Results Ansys vs Catiav5

[Lepetit]

Hello all,

I have made a FEM analysis in Ansys a Catia V5. I have different results and I can't explain why, I attach a picture first of one error in Catia that I don't understand so well.

The force applied had been 1000N.

I attach also a picture in Catia with the system analyzed where it's possible to see the geometry and where the reference axis is located, another where it's possible to see the name of every point analyzed, the results of the sensors in Catia.

The results for Ansys is:

FA(N) FB(N) FC(N) MA(N*m) MB(N*m) MC(N*m)

x -230.7 230.1 0.6 -4.4 -4.3 -18.2

y 133.2 132.8 734.8 -7.5 7.5 0

z -86.1 -86.5 754 -0.5 0.4 -0.1


Thank you in advance.

Cheers.

 

[irq]

If you do not attach any presentation, detailed description or comparison of your results I doubt if someone here can help you. Sorry, but I do not understand what is your issue and without additional effort from your side we can do nothing.

Regarding both tools - Ansys is a package for advanced simulations and analysis modul in CATIA V5 is in my opinion only usable in preliminary design and not suitable for certification purpose. Some times ago we did comparison between CATIA V5 and NASTRAN on not so complex structure (but not so simple like cantilever beam etc.) and the level of deviations especially in high stressed regions was unsatisfactory (non-conservative). Does anybody has different experience? I think CATIA results are generally not accepted by authorities even for the interface load extraction at least in aerospace industry.

 

[Lepetit]

Sorry I thought I was atached it!

Thank you

 

[rb1957]

the catai error is easy to explain ... 3D elements only have translational freedoms (ie x,y,z at each node) and no rotatioanl freedoms.

something looks "odd" with the reactions (other than the european "," for ".") ... if you applied 1000 N, it's hard to see how the reactions sum to react 1000 N.

i also would not trust a CATIA FEA ... i think there are many simplifications that can produce mis-leading results.

have you tried modelling a simple example problem ?

have you tried a hand calc ... for example, if the load was 1000N down, i'd assume a reaction of 500N at C and 250N at A and B (and of course the side component at these points).

your title says you're getting different results between CATIA and ANSYS ... what are the ANSYS results ?

 

[Lepetit]

hello,

Yes the values of Ansys are in the first post, maybe it's not so clear, otherwise all of you say that it's better to trust in Ansys so I will do this.

Thank you very much

Cheers

 

[rb1957]

ah, ok ... "none so blind (as those that will not see)"

so your reactions in both cases sum in Y to about 1000, which is clearly reacting your load.

other than a typo for Cz (754 ANSYS, 174 CATIA) ... CATIA nicely balances the z reactions at A and B.
the major difference seems to be the moment reactions ... if C is modelled with solid elements, i find it odd that there are any moments there
the moments (from ANSYS) in x sum to a -ve, which is offset by an asymmetry of A, B, and C
CATIA similar results, but larger (-76 Nm vs -27 Nm from ANSYS). the difference in moment is coming from somewhere in your model, CATIA isn't guessing the numbers. i think the problem is that CATIA FEA isn't very open, i don't think you can get a NASTRAN f06 file (ie all the internal results), i think you can only get a very limited set of "pretty pictures".

for both CATIA and ANSYS, the FEA understands that the model is in equilibirum, so the moment is somewhere in the CATIA model.

 

[Lepetit]

Thank you again!

So definitely I will trust more in Ansys than in Catia.

Thank you again!!
Cheers