Dear Baoyun, I too have been looking at this exact application. Building the mesh in deal.ii is difficult. Deal.ii only has some very fundamental shapes which require the use of boolean operations for increased complexity. This in itself is not challenging. The challenge is that the mesh nodes must line up exactly. This really requires a function which moves/merges/places nodes manually (i'm not sure if Deal.ii has such a "healing" function) or else an automatic mesh generator is required. I have, at least temporarily, given up on Deal.ii meshing and started looking at Gmsh as a possible option for automatic mesh generation, but I would rather stay within the deal.ii framework if I can figure it out!
Earl On Tue, Jun 4, 2019 at 7:37 PM Baoyun Ge <[email protected]> wrote: > Hi Daniel > > Thanks for your prompt response and guidelines. After going through these > examples, I have a better understanding now. > > Best, > Baoyun > > On Sunday, June 2, 2019 at 7:54:16 PM UTC-5, Daniel Arndt wrote: >> >> Bayon, >> >> What's the most appropriate/easiest way to setup the matrix for such a >>> multi-domain problem in the dealii programming environment? >>> >> >> Essentially, you are solving a Laplace problem with discontinuous >> coefficients. In case, there are no interface conditions between the >> different materials, >> you should be good to just look at step-6 (and the previous tutorial >> programs of course) to begin with. For periodic boundary consitionsyou >> might want to have a look at step-45. >> >> >>> In an actual electric machine, the air region is very thin, resulting a >>> large aspect ratio. What's the most appropriate way to mesh this region? >>> >> >> We provide generators for a bunch of meshes in the GridGenerator >> namespace. In you case, I would make sure that each materials are not cut >> by cells. >> You might want to merge triangulations via >> GridGenerator::merge_triangulations. With respect to the thin air region >> anisotropic refinement might be interesting (see >> https://www.dealii.org/current/doxygen/deal.II/step_30.html) >> if you can't afford a suitably (isotropically) globally refined mesh. >> >> Best, >> Daniel >> > > On Sunday, June 2, 2019 at 7:54:16 PM UTC-5, Daniel Arndt wrote: >> >> Bayon, >> >> What's the most appropriate/easiest way to setup the matrix for such a >>> multi-domain problem in the dealii programming environment? >>> >> >> Essentially, you are solving a Laplace problem with discontinuous >> coefficients. In case, there are no interface conditions between the >> different materials, >> you should be good to just look at step-6 (and the previous tutorial >> programs of course) to begin with. For periodic boundary consitionsyou >> might want to have a look at step-45. >> >> >>> In an actual electric machine, the air region is very thin, resulting a >>> large aspect ratio. What's the most appropriate way to mesh this region? >>> >> >> We provide generators for a bunch of meshes in the GridGenerator >> namespace. In you case, I would make sure that each materials are not cut >> by cells. >> You might want to merge triangulations via >> GridGenerator::merge_triangulations. With respect to the thin air region >> anisotropic refinement might be interesting (see >> https://www.dealii.org/current/doxygen/deal.II/step_30.html) >> if you can't afford a suitably (isotropically) globally refined mesh. >> >> Best, >> Daniel >> > > On Sunday, June 2, 2019 at 7:54:16 PM UTC-5, Daniel Arndt wrote: >> >> Bayon, >> >> What's the most appropriate/easiest way to setup the matrix for such a >>> multi-domain problem in the dealii programming environment? >>> >> >> Essentially, you are solving a Laplace problem with discontinuous >> coefficients. In case, there are no interface conditions between the >> different materials, >> you should be good to just look at step-6 (and the previous tutorial >> programs of course) to begin with. For periodic boundary consitionsyou >> might want to have a look at step-45. >> >> >>> In an actual electric machine, the air region is very thin, resulting a >>> large aspect ratio. What's the most appropriate way to mesh this region? >>> >> >> We provide generators for a bunch of meshes in the GridGenerator >> namespace. In you case, I would make sure that each materials are not cut >> by cells. >> You might want to merge triangulations via >> GridGenerator::merge_triangulations. With respect to the thin air region >> anisotropic refinement might be interesting (see >> https://www.dealii.org/current/doxygen/deal.II/step_30.html) >> if you can't afford a suitably (isotropically) globally refined mesh. >> >> Best, >> Daniel >> > -- > The deal.II project is located at http://www.dealii.org/ > For mailing list/forum options, see > https://groups.google.com/d/forum/dealii?hl=en > --- > You received this message because you are subscribed to the Google Groups > "deal.II User Group" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To view this discussion on the web visit > https://groups.google.com/d/msgid/dealii/e4082fe9-dcab-46cc-b38e-1adc87808733%40googlegroups.com > <https://groups.google.com/d/msgid/dealii/e4082fe9-dcab-46cc-b38e-1adc87808733%40googlegroups.com?utm_medium=email&utm_source=footer> > . > For more options, visit https://groups.google.com/d/optout. > -- The deal.II project is located at http://www.dealii.org/ For mailing list/forum options, see https://groups.google.com/d/forum/dealii?hl=en --- You received this message because you are subscribed to the Google Groups "deal.II User Group" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/dealii/CAK30MZ565J36dfOrsq872AVefYAPcBo3SNjtq0XpWYMME_uBxA%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
