take a step back and walk us through your calculation, the assumptions it makes and then the setup. feel free to post your model and the calcs using the advanced editor.
the reason i suggest this is i looked at your screenshots, some have 3000N, some have 6000N and i can't tell what you're loading or restraining in your model because they are all deformed models.
my suggestion for this is to fix (on a splitline) on one end, this will assume a super bond to the machine there.
and then on the other end displace the end the distance that it will displace in the test
I attached images showing the fixtures and loads more clearly.
Well using a force of 6000N and using the same value for youngs modulus as solidworks used I worked out that there should be a displacement of about .28 mm but fea in solidworks is giving me .19mm. Now is it as close as I'll get due to assumptions made in hand calculations that arent realistic or is there a better way to fix and load the speciemn to more closely approximate testing in machine.
what assumptions do your hand calcs make?
what do your physical testing results say? (this is the end all be all answer)
what are the assumptions your simulation makes?
a few things i see:
1. you're pulling and restraining only the very ends of the material, is that really how it is held in the machine?
2. simulation is taking into account the change in area, are your calculations taking that into account?