Learning Forum

[javat33489]

Hi all.

The question is this. I'm counting on fatigue. At the first stage, when I write to the binot file, singularities arise in the calculation. I would like to remove them from the binot file and d3plot file. Do I need some kind of filter for this or how can I edit these files after the calculation? For example, I want to remove all stresses above 500 MPa.

[igandiko]

Hello, I don't think there is a way to rewrite binout/d3plot data with a stress cutoff. Perhaps you can either make the analysis stable or reduce termination time if the singularity occurs after a certain time. 

Please take a look at *FATIGUE_LOADSTEP. LS-DYNA performs a fatigue calculation only for the stress / strain history extracted between TSTART and TEND defined in this card.

Note that *FATIGUE_ELOUT can be used to calculate fatigue results of a few elements in the model instead of the full model. 

[javat33489]

Thanks for the answer. I understood you. What can you say about the MAT ADD EROSION card? Maybe it's worth removing singular elements during solution?

[igandiko]

You can delete elements that are probematic, on the fly, to continue the solution. However, please refer the tips provided here (https://ftp.lstc.com/anonymous/outgoing/support/FAQ/instability.tips). 

[javat33489]

Maybe tell me from your experience? MAT ADD EROSION card?

 

 

 

​

[igandiko]

*MAT_ADD_EROSION is primarily used to define general material failure criteria irrespective of material type, not for handling instability. 

You can of course use *mat_add_erosion to delete elements that meet certain stress/strain cutoff or even minimum time step based deletion. The card supports several options. Please see the manual. However, you don't want element deletion to compromise the overall integrity of the analysis i.e. by deleting too many elements that could belong to primary load bearing stucture, unless the deletion makes sense from material failure point of view or for computational convenience that does not alter the analysis significantly while getting rid of distorted elements. I suggest you visually inspect d3plot animation at small intervals and check what is going on in the analysis that is causing the stresses to jump. The tips suggested in the link should help. 

[javat33489]

Yes you are right. But I can only remove singular elements. For example, I can remove everything over 650 MPa:

 

Otherwise, I cannot carry out a fatigue analysis, the stress exceeds the permissible limit according to the Wöhler curve.

 

[igandiko]

If you just want to ignore those elements for fatigue calculations, why not use *FATIGUE_ELOUT card like I suggested earlier? With *FATIGUE_ELOUT, only the elements that are in elout binary database are used for fatigue calculation, rest are ignored. You can control the elements that are dumped to elout database using *DATABASE_HISTORY_SHELL/SOLID. 

[javat33489]

 

It would be great to ignore singular elements. I tried to do this and it didn’t work. Can you tell me in more detail how to do this? Yes, I record data in a file when performing a fatigue analysis. I would like to skip recording stress that I do not need, that is, those exceeding certain values.

 

[javat33489]

I think it is very inconvenient to manually detect the number of the singular element and then enter them one by one into the HISTORY SOLID card

[javat33489]

If I used the MAT ADD EROSION card:
1. Is it enough to use only one SIGY parameter? The rest seem to be unnecessary or unknown. This parameter would indicate the voltage above which it should be removed.
2. Is it possible to turn on the MAT ADD EROSION card not immediately but after some time? For example, after half the calculation it turns on and starts working?

[igandiko]

1. Yes, the card allows several options. You can ignore the ones you do not need. 
2. Please check if FAILTM parameter in *mat_add_erosion helps. 

[javat33489]

I need help. Could you tell us in more detail how to do this?

[igandiko]

 

It is not possible to exclude those elements from elout database without knowing the ids or (areas of those elements) beforehand. Did you try element deletion option? 

 

[javat33489]

Delete after calculation or during calculation? Can you tell me more please?