Modal analysis (boundary condition problem)

  • 148 Views
  • Last Post 04 August 2019
AbdulahALMUTAIRI posted this 29 July 2019

Hi All,

Hope you are well

I am using Workbench to get the natural frequencies for the rail as is it shown in the below screenshot

I have tested the rail in Free-Free boundary condition in the lab where it is supported by foam and modelled it in Ansys workbench in Free-Free condition (without any restrictions on the boundary condition) and I got similar result with error percentage (difference) from 1% to 4% for the first 5 modes . when I modelled the rail where it is placed on hard wood sleepers (where it is placed in the lab, second boundary condition) the model result are completely different from the experiment result.

Can any one please help me wither is my boundary condition in Ansys is wrong or there is a problem in my contact (connection between the rail and the sleepers).

 

 

current Boundary condition : Fixed support

Order By: Standard | Newest | Votes
Wenlong posted this 30 July 2019

Hi,

Thanks for the interesting post. 

It would be easier to give an opinion if we can know the experiment settings. For example:

  • Is the rail fixed to the wood? or it is resting on the wood and there is no bonding in between?
  • Are you measuring the natural frequency of only the rail or the whole system (rail + wood)? In the current simulation you have, you will get the natural frequency of the whole system.

Bests,

Wenlong

AbdulahALMUTAIRI posted this 30 July 2019

Dear wenzhang Thanks for your interaction * The Rail in the lab is just resting on the wood sleepers, and the sleepers is resting on the ground.

  • I am measuring the whole system natural frequencies I have already measured the rail in Free-Free boundary condition (without sleepers) + no restrictions on the boundary condition in ansys

Best regards

Wenlong posted this 30 July 2019

In this case, I would suggest frictionless or frictional contact instead of bonded contact. And you probably need to change the fixed boundary condition as well. 

Bests,

Wenlong

AbdulahALMUTAIRI posted this 30 July 2019

Thank you But what should I put the boundary condition ? Other than fixed from bottom of the sleepers?

Regards

peteroznewman posted this 31 July 2019

Modal analysis is a linear analysis. That means ANSYS has to convert a nonlinear contact into a linear contact to do the modal analysis. Therefore a frictional contact that is closed is automatically converted into a bonded contact to do the modal analysis. That is why you get the unrealistic simulation result you show in this image.

It would be easier to match lab data to simulation data if instead of a wide flat wooden sleeper, you put a narrow rod under the rail. Then in the simulation, you could split the face where the rod touches the rail, and make that split line a remote displacement and constrain just 3 DOF and leave the rest free. Then you don't need the sleepers in the model at all.

  • Liked by
  • AbdulahALMUTAIRI
Wenlong posted this 31 July 2019

Oops, Peter is right. Frictionless and frictional contacts are nonlinear contacts, and in modal analysis, we can only have linear contacts. 

  • Liked by
  • AbdulahALMUTAIRI
AbdulahALMUTAIRI posted this 31 July 2019

 

  Thank you peter for explaining

 

But how can is it possible to replace the sleepers with a rod? I have inserted the sleepers material property in the Engineering data and if I change the geometry of the sleepers the mechanical properties will change. Is it possible please if you can explain it with an example ? I would really appreciate it 

AbdulahALMUTAIRI posted this 02 August 2019

 

Hi peter

 

I could not understand your idea, because my experience is limited in this subject

 

is it possible please if you could illustrate your idea with an example

 

it would be really helpful 

peteroznewman posted this 02 August 2019

Here is a rail with the bottom surface split at the two locations where, in the lab, the rail will be supported on cylindrical steel rods instead of a wide wooden sleepers. The rail is along the Z axis.

Here at the two Remote Displacements, one applied on the front, the other the back split line.

AbdulahALMUTAIRI posted this 03 August 2019

 

Dear Peter

 

I can not replace the wooden sleepers with cylinder rods because I want to investigate the rail behaviour when it is supported by the wooden sleepers.

 

I have included in the engineering data the sleepers material properties

is there any other way ?

Thank you very much

 

peteroznewman posted this 03 August 2019

You studied the agreement between a rail supported by foam and a free-free modal analysis.

I suggest you study the agreement between a rail supported by rods and a remote displacement supported modal analysis.

This is one step closer to support on sleepers, but is easy to do in the lab and in simulation.

Putting a rail on a wooden sleeper without any other constraint may not be that different to the rods.

If there is a significant difference, then you can no longer use modal analysis. You need to run a full transient analysis that can have frictional contact between the sleeper and the rail.  How do you excite the rail in the lab?  Do you hit it with a hammer?  You will have to have an impact in the transient simulation.  Extracting the data from the simulation will be similar to extracting data off the rail in the lab. How did you determine the frequency and mode shape in the lab?

But in reality, the rail does not rest on the sleeper, there are spikes that hold it down. When are you going to add the spikes?

 

AbdulahALMUTAIRI posted this 03 August 2019

Hi peter

Thanks for your grate support in helping solving this problem 

How do you excite the rail in the lab? 

*Using instrumented hammer ( excitation points are only at the top face of the rail )

Do you hit it with a hammer?

* YES 

How did you determine the frequency and mode shape in the lab?

* Using a laser vibrometer (capturing the response only in the vertical direction with its software ( it can give you : time domain or frequency domain )(FFT&FRF)

When are you going to add the spikes?

* The study considers just the rail resting on the sleepers without adding any spikes.

peteroznewman posted this 03 August 2019

I recommend you build the modal analysis model using remote displacement BCs and see how close it comes to the lab results of the rail resting on the sleepers. I expect it will be closer than the first trial you did with contact.

Then you can do all the work needed to simulate a full Structural Transient model and strike the rail with a hammer and simulate the ringing while extracting all the vertical displacements to feed that simulation data into the software that computes mode shapes and frequencies.

AbdulahALMUTAIRI posted this 03 August 2019

 Hi peter

So you mean I have to first try remove the sleepers and replaces them with a remote displacement by splitting the bottom face of the rail  at the place where the sleepers ?

but what about the sleepers engineering data ?

it will not be included in the model, I mean is it will be logical ? 

Thank you

peteroznewman posted this 03 August 2019

So you mean I have to first try remove the sleepers and replaces them with a remote displacement by splitting the bottom face of the rail  at the place where the sleepers?

Yes, that is what I mean. You don't have to do it, I just think it is a good idea because it is quick and easy compared with the next logical step.

but what about the sleepers engineering data ?

Will not be relevant to this model.

it will not be included in the model, I mean is it will be logical ? 

If the model matches the experimental data better than your first model, then it is a better model.

AbdulahALMUTAIRI posted this 04 August 2019

Hi peter

 I tried the Structural Transient, because it allows frication contact between the rail and the sleepers

the BC for the sleepers (bottom face ) I have put a displacement on the sleepers and made it free in all directions except Y direction to simulate its BC in the lab, and also simulate the hammer impact using nodal force,  the solution took 9 hours to finish and the result was not realistic, when I play the animation video the rail was flying away from the sleepers, I copied the time domain result (time S vs Acceleration mm/s^s) and then using MATLAB to get FFT

the results was completely unrealistic

 

AbdulahALMUTAIRI posted this 04 August 2019

Now I am trying to use remote displacement after splitting the rail bottom face where the sleepers interact

what settings should I put in my case ?

 

Regards

peteroznewman posted this 04 August 2019

You need two remote displacements, not one.  Delete the one you have because it has created a reference point between the two faces.

Pick one face, create a remote displacement and set X,Y and Z to 0 and Rotation Z to 0 and leave the other two rotations Free.

Pick the other face, create a remote displacement, set X,Y to 0 and Rotation Z to 0 and leave all others Free.

Close