Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Dimensionless 2-dimensional topology optimization
Posted 18 apr 2023, 03:37 GMT-4 Computational Fluid Dynamics (CFD), Heat Transfer, Optimization Version 6.1 3 Replies
Please login with a confirmed email address before reporting spam
Hi all We are currently working on 2D and 3D topology optimization of natural convection. The first common problems encountered are: 1. warnings that the orthogonal solution fails and requests a reduction in the step size; 2. errors that the solution cannot be found, the maximum number of Newtonian iterations is reached, the returned solution does not converge and all parameter steps are not returned. Finally, how should the dimensionless parameter Gr on buoyancy be entered into the model? Can anyone help me with this problem?
Thank best regards.
Please login with a confirmed email address before reporting spam
Hi
You can use move limits or more conservative settings for the nonlinear solver to reduce the risk that this fails. It is not clear from your question, whether you are solving the stationary or the transient case. If you are solving the transient case, you can increase the robustness of the adjoint computation by forcing the solver to use the same time steps going forward. It might also be beneficial to use a fixed time step, see tesla_microvalve_transient_optimization in the application gallery for an example of transient topology optimization.
If you need help with a specific model, I suggest contacting comsol.com/support.
Best regards, Kristian E. Jensen Technical Product Manager, Optimization
Please login with a confirmed email address before reporting spam
Hi
You can use move limits or more conservative settings for the nonlinear solver to reduce the risk that this fails. It is not clear from your question, whether you are solving the stationary or the transient case. If you are solving the transient case, you can increase the robustness of the adjoint computation by forcing the solver to use the same time steps going forward. It might also be beneficial to use a fixed time step, see tesla_microvalve_transient_optimization in the application gallery for an example of transient topology optimization.
If you need help with a specific model, I suggest contacting comsol.com/support.
Best regards, Kristian E. Jensen Technical Product Manager, Optimization
hi Thank you for your reply. Unfortunately, I am using a steady-state solver. Where should I improve on the errors that occur in this case? Does Comsol have examples and instructional videos related to convective topology optimization in 2D and 3D, and can you provide them? Best regards.
Please login with a confirmed email address before reporting spam
The microreactor_optimization application library model demonstrates steady-state topology optimization for a convective problem.