Brake squeal analysis

Altair Forum User

Hello,

I'm trying to make a complex eigenvalue analysis with a non linear quasi-static as preload for a brake model.

The aim of this work is to determinate unstable modes and to study squeal phenomenon.

 

Using the full version with solvers patch 14.0.210 I succeed to launch the analysis, but the modes extracted are all stable.

I tried lots of contact configurations: with the HM contact manager, using gap elements and using superelements with DMIG matrix, but Optistruct always gave the same results.

 

The only way to have unstable modes is to use frictional elastic slip distance , but in this way the calculated modes have too low frequencies and too high dampings that is not correct (I have correct results for this model from someone who used another software).

 

All the used values are appropriate to simulate squeal phenomenon and congruent compared to the respective units of measure.

 

Does Optistruct allow to perform this kind of work??

If yes, could anyone help me, please??

 

Thanks

Alberto

Altair Forum User

Hi @Alberto B.

 

How are you performing the analysis?

 

Brake squeal can be studied as a quasi-static process – a two step solution:

1. A nonlinear static solution to obtain the static equilibrium of brake assembly under the applied loads

2. This is followed by the extraction of the complex eigenvalues

Altair Forum User

Hello Prakash,

yes I set a non linear quasi static as first loadstep and then a complex eigenvalue analysis as second loadstep with the nonlinear loadstep in STATSUB (PRELOAD).

I used a sample solver deck, that Rahul R sent me, as guidelines (the name of this file is brake_squeal_example.fem), but with the same settings my model gave uncorrect results.

 

In the image attached, you can see the results without FRICESL (fixed slope KT approach from the online help) on the left and those with setting of FRICESL on the right.

The results on the left are bad for the too high and too low values (for example frequencies are too low); those on the right are bad because there aren't unstable modes (real parts of eigenvalues are all negative).

I tried to change ND values in EIGC and EIGRL ,but I always got uncorrect results.

 

Let me know if there is something i could try for solving this problem.

 

Thank you for the help!

Alberto

 

 

 

<?xml version="1.0" encoding="UTF-8"?>

Altair Forum User

@Alberto B.

 

Can you share the model with us?

 

Use this File Transfer LINK

Altair Forum User

Hello Prakash,

I sent the model.

 

Thanks for the help

Alberto

Altair Forum User

Hi @Alberto B.

 

The reason for such low frequency or I should say it as rigid body modes is because there are free nodes every where.

 

Have you modelled (meshed) in HyperMesh?

 

My suggestion is to create a fresh mesh and run the analysis.

Altair Forum User

Hello Prakash,

I imported the geometry from a STEP file, then I meshed all on HyperMesh.

I did the equivalence for the nodes in common for each body and imposed SPC as the model that i saw on other software.

 

How is it possible that there are free nodes??

 

Thanks 

Alberto

Altair Forum User

What is the version of HW are you using?

 

<?xml version="1.0" encoding="UTF-8"?>

Altair Forum User

Hypermesh v14.0 full license and solvers patch 14.0.210

Altair Forum User

I've solved the problem, with optistruct 14.0.210 the results are uncorrect, but updating with 14.0.211 the results become correct!

Thanks for the support

Altair Forum User

Hi Alberto,

 

Thank you for the update. I will report this.