FEA in strain analysis

[EngineerDave]

I am a new FEA user. The question I have may be really simple, but here it goes:

If you have the dimensions of an object in the undeformed state and also the dimensions it obtains in the deformed state, how can you back calculate the local strains in ANSYS, Ideas and/or NASTRAN?

Basically I would like to back calculate strain using ANSYS, with only two files, one of the undeformed object dimensions and one of the deformed object.

Does anyone know hot to do this?

Thanks in advance for your help

 

[prex]

Your problem should be solvable by considering the deformed shape as a field of imposed displacements: the code will calculate the associated stress and strain.
The tricky point is that the result will critically depend on the values specified as displacements and how they are distributed. Assuming a 3D model, you would need to specify the displacements in a quite high number of points on outer surfaces and inside, and the displacement field should be coherent to a good precision (say down to the 4th significant figure at least), otherwise the calculated strains will be unrealistic.
Apart from very simple cases (e.g. hydrostatic pressure over a sphere), I think the only way to obtain meaningful results would be in using the results of a normal calculation to feed your reverse procedure: of course the latter would become useless! prex

http://www.xcalcs.com

Online tools for structural design

 

[EngineerDave]

Do you know the logistics of how to do this?

I could for instance note the change in dimension between nodes and use this for the basis of calculation.

I am having trouble finding where in the program ANSYS and or ideas I could accomplish this.

There are Update Geometry commands but I dont' believe they do what I need.

 

[EngineerDave]

Anyone else have some ideas?

I'm trying to force displacments on nodes. One other frustration is that I need to manually enter dimension (x,y,z) data. Seems like there should be a way to import this in a txt or binary file.

I'm getting used to Ansys 6.1! It seems there are some changes.

 

[idly123]

Since u talk of a deformed and undeformed state, first manually compute the approximate loading that would produce this deformation. Having done this, Model ur original undeformed configaration in Ansys, Then perform a geometric nonlinear analysis . For doing this give the load increaments suitably such that u go few load increments beyond the load for which u have deformed configuration.Note : The above holds good provided the material is linearly elastic. In case the material is nonlinearly elastic or if it has an inelestic response then u also have to do simukatneously a material nonlinear analysis along with geometric nonlinearity.

The earlier suggestion of prescribed diplacements as a Multipoint constraint problem works only if u have displacements prescribed at few locations and not the entire deformed configuration. also themultipoint constraint problem does not always lead to correct solutions.
in view of this for strain analysis a rigrous material or geometric nonlinear analysis is essential
hope this gives an insight in to ur problem
regds
Raj

 

[EngineerDave]

I don't have a good estimate of the loading. I am following some previous work that wasn't well documented.

I am currently forcing displacements on nodes, although the material properties aren't clearly known.

We are looking for just the strain calculation. I'm concerned about the accuracy of solving it this way.