Member Posts: 9
edited February 16

Hi all,

I am given the task of FEA of a rotating shaft. Basically I have a belt driven shaft supported by two bearings on two ends. The structure on which the lower part of the bearing is attached is vibrating (using a rotating eccentric mass whose specs are unknown) and is connected to the ground using a number of strong springs. However, we know the amplitudes in x, y and z directions and frequency of vibration of the structure.

My intention is to find the critical speed of the shaft , and to locate the weak points in the entire assembly.

So far, I tried with a simplified model as attached below:

Following are the inputs I have provided:

## For MODAL:

1. Geometry : stepped shaft, face split at the two ends to provide bearing connection
2. Connections :Ground to body bearing at both ends selecting the split faces (purple color in the above image)
3. Cylindrical support with all axial, radial and tangential directions fixed.
4. Both end faces fixed
5. Center of the end faces with remote displacement (this worked with static structural but in model it gave a warning "Participation factors may not be accurate if Remote displacement is applied as constraint using deformable behavior. Please use rigid behavior")Rotational velocity 0.21 rad/s (shaft is supposed to rotate at 2 rpm) [ I read in some previous discussions that rot. vel. in modal affects the resonant frequencies through gyroscopic effects. However I got no change in the frequency values for even a rot. vel. up to 10000 rad/s. Is it because I have not given any damping input, if so where would the input placeholder be?]
6. In Analysis settings : Damped - Yes , Coriolis effect - On, Campbell diagram - Off. Is the methodology correct? I am getting frequencies > 1250 Hz

## For Static Structural:

• 1 ,2,4,5,6 repeated
• Standard earth gravity for weight effect
• Number of steps = 1 , step end time = 120 sec (because I thought to complete at least one rotation since the speed is 2 rpm)
• Getting very low deformations ( ~ 0.165 micro meter ) and 0,24 MPa max. stress (these seems reasonable considering the material and geometry)

Queries:

1. Am I doing the right steps?
2. If not, kindly guide _/\_
3. How to simulate the effects of the sinusoidally vibrating base (single speed)? (I checked the harmonic analyses section which left me more confused as it was supposed to give a range of frequencies and a single phase angle - whereas my x , y and z vibrations differed with each other in phase)
4. What about the belt drive ? should I add a moment for that condition? Or does the rotational velocity already consider the moment applied?
5. I want to estimate the fatigue life of the components/ assembly as well.
6. What all information can I extract t be useful for my scenario?

Tagged:

• Posts: 262Ansys Employee
edited February 19

Hey @psansys,

Please check out this link - https://www.youtube.com/watch?v=XHDY6Z92GE4. For rotor dynamics it is preferred to add a bearing connection since it takes into account the elastic behavior of the bearing by additional stiffness and damping inputs. You could check the help documentation for additional info on bearing connections Bearings (ansys.com). Also you could simulate the vibrating base by adding sinusoidally varying forces in harmonic analysis. Add force defined by components and set a magnitude and phase for each of the components independently. Let me check about Q5/6 and get back to you.

Regards,

Ishan.

https://forum.ansys.com/discussion/3978/how-to-access-the-ansys-online-help

• Posts: 22Member
edited February 23

@1shan Thank you so much! I'll check it out!

About the force and it's components, I have the lab data of displacement values at a particular sensor location only and not the force as such. So, will using the displacement give equivalent results?

TIA.

• Posts: 262Ansys Employee

Hey @psansys,

Yes, that would work. You can specify a displacement base excitation. Insert displacement>baseexcitation "yes">insert sensor data in a tabular format.

Regards,

Ishan.