Torsional Natural Frequency

[JonReynolds]

Hello all,

This is my first post.

I am trying to perform an analysis on a model of a crankshaft which will tell me the natural frequency of the shaft in a twisting manner, i.e. torsional. I have calculated this via calculations using Holzer's method and I want to verify this with the FE analysis.

I believe this needs to be a modal analysis. I have done a few experiments, but do not seem to be having much luck with constraints and loads necessary. I am really stabbing in the dark as i have no guideance.

Hopefully someone can point me in the right direction as to what I need to do?

Many thanks in advance.

Jon Reynolds

 

[mch]

Dear Jon,


Try doing a 'free-free' modal analysis, which is enabled by selecting 'Unconstrained' in the modal analysis definition box. You'll get lots of bending modes as well, so ask for 10 ro 20 modes. If you have a rough idea of the frequency from the Holzer method, then you could ask for modes in a range which brackets this value, but I prefer to get all the modes and look at them.


Carrying out a constrained modal analysis of a crankshaft is particularly tricky as the crank is restrained on bearings, and calculating the bearing stiffness (especially bending) is hard, and then there's the problem of connecting the crank to the bearing springs.


We offer an advanced Mechanica course which covers these aspects and a lot more. If you're interested, give us a call off-line.


Regards,


Rod

 

[JonReynolds]

Brilliant!


Thank you very much for your reply, makes much more sense now.


I had 4 modes showing and one which was closest to the frequency i am looking for was basically showing the flywheel flapping up and down, so it was not the mode I was looking for.


So am I right in thinking that different modes are different frequencies in different directions...one of which will be the torsional one alone I am seeking?


Again, thank you for your quick response, I appreciate it greatly.


Regards,


Jon Reynolds

 

[mch]

Yes the torsional mode will be the one which looks like it's growing radially. It's easy to see this by animating the modal displacements. Other modes will be bending modes of the crank/flywheel assembly.


Cheers,


Rod

 

[JonReynolds]

Cool.

So is there no way of telling it to just examine the torsional behaviour and ignore all bending possibilities?

I picture it in real life as grabbing the end of the crank and twisting it and then suddenly releasing it so that it vibrates axially. Is it not possible to set up a test like this?

It seems to take an age to solve as I guess there are many possible modes it is simulating, when you are only interested in one.

Regards,

Jon Reynolds