequencies will be of the audible range 20 to 20,000hz and will be structural borne. I am designing a range of platforms on legs which will absorb vibrations from components placed on the platforms.
I contacted DST and they said it is possible but it will not take into account the natural frequencies.
I've looked at FEMdesigner (waiting for a demo download) it sounds okay but does it simulate assemblies and if so how does it calculate the joints of each component.

I've looked at ANSYS mechanical which seems very good but a bit out of my price range.

Any help would be appreciated.

 

[Lew_Merrick]

Alexander -- You end up having to "re-constrain" the Parts making up an Assembly in FEMdesigner manually to get things right. It is somewhat of a PITA, but apparently the Constraints were not well documented back when it was created. The truth is that nobody (except for a couple of USAF and NASA programs I have run into over the years) does a good job of calculation stresses or strains in joints held together with screw threads. [I have a UN Thread Data sheet that provides the shear area per inch of engagement for internal and external threads. It would require 25 such Tables to cover Metric threads properly.]

The main issues for Welded Joints is to get the Material Property definitions correct for the welding filler.

 

[Alexander]

Thanks Lew, it does make me wonder how any of the FEA software can be reliable/accurate when most of the joints can't be simulated.
Probably one reason why your friend Alex hates all analysis systems.

 

[Lew_Merrick]

Alexander -- Take a look at the difference between a screw's Major Diameter Area and the Effective Tension Area sometime...

 

[Plista]

To be clear, Dynamic Simulation Technologies should stay away from things they do not understand!

I will suggest to make technical questions tosupport@design-simulation.com for serious investigations before coming to such conclusion.
Let me point You out what I mean with a classical example in http://www.simwise4d.it/CONTATTO.WM3 (download a demo of SimWise from http://www.simwise4d.it/SimWiseAlibreEN.htm in case You need the code to run it).
The final result out of the drop of the green (verde) and blue (blu) balls is the same, but clearly the story (transition) to reach it is completely different depending on the two different contact models (impulse and custom) used in the same simulation.
If You are looking ONLY for the final deformation of the beam due to the weight of the ball, then both approaches are fine, the impulse model beeing much faster and easier than the other.
But it is clear that such final deformation is not the biggest one during transition, so to model the force distribution vs. time during contact a custom model is proposed assuming very small deformations of the beam.
But of course even this second approach doesn't include the fact that, due to the much bigger deformations of the beam, the ball doesn't jump in the same vertical line but start moving behind and falling out of the beam.
So in this case neither of the two approaches can hit the goal of investigating where the ball will finally go. But even this third case can be simulated as well with this software (is part of the training we can provide), modeling a flexible beam instead of a rigid body in SimWise, so I cannot conclude that this software cannot make the required simulation (finding where the ball will go and the story of deformations of the beam). Of course within the limits of the material properties and modelization of it (linear, non-linear, etc.).
Therefore, my conclusion is that:
- first, think about what are the mathematical models that can accurately describe the aspect of the simulation You are looking for;
- second, select the most efficient and simple mathematical model to reach such goal;
- third, ask the developers if such mathematical model is built in the software You are selecting (if You don't have enough documentation to check by yourself).
Ciao!
P.S.
I placed some other models to rise up curiosity about simplified approaches to typical issues in:
www.simwise4d.it/BALL_FAN_VECT.WM3
www.simwise4d.it/COASTER.WM3
www.simwise4d.it/PSNAP_WITH_ROLLER.WM3
www.simwise4d.it/PRODUCTION_LINE_CONVEYORS.WM3
www.simwise4d.it/TELAIO.WM3
For example, please enjoy travelling with the camera inside the tubes of TELAIO while deforming, to see the simplified (and automatic!) solution for the welding thickness.
What is up to You, again, is to justify if the model is adequate for Your needs, balancing time, economical investment and precision, which is not as simple as a keystroke!

 

[rockfusion]

I was experimenting with your Contatto.wm3 file; does this require Motion + FEA to get the ball to fall under acceleration? I tried to recreate it from scratch but I cannot figure out how to show the ball in motion (only have FEA).

 

[Plista]

Correct.
Please write to sales@design-simulation.com for a free trial of SimWise4D for Alibre (such as, Motion+FEA) on top of Your actual license of FEA only.
Nevertheless, any version of SimWise can play all files, I mean in this case Your (FEA only) license can play a (Motion) or a (4D=Motion+FEA) file (but cannot modify or create the kind of analysis You are not licensed to, as You are trying).
Therefore any version, Motion, FEA, 4D or 4D+ (which is 4D plus fatigue) can play & investigate the above files, but to modify/recreate them from scratch You need the license for the side of analysis You are using, or a free trial demo which is save capable for a limited period of time.
I am going to fulfill more industrial cases in www.lista.it/atnet/index.htm, stay tuned!
Ciao!