Showing results for 
Search instead for 
Do you mean 
Reply

Solution 601/106 problem

Dear world,

I am having the following problem: 

I want to model the flattening of a pipe, using solution 601/106. I got some results. My question is:

I have used a plastic material with the stress-strain curve prezented in the Elemental Nodal stress.rar file (stress-strain.jpg). The final value for stress is 490 MPa.

The two archives present the results for a total deformation of 85 mm, that is elemental and element nodal stresses (see the archives with the corresponding names). As you see, values much greater than 490 Mpa are present, although  the XTCURVE is set to Not Extended, and I would have expected that values greater that 490 will not be present.

Do I make a mistake? 

Many thanks

Ionut

4 REPLIES

Re: Solution 601/106 problem

Hi Ionut,

 

Stresses are dependent on loads, so if you are getting high stresses your loads are high compared to the laod carrying ability of the part.

 

In your model, to quickly get a ballpark stress value check the P (load) / A (area) at the critical cross-section. This will only give you a ballpark at this does not account for the fact that as the sides are compressed the critical sections will see some hoop effect from the compression.

 

Also when comparing elemental vs. nodal results make sure that for elemental results do not compare the average or minimum stresses.

 

Finally, check the element quality and if necessary use hex elements instead of tet elements.

 

Regards,

 

Saptarshi

 

 

 

 

Re: Solution 601/106 problem

Hi,
I do nit have loads, only a prescribed displacement. My reasoning us that,
as long as the stress-strain curve is not extended, in NX, the elements
where stress goes beyond the ultimate stress limit are canceled, so it is
not normal to have elements/loads with steesses larger than the ultimate
stress 490MPa in my case).
Ionut

Re: Solution 601/106 problem

[ Edited ]

Hi Ionut,

 

I am not sure what you mean by: "as long as the stress-strain curve is not extended, in NX, the elements where stress goes beyond the ultimate stress limit are canceled".

 

Even if you apply a displacement, (representative of compressive load) at the flat faces, at the critical section (curved section) it is not just compression that is causing the high stress. 

 

If this was a flat tube with one end fixed, and apply load on the other end (similar to a column), you will notice that the deflection matches the stress strain curve.   

Re: Solution 601/106 problem

Is the NX forum a little short on Advanced Nonlinear expertise?

Anyway, firstly, I would strongly suggest you run this problem with a brick mesh, which would allow you to get excellent results with less elements. It's a tad more meshing effort, but it should be very easy to do and control for that shape, and your runtimes will be shorter.

You should look very carefully at the Advanced Non Linear Theory and Modelling Guide docs, but I think either the entire element or (at minimum) the element centroid must exceed the material failure threshold before the element will be removed from the analysis. This means that certain points in the element could exceed the material failure threshold before the whole element is removed from the model. You should look at the plastic strain results as well. Plastic strain might be only an element centroidal result, not an element nodal (corner) result. If so, I am not sure how NX will show (if it can show) an element nodal result for a result type which is only element centroidal. If it can show it, then the unaveraged element nodal result may show a single value for the whole of each element.
Regardless, if the plastic strain result is only a centroidal result, this may also explain why corner stress results can exceed the material failure threshold - which does not occur until the element centroid reaches the stress value.
I hope that helps.