from many screenshot provided, i"m only guess the model is box shape with diaphragm at every space and shell element being used. Load and support placed at pin holes with rigid constraint. Also,some transform function was assignee at inclined support.
Other FEA software reported stress of shell element at every layer position (top, mid, bottom) in which CalculiX is not due to expanding to solid element.
If the model with shell element is in bending dominate, someone need to properly displayed the result in layer position. This is may negligible for the case of membrane dominate since the the stress nearly uniform at every layer.
If the results of model look reasonable overall except at some location even without transform function, maybe the model of shell element is not using continuous mesh or glue/tie constraint is not properly assigned.
So, i will double check at support boundary condition of two pin at mid length. Creo Sinulate shown rotation released make it produced membrane stress in highly compression at bottom. In contrast, CalculiX model result has not.
¿Is your discrepancy smaller than, let’s say, a 4%?
If yes, your design code safety coeficients should parefectly take care of it.
By other hand, I understand there is a limited time for each project but, I would recomend you to keep investigating up to clarify the discrepancy. One almost never know where it is going to end but in my experience, I have to say it is always worth and gratifying.
Do not assume that the commercial option is the correct one.
In general, I have not found a problem, where the result would be OK for one part of the model and not OK for the rest of the model. I did have some problems with the PaStiX solver or even Spooles solver and in such cases, the Pardiso solver gave me the correct results. Did you try using the Pardiso solver?
Since only a part of the solution is not OK and you are using a Transform keyword, try comparing the stress invariants or displacement invariants. Those are the stress and displacement components independent of the coordinate system. For stresses, you can compare Mises, Tresca, and principal stresses, and for displacements the displacement magnitude.
Discrepancy shown too large since it occur at support as critical locations, maybe it’s about 80% and more so any tolerance does not help.
Creo Simulate shown reasonable results. I still doubt in PrePoMax or CalculiX models, better approach may reach by including model of rod pins with contact analysis. This can eliminate inappropriate assignment of transform function and rotational release.
80%?¿ . How did you deduce that ???.
If discrepancy is about 80% (don’t know how you deduce that) it needs to be checked for sure.
The deformed view looks like the lifting maneuver of a jib prior to its assembly in the main body of the crane. Stress distribution doesn’t seem the natural working position of that kind of pieces .¿isn’t it?
I see uneven shears at the bottom and some transverse displacement. Maybe it’s a piece intended for demolition. Don’t know.
Oh wait a minute.!!
By other hand , why displacements with respect to the undeformed position are shown at the opposite side in Prepomax with respect the simplified version. ¿?¿
Only quick viewing based on screenshot provided, PPM/CCX only produce stress in average of -24MPa which unreasonable. In contrast Creo shown high stress about -120MPa and more.
So this discrepancy may not be ignored or covered by safety factor of code regulation.
Thanks for giving sketch of problem. Looking on support descriptions, the stress in location between pin support should not too high since an inclined support resist horizontal movement.
But i did not known, why Creo Simulate shown too high. Maybe cause of one pin support being released in horizontal movement.
Thanks for sharing. Solving your issue will be rewarding. I’m completely sure.
One check you can do is to see if the modal vibration analysis of your first rigid body modes is compatible with that peculiar sliding support at a certain angle.
I agree with you it is not easy to see without clear references. Try to build some dummy elements.
Model completely free . Then constrain rotations of the REF node. Check the translational modes to be sure you have set up properly the *transform card. Problem most probably will be there. This is something you can also ask to Creo.
I will try it later with prepomax. There should not be any problem to get the same result.
What causes that support to be so inclined ? Is there a hydraulic cylinder attached to it ? Maybe you could try modeling it (in a simplified way of course - just a cylinder with a hinge attached to the beam - that connection is crucial here) to compare with what you get when *TRANSFORM is used. I would also advise a comparison with one more solver if possible. This may show which solution is the correct one.
model is supported as shown on graph
graph represents loads only in xy plane since values-bending moment from presented beam divided by bending factor (Wx) and ± axial forces (N) divided by cross section area (A) gives us sigma xx on top and bottom flange (item where is big difference) this calculation was made by me - hand calculations and are correct when compare to creo simulate.
I n real conditions are lateral forces which causes bi direction bending and twisting of structure and axial copression/tension
Hi Everyone
I recalculated it again and I see very strange things: reaction force where is only one support in lenght direction( Rax) should have opposite direction thats why rest of result is simply scrap, I have calculated similar construction and the direction and values of Rax were ok as rest of results. Any idea ?
. How did you deduce that ???.
