I stimulated a section of roller chain under the stress of 100N. The resolver sorted the length of its stretch is 0.00246mm. And then I created another very similar analysis. Set the roller chain is stretched to 0.00246mm. However, the calculation result shows it only need 74Newton pull force to stretch 0.00246mm.
The model of roller chain comes from the tutorial book CD.
The steps of my operation lists as following:
Phase-1
open the roller chain.asm
1. creat a new static study
2. set all parts material AISI 304 (steel)
3. set the connections: 16 sets was seleted and set as no penetration all the rest set as default (global contact bonded)
4. give a boundary condition: soft spring to
5. set the exteral loads and fixture: from this step I tried two settings for comparing. The frist one is both sides of the model was conducted a 2(faces)x400N force. The second one is one side 2(faces)x400N force, the other side was set as a slider fixture 
6. creat mesh: curvature based mesh
7. Until now the final results is not very different from the tutorial.
Phase-2 (symmetry model)
because the model is symmetrical, it was cut into one eighth for time saving before calculation.
1. cut the model into eighth. as the following pictures display
2. give the same meterial and same connection settings
3. fixtures: advanced fixture and select all the symmetrical face (shows in the following picture)
4. then give a 100N pull force on the side of model
and then I created another study for trying to use the displacement to calculate the pulling force.
1. I suppressed the extermal loading (the 100N pulling force).
2. And then through the fixture setting I designated a 0.00246mm displacement on the face was set a 100N pulling force.
Then according to the tutorial, we can get the pulling force by: list result force
but the result force is only 74N it is very different from the tutorial result 117.08N
That is my problem. I don`t know why did it happen? It is because my misoperation or its the solidworks` error?
thanks for your help!
Can you post your model? Assuming everything was set up correctly, it sounds like an issue with SolidWorks. The tutorial's result of ~117 N makes since (in that it is higher than the 100 N applied load from the previous analysis) because you're forcing the entire surface to displace 0.00246 mm.
And thanks for your reply. It remind me that the deformations are different between two studies. That could be the foremost reason for the difference of the results.
But I still feel confused that the stimulation of 0.00246mm displacement studies unexpectedly achieved by only 74N pulling force according to the calculation result by my operation. The tutorial shows its result is 117N. I mean even my result is smaller than 100N, it should be bigger than 100N.
Do you mean to put the model files on here? I`ve tried to, but it seems there is no option for uploading files.
Feizc Zc wrote:
But I still feel confused that the stimulation of 0.00246mm displacement studies unexpectedly achieved by only 74N pulling force according to the calculation result by my operation. The tutorial shows its result is 117N. I mean even my result is smaller than 100N, it should be bigger than 100N.
I agree; it seems to me that calculated reaction load should be higher than 100 N if everything between the two analyses is the same. Speaking of which, you can attached using the following process:
Sorry for the late reply. I just finished my holiday.
Please check the attachement which content the model files.
5. mesh: default mesh settings (curvature based)
6. the result shows the displacement is 0.00246mm
what displacement do you get with a 100N?
Hi Jared, The displacement was 0.00246mm when a 100N pull force conducted on that face.
Have you refined the mesh a few times (in both studies) to show that the stress is convergent? You have 1mm elements to resolve .001mm strain. It's easy to imagine that introducing the amount of error you are seeing.
Its possible everything may be correct, but you just haven't gotten to the answer yet. The force will follow right from the stress, and it is an unfortunate fact that stress (and temp) converge from the bottom.
Just a thought.
Also, is the symmetric study still set to direct sparse? I can remember a time when I was calculating some motion planing thing and there was a huge 30deg error. After flailing to find something wrong with my model, I changed the error tolerance from something like 10^-4 % to something like 10^-6 % and it fixed the problem.
Thanks for your reply. I`ve tird to mesh the model with the finest grids. But it didn`t change a lot. I would chose to continual my tutorial study first. Maybe after I go through the whole tutorial I can find the answer.
Do you still have soft springs turned on?
post the details of your analysis setup
maybe a picture of your model
and your hand calculations
this sounds like a classic example of setup doesn't match calculations