Convergence issues with electromagnetic simulations based on the specific input parameters

Please login with a confirmed email address before reporting spam

Hi everyone,

I'm performing some simulations with the efwd module, to see the effects of nonlinear phenomena on the optical properties of infinitely periodic electromagnetic scatterers, but I'm encountering issues with the convergence of the solver under specific conditions. Unfortunately, for a matter of secrecy I cannot upload the models or some screenshots, but I'll try to be the most precise possible in describing the situation.

The workflow I'm trying to implement is the following: the model has two components, one that does not consider the nonlinear effects on the refractive index (let's call it "unperturbed"), and another for the nonlinear phenomena (let's call it "perturbed"). The second component must use the solutions of the first component (more precisely the electric field) to modify the refractive index of the materials, and then compute basically the same physical quantities of the first component. The key point is that this "perturbation" depends, of course, on the intensity of the incident light, which is one of the input parameters of the model.

The model I've built in this way works perfectly fine, as long that the incident intensity is not too high. Above a certain threshold, however, the model stops working, and it gives the following error: "Maximum number of Newton iterations reached. Returned solution is not converged. Not all parameter steps returned". After reading several articles and forum posts, I've tried many alternative approaches (including finer meshes, greater distances of the two ports from the scatterer, load ramping), but none of these actually worked. The last approach I've tried, which seemed to work at first, is to change the kind of stationary solver: the default setting uses a Fully Coupled solver, but now I've tried setting up a Segregated Solver with two different steps, one for each of the components (I followed the instructions reported here: https://www.comsol.com/support/knowledgebase/1258). With this new approach, the model now runs also for higher incident intensities, but it gives me a result that physically makes no sense at all.

Does anybody have any idea of what's going on? Is this "segregated approach" way the right one, or am I missing something else? And if it is correct, why isn't it giving me the correct result?

Thank you in advance,

Pietro


Reply

Please read the discussion forum rules before posting.

Please log in to post a reply.

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.