How to simulate a rotor-shaft system? 1

[Lirock]

Hi,everyone:

I want to simulate a rotor-shaft-bearing system,such as the shaftline of a hydraulic turbine, to determine the deformation and critical rotate speed.The Some guys said mass,beam and spring elements can make it. But I didn't catch it.
Is there someone has any idea about how to do it?

 

[cbrn]

Hi,
yes, as a general input, you've been told correctly: shaft -> beam188 elements (must support gyroscopic effect) for example
shaft mass -> directly determined by FEM as soon as you input a valid material
added mass(es) -> concentrated masses on shaft node(s)
distributed masses -> beam188 elements with "fictitious" material density and "fictitious" section
bearings' oilfilm -> combin14 spring/damper elements in 3 orthogonal directions
bearings' structure -> combin14 in 3 orth dir

Don't forget to activate Coriolis effects and input a rotation speed.
You can also calculate a Campbell diagram in this way... There's an example in the Help file.

If you need deeper explanations, please better specify which are the troubles you are experiencing. Is it the first time you perform rotodynamics with Ansys? If so, start with something VERY simple, something you can benchmark against handcalcs.

Regards

 

[Lirock]

Hi,cbrn
Yes,it's my first time to perform ratodynamics with ansys. Thank you very much for your explanation. I will work according to your advice.
But I still can't understand why there must be a rotation speed. As I know, critical rotation speed is the first set natural frequencie of the structure in fact. So the simulation could be treated as a modal analysis in which all loads except constrains are useless.
Then why ?

 

[cbrn]

Hi,
no, you're not correct. As long as you speak about "natural frequency", then yes, it is the natural response of the system. The first natural frequency is the lowest eigenvalue which solves the eigenproblem in a non-trivial way (the trivial solution is(are) the rigid-body motion(s) ).
But the "critical speed" involves the gyroscopic effect, which varies with the rotation speed, and by definition is "the rotation frequency at which the natural frequency equals the rotation frequency itself". This applies of course only to bending eigenforms.
For each bending eigenvector, there are two critical modes, one with the gyroscopic effect "in phase" and the other "in antiphase": it's also said "forward-" and "backward-whirl". One tends to "unstabilize" the rotor (thus lowering the frequency of the critical speed wrt the natural frequency), the other to "stabilize" it (thus making the critical frequency higher). There are physical reasons why the only whirl to be considered is the "stabilizing" one (try getting some good rotodynamics books).

Regards

 

[Lirock]

Hi,cbrn
Thank you again for your excellent explanations from which I can learn a lot.