I am still working on this problem *push*
From what I understand, a virtual wall would be adequate to use on the bottom supporting surface to simplify the simulation and still provide the relative constraints for the lower portion of the model and not allow penetration. For the reasons you mentioned above, you would require some form of second body/surface to contact the wave spring and compress it as the contact area the force is applied over varies as the spring contorts and twists out of plane when compressed. If you're running a static analysis you have to remember the loads are considered to be applied slowly and remain constant in there direction of application.
To allow some twist out of plane, you could could try splitting the top face of the spring and applying point displacements/forces at three points that coincide with where the 'peaks' of your wave spring are. Or you could modify the geometry of the top body so it is only in contact in the necessary area of the spring. However this then defeats the purpose of having the spring 'sandwiched' between two planar faces and being held there as the planar faces would limit the 'out of plane' motion when contact occurred between the two - so I'm not sure if this works for you.
Personally I think the setup you have works well enough and only if you are getting excessive displacements and strains/materials are non-linear would I consider switching to a NL analysis to tackle this problem as the NL solver can re-evaluate the loading direction (can enable this in the properties) and stiffness of the model at each iteration.
Looks okay to me!
even your setup and simulation is successful, there is a high chanse that it wont give the same result as a real life measure (it shows far less reaction force):
Almost all of these spring-washer-thingies (but also the ones made from wire) after the heat threatment have a process called "Scragging" :
During this process the part will be squeezd (sometimes until it will be absolutely flat). That couses plastic deformation in the material and completely changes the load/displacement curve of the Spring.
we measured the real spring and the force was (as you mentioned) far below our expected reaction force.
Is there any way to take the scragging process into account?
In the meanwhile i found an interesting paper which compares the analytical (linear) calculation with FEM results (Ansys) of a wave spring.
So i rebuilt the wave spring (Step file attached) and tried to get the same results with SW Simulation -> but i failed.
Are you guys getting the correct results? Would be great if someone would also evaluate the geometry.
E = 197 GPa
Poisson = 0,282
WaveSpringN5.STEP.zip 243.5 KB
the free heigh in your model is 5,1 (between the flattered faces) and the paper says it should be 4 mm
I changed the height to 4 mm and re-uploaded it!