fem analysis

[Leonardo.G]

Bye to all,
I would like to ask (simply) if the cat's fem module is internationally recognized as cad/cam and not just as fem. As I'm doing a job abroad, I don't want to have any problems at work.

 

[felperna]

carissimo, the solver is recognized but the + important thing is the approximation (the essay) that uses in the mesh and the constraints... in a few words the finite element model.
the solver is always based on the same mathematics.

 

[Leonardo.G]

thanks felperna,
At least now I'm quieter!

 

[panormus]

carissimo, the solver is recognized but the + important thing is the approximation (the essay) that uses in the mesh and the constraints... in a few words the finite element model.
the solver is always based on the same mathematics.

I'd like to know what you're talking about.
essay, solver.. mhh

 

[www.cad-3d.net]

Take it easy with the chain of cathe... use simulia!

for luca (panormus):

the sagitta is the distance between the face of the model to tetrahedri (so flat) and the surface of the object (which can be curved).

Good evening!

of the noble

 

[Leonardo.G]

during the meshatura it is possible to set the maximum size of the element (dimension) and its accuracy of approximation of the original geometry (shocket), in addition to the type that characterizes it (line 4 knots or parabolic 10 knots).
the sagitta represents the maximum deviation measure of the constituent elements the model from the original geometry.

 

[panormus]

therefore less is the sagitta, the greater the accuracy of the graphic representation. is it a concept that limits itself to video display?

a little like the circle concept consisting of a finite series of segments, the greater the number of segments, the greater approximation to a circle we have, only that you have to deal with the 3d where I have surface element is composed of a tetrahedron: Am I right? I know a lot about infinitesimal analysis.

but why interest in this in the design or production phase? What's wrong?
if a design is parametric should be composed of mathematical functions of which you can have point-by-point or better performance in any local area of the point.

 

[Leonardo.G]

therefore less is the sagitta, the greater the accuracy of the graphic representation. is it a concept that limits itself to video display?

the less the sagitta and the greater the precision of the results of the fem analysis.
to obtain an accurate solution it is necessary that the finished elements have as little dimensions as possible. On the other hand, however, in order to limit the calculation times, the finite elements should not be too small in size. optimal spatial discretization of the domain must be a compromise between two opposite needs: that of an accurate solution and the need to limit calculation times.

 

[panormus]

I learned another one. What time are we talking about?
If I had to do an analysis of the piper structure that I have made, for example, what time do we talk about?
https://picasaweb.google.com/lh/photo/cjzekaaya8mxzlnrh_i1inmtjnzetymypjy0liipfm0?feat=directlinkcan you make a relationship together/detail/time?

 

[Leonardo.G]

What time are we talking about?

we talk about the time it spends since you launch the simulation to when you get the result, in certain simulations it may take even days. therefore is in the skill of the designer the choice of the mesh more appropriate.

 

[www.cad-3d.net]

depends on the processor, the clock frequency of the processor, the ram (and its speed), the number of finite elements (and the type: 4 knots or 10 knots, as mentioned by leonardo) in which the model is "fragmented" and therefore it is a little difficult to predict.
I would say from about twenty seconds of cpu to a maximum minute (i7, 8 gram, 1333 mhz).

Good evening!

of the noble

 

[panormus]

ah from a minute, but also 5 .. are not days. Okay.
but do these calculations give a real idea of the estate of a structure? Are they really reliable? I mean, sometimes the material you use can also have different characteristics from the one on the casket.

 

[PierArg]

ah from a minute, but also 5 .. are not days. Okay.
but do these calculations give a real idea of the estate of a structure? Are they really reliable? I mean, sometimes the material you use can also have different characteristics from the one on the casket.

Hello, luca.
the problem is not the calculations, but how precise do you model what you want to simulate.

Trust that if you want to simulate something like an impact analysis of an entire car, you won't be enough months.
The calculations you refer to are well-known physical laws...and there is so much of that mathematics that you don't need 3 university exams to see it all:biggrin::tongue:

It will sound strange but there are problems where the fem simulation is the only one able to provide directions.
for example some phenomena on the processes of removal of truciole (I know something ) or the study of combustion in combustion engines (for these there are only theories that have never actually been verified).
fluid dynamics then we don't talk about it... or the problems where you want to study a transient.

 

[PierArg]

the fems are born to solve, through a computational calculation, problems that, by their nature, do not have a real known solution.
this depends mainly on the geometry of the piece. at school they always teach us the famous "ideal cases" with very simple geometries.
but what happens when geometry is quite complex? theoretically you have no way to solve the problem.

However, fems do not provide the solution of reality, but only a "simulation" of a physical phenomenon.
fundamentally take the piece you want to study and subdivid it into many elements (your mesh) of geometry (this time) note.
these elements, being known and easy to "study", being connected to each other, allow to apply the simple laws of physics and get to the desired solution.

I hope I explained. I used a simple language to make me understand better.
Unfortunately I am not very good at explaining things

 

[panormus]

It is clear that you have been clear about the removal of the trucile (I would like to say more :-d ), but that ... of work do :-d ? (with all the.

 

[Counterblow Hammer]

depends on the processor, the clock frequency of the processor, the ram (and its speed), the number of finite elements (and the type: 4 knots or 10 knots, as mentioned by leonardo) in which the model is "fragmented" and therefore it is a little difficult to predict.
I would say from about twenty seconds of cpu to a maximum minute (i7, 8 gram, 1333 mhz).

Good evening!

of the noble

5 minutes... oh, my God, it always depends on what you do!
In simulation of sewage processes, for example, even with a 32 core can take full days to complete a simulation. . .

 

[www.cad-3d.net]

I add:

we are talking about stationary geometries (not dependent on time) and in the linear field easily traceable to shell elements and however without noticeable geometric jumps and/or cfd or other interactions. in fact if you want to make calculations of transients (combustion, phenomena of instability etc.) then things change much and linearity is no longer allowed.

of the noble

 

[PierArg]

It is clear that you have been clear about the removal of the trucile (I would like to say more :-d ), but that ... of work do :-d ? (with all the.

hi luca, we'd miss more:finger:

during the university thesis I had to create a fem code for simulation
a particular phenomenon (I do not go into detail otherwise we arrive to tomorrow) that occurs during the processing for removal of truciole.
the only way you have to study it is through the fems and make comparison with experimental evidence to attest their veracity.

 

[panormus]

hi luca, we'd miss more:finger:

during the university thesis I had to create a fem code for simulation
a particular phenomenon (I do not go into detail otherwise we arrive to tomorrow) that occurs during the processing for removal of truciole.
the only way you have to study it is through the fems and make comparison with experimental evidence to attest their veracity.

urca... tesi in informatica?

 

[Leonardo.G]

the only way you have to study it is through the fems and make comparison with experimental evidence to attest their veracity.

I fully agree with pierarg.
virtual prototyping is the most successful attempt to faithfully represent the idea of the designer, in fact it is possible to carry out with appropriate software a wide type of analysis that include, for example, structural analysis, simulation of technological process, simulation of use of the product and many other topologies of problems.