Quick question about "non linear material effects" (Ansys Workbench)

[kfro]

Hi everyone,

I have a question that I feel I should know the answer to myself, but unfortunately I feel I have a total black-out in the subject...

Well here it goes;

What exactly does "non-linear material effects" (geometry branch settings in ANSYS Workbench) means, and should it be turned ON or OFF if I am making an static structural analysis of a Pod-structure made of aliminium?

I'm using tabular boundary conditions since I have 26 different load cases, but I don't want the different load steps to be dependent of each other. My first thought is that I don't want non-linear behaviour, but since it's not the default and there was such a huge difference in the results, I got confused and I would be sooooo, so greatful if someone could help me out!

Thanx in advance!!!!

 

[hannis]

Hi Kfro,

I've not used ANSYS, but generally non-linear effects would mean the behaviour of the material beyond the elastic (linear)limit, either in the elastic/plastic region where permanent strains are setup e.g as in proof stress. Or even further in to the plastic region where plastic deformation will take place. Which is usual if one needs to now likely fatigue boundaries in a component or structure e.g development of plastic hinges etc. In order to set the parameters for the non-linear effects you will need the plastic mechanical properties of the material.

Personally non-linear FEA makes me nervous. It is generally ok for homogeneous materials such as plastics and material etc where the mechanical properties are tested out of sight and the material is a well defined recipe. But non-linear FEA is being rolled out more for reinforced concrete design which is a quite different material and can be subject to a vast number of variables. In order to get convergence on the stiffness matrix the FEA software has to do a lot more number crunching producing a greater number of iterations. Generally 4 or 6 iterations for linear FEA but possibly up to 80 for non-linear, and this has always seemed to me a less stable analysis.

Regards

Hannis

 

[kfro]

Thanx a lot =)

I'm thinking that since I'm only interested in knowing IF the yield limit is exceeded, and not in knowing what happens afterwords, it should be alright to assume linear behaviour of the material???

 

[hannis]

Agreed, serviceability often limits, since we do not want a structure or component to deflect causing cracks and elastic deformation etc.

Hannis