different results from this same simulation

[marcin.blaszczyk]

hi,

for more complicated shapes (see attachment) I get strange results... I mean it is like this:

Fixing, Moment, Material and Force are not changed during this test.

I turned on "create curved mesh elements", Run Simulation: safety factor result: 4,5

I turned off "create curved mesh elements", Run Simulation: safet factor result: 11,4

I turned on "create curved mesh elements", Run Simulation: safety factor result: 4,23

I turned off "create curved mesh elements", Run Simulation: safet factor result: 11,4

I turned on "create curved mesh elements", Run Simulation: safety factor result: 4,23

I turned off "create curved mesh elements", Run Simulation: safet factor result: 11,27

etc. values are not the same all the time! they're floating

for less complicated models values are stable

what is wrong here?

[JDMather]

---

@marcin.blaszczyk wrote:

.....what is wrong here?

---

Probably nothing.

Are you running Convergence Analysis?

Attach your file here.

[ccamara01]

I would assume that a slightly different mesh is created on each mesh update.

[marcin.blaszczyk]

JDMather: unfortunately this particular one can not be shared with outside world

Ps. "Covergence analysis" - which option are you writing about?

ccamara01:  I thought that algorithm is an algorithm... so you say that results depends on, hmmm, computer's mood ? 🙂

[JDMather]

---

@marcin.blaszczyk wrote:

JDMather: unfortunately this particular one cannot be shared with outside world

Ps. "Covergence analysis" - which option are you writing about..

---

Do you see this behavior only with this particular file?

Are you saying it can't be reproduced in a simple example file?

What book are you using to learn Inventor FEA?  Most books would cover the topics of mesh and Convergence.

[marcin.blaszczyk]

this happens only for more complex shapes

see movie attached

[JDMather]

I was expecting your file so that I could demonstrate why it is happening with your file.

http://forums.autodesk.com/t5/inventor-general-discussion/easiest-way-to-learn-fea/td-p/5484707

[marcin.blaszczyk]

I'll prepare another model with the same effect soon and I'll upload it here.

[JDMather]

Before you spend a lot of time on this, you might want to go through this -

http://au.autodesk.com/au-online/classes-on-demand/class-catalog/2014/inventor/md6583

especially Reconciling Differing Results video and the Handout(.pdf)

Basically it is all in the mesh and you have changed the intial condition.

Doing a Convergence Analysis should converge on "the answer".