Open m-greig opened 4 years ago
Hi @m-greig
The behaviour you describe is something I've struggled with myself and unfortunately I haven't been able to resolve it in a good way yet. It is related to this part of the code for the beam goal: https://github.com/CecilieBrandt/K2Engineering/blob/5adf97512821f533316f37667435a6a578be00d6/K2Engineering/K2Engineering/Beam.cs#L112-L127
As you can see, the split into a move vector and weighting is quite arbitrary at the moment but I found that those values made it work most of the time (not ideal and hence why I keep the WIP message for those components). The split has an effect on the convergence speed and relates to the step size for each iteration. You can try to change the split by e.g. changing the multiplication factor on the bending stiffness and see if that solves the problem in your case. I have not exposed this factor to the user to avoid confusion so you will have to fork the project and change it in the code before you build a new .gha.
//Cecilie
Hi Cecile
When using the plugin I noticed that the beam element component seems to cause an issue of non-convergence within in the solver when modifying the section properties. It seems to non-converge when I exceed 200x250x5 RHS. At very high stiffnesses model disappears completely and provides null results. Attached a couple of screenshots.
Not sure if an issue within the Kangaroo2 solver itself. Using Rhino 6 + Kangaroo2. Any help would be appreciated.
Thanks
Testing the Frame 6DOF example
200x250x5 RHS
250x200x5 RHS