errore fe analysis

[Stan9411]

It's called convergence analysis.. It's a very sensible thing. but on a 2 meter long deal, 11 mm tetrahedri side for me are sufficiently small .. also because reducing them computational cost increases so much

 

[Onda]

in general the analysis of convergence makes sense for stress, but for deformations it is difficult to have variations if not a few percentage points,.
In principle, if there is a big difference with the building, the first thing to check is the assumptions on the constraints.
then it could be that idealizations have been made that make the structure more rigid than it is in reality: continuous welding for points, full penetration or minor welding.
It may be that the elements used are linear and non-square, which give more rigid results, but the difference seems enormous.
can also be that the measurement mode is not correct. very difficult to assess zero, because influenced by the elimination of games. it would be better to try to make different measurements, to different load values, and then linearize everything, you find surprises sometimes.
It is also necessary to ensure that the applied load is correct and how it is evaluated, with which precision. Can I miss the area of piston plate? pressure?
then there are trivial errors, elastic module, model thicknesses, matching geometry, etc.

we, without the information you keep jealously hidden, can not give much help. Keep in mind that any finite element analysis, in the elastic field, if done correctly, must give results at least in the order of 10%.

 

[Cadinpiedi]

can also be that the measurement mode is not correct. very difficult to assess zero, because influenced by the elimination of games. it would be better to try to make different measurements, to different load values, and then linearize everything, you find surprises sometimes.

Right, how was deformation detected? machine mounted or on a granite test bench?

 

[Stan9411]

in general the analysis of convergence makes sense for stress, but for deformations it is difficult to have variations if not a few percentage points,.
in principle, if there is a big difference

the fem solves a matrix of rigidity .. so the direct result is the movements of each knot. from there to dimensionalize the deformation and through elastic bond calculates stress. things are connected so it always makes sense.

then it could be that idealizations have been made that make the structure more rigid than it is in reality: continuous welding for points, full penetration or minor welding.

of the European Union.

It may be that the elements used are linear and non-square, which give more rigid results, but the difference seems enormous.

right but can't make such a difference.

 

[Onda]

the fem solves a matrix of rigidity .. so the direct result is the movements of each knot. from there to dimensionalize the deformation and through elastic bond calculates stress. things are connected so it always makes sense.

this is true, but the peak of tension in a concentration area, which is seen by making a thicker mesh, does not change the total deformity, so the quality level of a mesh necessary to calculate a deformed is generally lower than the quality level necessary to calculate and then converge in the stress zone.
Just try. .

 

[rapa]

the circular surface that in the incastrious analysis, in the reality how is it bound?

si la ralla è imbullonata e compresseime una matassa di filo

 

[meccanicamg]

Is it not that the Chinese ralla rotates of a degree that on a distance so....by millimeters?

 

[Stan9411]

the indignant number one is he give. is always the way?

all other observations are valid but easy to prevent/control (size accuracy, structural defects of the real object) or relatively low (mesh order, mesh size).

We wait for news.

 

[rapa]

update: thanks to everyone for interest
calculation with mesh equal to half sheet thickness and shift changed from 1.6 to 1.8mm
computation with ralla bond on line rather than shift surface changed from 1.6 to 5.4mm !!!
I wait for the software support response

 

[Stan9411]

the assistance will tell you that this is right. it is you who must check why on the real piece melts 9 mm.

 

[meccanicamg]

update: thanks to everyone for interest
calculation with mesh equal to half sheet thickness and shift changed from 1.6 to 1.8mm
computation with ralla bond on line rather than shift surface changed from 1.6 to 5.4mm !!!
I wait for the software support response

You are not binding properly.... there are too many wrong things. as you see you approach the real value.... there is a lot to work, do some tutorials and some basic fem courses.

sheet metal mesh should have at least three elements on the thickness.... otherwise do not read the bending.