3 Replies Latest reply on Nov 14, 2016 4:48 PM by Gian Flavio Violi

    instable model failure

    Maurits De Wolff



      I've been struggling with a simple bicycle crankset. The goal is to run a simulation, evaluate the outcome and adapt parts or materials were needed so as to meet the proper criteria. All other students have finished their assignment and submitted a report but I can't seem to finish the simulation in the first place.


      At first the solver was at it for 5 or 6 hours. It didn't fail at the time, it just never finished the job. Then I found I had some interference issues. I solved those and I speeded up Windows 7 by unchecking all unuseful fancy stuff. There's several Solidworks instuctions floating around on the internet on how to tweak swap space, and Windows settings in general. I also tried alternating between Direct Sparse and FFE Plus. Now the solver fails within the first 5 minutes due to instability of the model. All students modeled their crankset to the same dimensions and nobody I heard of haviing such failures.


      What causes these failures and how can I prevent this from happening? I've read the go-engineer link but I do not comprehend and this point.


      Thank you for your time.



        • Re: instable model failure
          Gian Flavio Violi



          I reviewed your model and I'd be happy to provide some pointers so you can get your sim up and running. Am I correct to asume you are simulating a bike pedal/lever?


          1- I'd redo all you contacts. Use contact sets instead of component sets for the faces that are transfering forces between the spies-trapas-crank. Also, don't use no penetration for global contact, it just drains to much computer resources without actually adding value to your simulation, specially when you do the proper contact sets to override the bonded global contact.

          2- Check if the solver you are using is compatible with your contacts (and simulation set-up in general).

          3- Re-check the restraints... you could do a little better on that regard.





          Update: review your geometry/material. Large displacement were needed to run the model with forces far bellow critical cases (I've actually performed this sim before). My set-up has already been validated experimentally, so I'm sure what fixtures/contacts/etc I used in the full sim.