Operational Deflection Shapes ODS

Hey all!

I'd like to enhance the investigation of the resonances of a mechanical system model incluing flexibles structures. Is it possible to extract operational deflection shapes from the System after the simulation using the Adams PPT? Or should I stop the Simulation at a certain time step(system state) and do a linear mode analysis?

What are your thoughts on this?

Kind Regards

Nils

0 Kent West over 7 years ago

Hi Nils,

Yes: the typical way to do this is to simulate the dynamic model until it's in a configuration of interest, then linearize at that point and inspect the mode shapes. You can repeat for different configurations that you reach from dynamic analyses.

The 'next step' here is to use Adams/Vibration on the linearized models and apply/measure inputs/outputs in the frequency domain..

HTH,

Kent
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel
0 nils_s over 7 years ago

Thanks for the reply!

Can I define a solver simulation script that

1. Runs a dynamic simulation to preload the system via constraints and external loads

2. Deactivates the forces and constraints

3. Linearize about the last configuration and calculate the normal modes

..?

I just looked up the the commands and I don't find the one for the normal mode Analysis. What does eigen solution calculation do? Can I investigate frequencies and mode shapes with it?

Kind Regards

Nils
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel
0 Kent West over 7 years ago
Hi Nils,

A Solver simulation script for this would look something like this:

sim/dyn, end=5, dtout=0.01 ! Get model into position

deactivate/joint, id=44 ! Remove that pesky joint..

sim/init ! or sim/static Need to do some kind of sim before the linear..

linear/eigen ! produces the 'normal modes' that I think you want.

You might not need to do the sim/init, but consider this:

after you deactivate the joints/forces your model is no longer in a feasible configuration. It *was* in a solution configuration, but now you've changed it. So Solver must do some kind of analysis (static, dynamic or just initialization) to get into some kind of feasible configuration before the linear analysis. You typically have to think about this: do you want a static analysis at that point or a short dynamic? Basically what happens to the real system after you've deactivated things?

HTH,

Kent
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel
0 nils_s over 7 years ago

Thanks!

So if I get you right I'll try the following.

1. Set up the solver simulation script like you proposed

2. Run a scripted simulation

3. Run a normal mode analysis using the simulation control (by hand)

..I just want to get sure that I linearize the model for the time step after which the simulation script ended, and not in initial conditions. How can I actually double check this?

Kind Regards

Nils
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel
0 nils_s over 7 years ago

I am sorry, your answer was partly hidden and I did not realize. :/

I guess in my case it will be better if I replace the constraints with force elements after the preloading, copying the constraint forces at that particular point in time in order to preserve the degrees of freedom of the model.

Kind Regards

Nils
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel
0 nils_s over 7 years ago
So my model is running properly right now, usin the following steps.

preload the model using applied forces and global constraints at the system boundary in a dynamic simulation step

replace constraints by applied forces to regain the DOFs of the system

run a static simulation

linearize the model about the static equilibrium

First the linearized model did not show the resonance frequencies that I see in the dynamic simulation in steady state condition, so I had a look at the documentation to find out what could have gone wrong.

I came over the RM option of the LINEAR/EIGENSOL command and what I found out is that there is a significant dependency of the calculated eigenfrequencies to the reference marker you chose for the linearization of the model equations. In my case I did not even chose a reference marker on a moving part. I just used one ground marker which is at the force interface where the resonance frequency showed... Now I am struggling to understand the difference

Do you have experience in doing this? What are your lessons learned?

Kind Regards

Nils
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel