Dear Colleagues,
The short question I am having is: can I restart an Abaqus simulation that failed previously with modified step parameters? For example, at the end of the last simulation, the time increment was 0.2. I want Abaqus to continue this simulation with a time step of 0.01 instead of 0.2. If that is possible, please kindly show me the command I should use.
Here are some details of my problem for anyone interested.
As I try to overcome some convergence issues in my Abaqus/Standard model, I noticed a very interesting phenomenon:
The model encounters convergence issues at a time point. I modified the step to let it stop at that point and run restarts from there. At this point, the model is having difficulty converging even with a tiny time increment of 1e-11. However, with a much larger time increment of 0.01, the model can converge and continue. Based on my understanding, this could be that since the model is non-linear, the "initial guess" has to be very close to a solution for the Newton method to work. However, Abaqus automatically cuts back the time increment size whenever it is having a convergence issue, which works against my favor in this case. Thus, I feel I need to adjust the time increment manually and gradually drive the simulation forward.
1) yes... restart is same as adding new step in the analysis
2) you got it right. initial guess is very important for newton's method to converge
3) if the analysis is not able to converge with 1e-11 inc size...it is mostly genuine... check your model again... and simplify or change analysis type if possible based on available results as it'll take forever to run an implicit solver if it's working with extremely small increment size
- Badges
- Science topic