gnikit
commented
3 years ago I think we (FETCH) does exactly that with radiation particle energies, when we want to adapt across multiple particle energies. Unfortunately, our solution space is 5-7D which means that some of the higher level functionality with state and scalar/vector/tensor fields does not work the same as in Fluidity and a lot of the underlying Fluidity code has been changed to fit our needs.
Do you want to upload a test case with which I can play a bit and get a better understanding of what it is that we need to change?
Would it be possible to have the adaptivity interpolation error bound fields depend on the values of other fields? Currently, these fields can be set to depend on time and space, but I believe it could be useful to have them depend on other fields. As an example, I have a Stokes flow simulation inside which multiple thermal instabilities with different intensities develop. They are all of importance and need to be properly resolved. However, either I am satisfyingly resolving the most intense instabilities and missing on the smaller ones, or I am resolving everything but the most intense instabilities are absurdly over-refined, dramatically increasing the time spent in the solvers. I am potentially missing something, but I believe that if I were to be able to define the interpolation error bound fields based on the value of, say, the temperature field, that could easily solve this kind of problem. Happy to hear about if this is possible, or if there are other ways to deal with such a situation.