You can use CBUSH elements. Create a PBUSH and you will be able to directly use the Force/Displacement diagram you have created under functions.
Go to Nonlinear/Freq Resp / Nonlinear Function Dependence. You can vary the stiffeness for each DOF independently:
where can I find Force vs. Displacement function in Femap? When I use 0..Dimensionless function and write 1 to Stiffness, Femap always calculate with Stiffness value equal to Y value at X=1. In this case Stiffness didn't depend on Force value in my Load Set. This happenned with Analysis type 10..Nonlinear Static.
I would like to use Stiffness from Force/Displacement diagram in my analysis.
I see from your posts that you have figured out that you must use non-linear analysis (not SOL101) in order to use a force-displacement function for a bush.
Note that various Function types will work (Dimensionless, Force vs Frequency, Displacement vs Frequency (!!), Stress vs Frequency. All of these produce the same effect in the Nastran .dat file, writing the required PBUSHT and TABLED1 entries. Note that the X in your function is displacement, and Y in your function is force. However, for you to see non-linear effects, you need to carefully choose the pairs of force and displacement values in the function and calculate intermediate results which traverse gradient changes in the force-displacement function.
I made a model to testing Spring with variable Stiffness. Here is my Spring property:
Here is my function:
I launched a nonlinear analysis:
The resultat was with F=25 N the following:
Displacement of Node 51 was 25 mm, this means, that the Stiffness was 1 N/mm. I tested with lot of forces, but the stiffness was always 1 N/mm, the model didn't use the function.
What's wrong with my model?
The nastran element force sign convention is causing your spring to be in compression, while your force displacement curve is always positive. Notice in your f06 file, the messages about table lookup erros, nastran is trying to look up negative values which do not exist so it is extrapolating on the negative side of your curve which is always 1.
On your spring entry, reverse the order of the nodes.