Learning Forum

[javat33489]

Now I am solving the problem of teeth interference fit and their failure.

Now I'm solving the problem of engagement. And it is important to me with what force the teeth will jump. For this I use force reaction. I use metal ductility, bilinear isotropic hardening and large diflection. I have a constant error with the grid DISTORTED 9999999. If I set the stiffness factor to 0.01, then the problem is sometimes solved, but this gives a strong error in the calculation, refining the grid does not help, I tried different triangular and square grids. I also tried using NORMAL LAGRANGE.  I use friction contact 0.1. Augmented Lagrange, Axysymmetric, Nodal Normal to target, Add Offset ramped effects.

All teeth are rounded.

Please give me advice on how to solve this problem. Thank you.

[Ashish Khemka]

Hi,

Can uou share snapshots of the set up? Please see if the following link helps:

 

Dealing with Convergence Issues - Element Distortion Error - FEA Tips

 

Regards,

Ashish Khemka

[javat33489]

Thanks for the info, but this is for beginners.

Attached screenshots, maybe it will help:

Stage 1 ekill is successful, stage 2 elive successful, when the teeth are stretched, the movement begins and 80% of the grid error is resolved:

 *** ERROR ***                           CP =    1909.812   TIME= 15:02:06
 Element 3170 (type = 2, SOLID186) (and maybe other elements) has become 
 highly distorted.  Excessive distortion of elements is usually a        
 symptom indicating the need for corrective action elsewhere.  Try       
 incrementing the load more slowly (increase the number of substeps or   
 decrease the time step size).  You may need to improve your mesh to     
 obtain elements with better aspect ratios.  Also consider the behavior  
 of materials, contact pairs, and/or constraint equations.  Please rule  
 out other root causes of this failure before attempting rezoning or     
 nonlinear adaptive solutions.  If this message appears in the first     
 iteration of first substep, be sure to perform element shape checking. 

I also tried making the grid more frequent:

But it doesn't help. I solved the problem in static structural. Maybe I should switch to transient?

[javat33489]

managed to solve most of these problems in transient analysis. The calculation took much longer. Some issues are still unresolved. The problem is either in the grid or in contact:

small mesh:

I will try to solve the problem on Augmented Lagrange and using predict for impact. I don't know if it will help.

[javat33489]

I need your advice on how to solve such problems! Help me please.

[peteroznewman]

I don't understand the purpose of ekill and ealive.  Which elements are being killed?

Without reference to ekill and ealive, what physical process are you trying to simulate?

This looks like it is a small angular slice of an axisymmetric model.  Are you trying to simulate a complete cylindrical pipe with teeth on a ring with teeth, or are you trying to simulate just the geometry you show?

[javat33489]

I count one segment, in total there are 10 of them in the estimate, they form a cylinder with cutouts. This is a collet. I'm simulating preload because first the collet teeth will hit the teeth of the lower element and then they will break. I simulate a breakdown. I do the tension with the help of ekill. I try to use square elements rather than triangular ones. In the course of calculations, grid errors of 80% constantly occur. After spending 20-30 calculations, I realized that the matter was in contact. I used Augmented Lagrange - Nodal Normal To Target - Stiffness Factor 1 and I managed to get a solution.

But I would like to receive recommendations on how to proceed in such cases of errors? What life hacks to use and what to apply first? Share your experience? It is very difficult to carry out 20-30 calculations without knowing where to start.

[peteroznewman]

I recall you showed in another thread slots in the outer cylinder (collet), so this is one leg with a slot on either side, but the teeth form a full ring. 

When you say you want to create tension, do you mean that there is radial interference when the teeth are nested peak-to-valley?

Is it frictional contact elements that are being killed in Step 1?  If so, why is that necessary?

How do you know that the collet teeth will break off? What if the finger of material simply flexes up so the inner ring can get to the peak-to-peak position before sliding down the back side of the tooth?