Hi Konstantin, Very very long time not contact with you, how are you doing? I am sorry I really have not too much time that could spend on Deal.II project, the Maxwell equation's problem and so on.
what you propose here I think is right, at least in the view of very beginner of deal.II like me, the documentation in deal.ii is not hard to start with but really hard to understand and implement more advanced techniques like Adaptive mesh, hp refinement, PETsc / Trilinos. Since these techniques are introduced in different problems, very beginners may be expertise in optics but have to learn these techniques from tutorials of fludic-dynamics. Although from mathematics point of view, this is simple. I believe if there are some kind of simple examples implementing these techniques gradually would greatly decrease the learning and thinking time of beginners. By the way, right now I am calculating temperature in MPI and unfortunately stucked. 😂 Best, Mark ------------------ Laboratoire Hubert Curien, UMR CNRS 5516, Bâtiment F 18 Rue du Professeur Benoît Lauras 42000 Saint-Etienne FRANCE 在 2015年9月7日星期一 UTC+2下午11:26:55，Konstantin Ladutenko写道： > > Hi, > > Sorry, it was hard for me to find enough time to get really started with > dell. Few things I was able to learn this summer you can find in this group > and in updated step-6. > > The next possible step for me is to solve Mie problem using deal.ii and > compare against analytic solution. It something similar to H. Fahs > self-written code you can find at http://hfahs.free.fr/software.html (and > related bunch of publications http://hfahs.free.fr/pubs.html ). For sure > I will present it here as soon as I have something to solve the problem of > EM scattering correctly (and probably after that it can be added some > nonlinearity for plasmonic nanoparticles). > > I also found a book of Peter Monk "FEM for Maxwell equations" that is > rather hard for me to read, however, I hope, it will give me answers to all > my basic questions. There is also "Introduction to the Finite Element > Method in Electromagnetics" by > Polycarpou which is much easier to read from my "applied" point of view. > > Best regards, > Kostya > > > пн, 7 сент. 2015 г. в 21:21, JT Hu <jingt...@gmail.com <javascript:>>: > >> Hi Kostya, >> >> It has been a year since you posted this. Do you have any updates on >> developing Deal II methods for plasmonics/photonics calculations to share? >> If you have published any related work discuss the methods and coding, I >> will also be glad to read and cite! >> >> Best, >> Jingtian >> >> >> On Tuesday, September 16, 2014 at 4:44:42 AM UTC-5, Konstantin Ladutenko >> wrote: >> >>> Hi Wolfgang, >>> >>> Thank you for your reply! Deal.II documentation is great, I really like >>> your lectures, I will recommend them to my students. >>> >>> 1) My goal is to correctly solve my problem as fast as possible, time >>> needed to learn new tool included. >>> >>> 2) My choice is open source. We have few Comsol licenses usually not >>> available for students (taken out for 'great and terrible' top science >>> projects), this also leads to some troubles than you apply to use external >>> cluster for computations. There is a number of other open source FEM codes >>> (Elmer, GetFEM++, Moose, etc.), however as to me deal.ii seems to be a >>> mature tool which I am learning with hope to use it many years after that. >>> >>> 3) I suppose that almost any project can benefit from large user base. >>> It seems that deal.ii can solve many problems in different research areas. >>> However average researcher with need of calculations will prefer to use >>> matlab (or octave, scilab), may be some python codes (e.g. with help of >>> sage) and sometimes some ancient Fortran codes - too large and awful to be >>> maintained. GUI programs (Comsol, Ansys etc) are also an option. >>> >>> It is a good way to start from basic elements of deal.ii, as it is done >>> in stepXX tutorials and to go to some real world problems step by step. >>> However you will need few weeks before your can start with your problem. >>> Many possible users of deal.ii do not have so much time to be spend without >>> any practical results. May be I am wrong, but a number of simple applied >>> examples can be used as a kick start entry point to deal.ii. You can use >>> them to copy-paste as building bricks for your exact (may be simplified) >>> problem. It will not be solved in the best way deal.ii can provide, but it >>> will motivate to learn more details about how it should be done in deal.ii >>> way. >>> >>> 4) My current research area of interest is metamaterials (with nonlinear >>> effects coming soon), so Maxwell equations are one the first place. I am >>> mostly interested in transient problems, however after 20 video lectures I >>> am still not sure how i should treat them in deal.ii). This is my entry >>> list of examples for electromagnetism (all of them can be found in almost >>> any good book on the topic): >>> - Simulate dipole source switching on in mirror box (zero boundary >>> condition) until the field reaches corners of the box >>> - Same in free space (PML boundary condition? I suppose this should be >>> done with constrains in deal.ii, still do not know how). >>> - Steady state field of dipole source if free space (may be this should >>> be the first example) >>> - Plane wave source in free space. Gaussian beam in free space. >>> - Transmission and reflection of normal incidence on dielectric >>> interface, thin dielectric plate, multilayer coating. Results should be >>> verified analytically. This can also be used as a high level test case for >>> deal.ii >>> - Get spectra from the Bragg mirror. >>> - Field distribution in Bragg resonator. >>> - Scattering from the sphere (Mie problem). Dielectric material, same >>> with losses. Calculate total scattering, back- and forward- scattering, >>> scattering angle distribution, verify with analytic formulas. >>> - Same with sub-wavelength metal sphere (Drude model of metal, Lorentz >>> model). Get out spectra of plasmonic resonance, verify position of >>> resonance peak against formula. Change the host media. Calculate field >>> amplification inside the sphere. >>> - Non-linear media: temporal and spatial solitions. >>> - Metamaterial cloak with Maxwell equations, as far as Comsol uses >>> particle trace module for some reason >>> http://www.comsol.com/model/ideal-cloak-11721 >>> - etc >>> >>> I can continue with this list, most of the situations are well known and >>> has analytic solutions, so they can be used as a test case for deal.ii. I >>> hope i will be able to contribute some of this examples as I learn deal.ii. >>> >>> >>> Best regards, >>> Kostya >>> >> 2014-09-16 1:36 GMT+04:00 Wolfgang Bangerth <bang...@tamu.edu>: >>> >> >>>> Ladutenko, >>>> >>>> >>>> My problem is that I am overwhelmed with all these details needed to >>>>> use deal.II >>>>> >>>>> I should note that previously I had no problems learning new methods or >>>>> software I want to use, no matter is it in C, C++, Fortran, Python, >>>>> Bash, Lisp etc., using large and small libs in any case, software >>>>> packages with and without GUI and so on. Usually at takes few hours or, >>>>> if my goal is big, few days to learn what I need. However, after >>>>> learning first steps from deal.II tutorial and brief look to some >>>>> others, watching about 15 lectures of Wolfgang I still not really sure >>>>> how to "deal" with my task. Yes, I suppose that deal.II will be rather >>>>> useful for me in long term and I hope I will mange with it one day, >>>>> but... Hey! If you can teach your students to use deal.II than it >>>>> should >>>>> not be a problem to teach them to use emacs (and with help of CEDET you >>>>> can have whole project aware code auto complete, project source >>>>> navigation, refactoring and so on in emacs too, remotely over ssh :) >>>>> >>>>> My point is that there is a need for number of simplified examples, >>>>> which will not use all these advanced features, that deal.II has. I >>>>> suppose, that a regular user do not want to become a FEM guru, he just >>>>> want to solve his problem. And often there is no FEM guru near by... So >>>>> it is important that all of these simplified examples should produce >>>>> some meaningful result from practical point of view. >>>>> >>>>> It is easy to find many hundreds of such examples and application >>>>> notes, >>>>> that come with Comsol, Ansys, CST, Lumerical and so on, ranging from >>>>> simple beam deflection to complete device modeling. Usually it take >>>>> less >>>>> then an hour to follow guidelines and reproduce the result, including >>>>> to >>>>> play around changing this or that... >>>>> >>>> >>>> It all depends on what your goal is. If you want to compute the >>>> deflection of a beam, or want to solve a Laplace equation with a standard >>>> discretization on a mesh you have little control over with a solver you >>>> can't change, then deal.II is not the right tool for you. There are many >>>> good codes that can do that -- Comsol is one, but also Ansys and Abaqus >>>> are >>>> excellent tools if you have an application that fits what they can do. >>>> >>>> The point of deal.II is a different one, though: we provide the tools >>>> if you want to (i) build a FEM solver for problems for which there are not >>>> yet any commercial and well developed codes, or (ii) you are a method >>>> developer and want to play with different discretizations, different >>>> solvers, etc. If these are your goals, then you will need to be able to >>>> control the different parts of your application yourself, and in that case >>>> a library like deal.II is a tool much better suited to the task than is an >>>> application like Comsol. >>>> >>>> >>>> Last but not least - there should be less math and more physics in the >>>>> description, which is intuitively can be understood with applied >>>>> scientist of engineer. Yes, same kind of equation can describe many >>>>> effects in very different topics... But it should be much more easier >>>>> to >>>>> understand what is going on if you put Maxwell equations into deal.II >>>>> instead of writing many math around "u" and "v" for the same problem. >>>>> >>>> >>>> I think that depends on where you come from. In the end, handwaving is >>>> not enough in documentation but one has to be precise to explain what >>>> exactly a function does. This being mathematical software, "intuition" >>>> does >>>> not suffice, but formulas are frequently necessary. >>>> >>>> >>>> All of this said, there is no question that providing adequate >>>> documentation for a tool like deal.II is about the biggest challenge we >>>> have with software this large. This is why there are many layers of >>>> documentation (videos, tutorial programs, documentation modules, >>>> class-level documentation, function-level documentation) that provide you >>>> with different views. We have spent many years writing this documentation >>>> (my guess is that there are 5+ man years of work in the documentation >>>> alone) but it is a process that is never finished and where one can always >>>> (i) write more to explain better, (ii) provide alternative points of entry >>>> to those looking at things from a different direction. Any concrete >>>> suggestions on how to (and, in particular help with) improve the >>>> documentation are always welcome, and we'd be happy to work with you if >>>> you >>>> have ideas! >>>> >>>> Best >>> >>> >>>> Wolfgang >>>> >>>> >>>> -- >>>> ------------------------------------------------------------ >>>> ------------ >>>> Wolfgang Bangerth email: bang...@math.tamu.edu >>>> www: http://www.math.tamu.edu/~ >>>> bangerth/ >>>> >>> >>>> >>>> -- >>>> 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 a topic in >>>> the Google Groups "deal.II User Group" group. >>>> To unsubscribe from this topic, visit https://groups.google.com/d/ >>>> topic/dealii/oNmugeCj9Ho/unsubscribe. >>>> >>> To unsubscribe from this group and all its topics, send an email to >>>> dealii+un...@googlegroups.com. >>>> >>> >>>> 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 a topic in the >> Google Groups "deal.II User Group" group. >> To unsubscribe from this topic, visit >> https://groups.google.com/d/topic/dealii/oNmugeCj9Ho/unsubscribe. >> To unsubscribe from this group and all its topics, send an email to >> dealii+un...@googlegroups.com <javascript:>. >> 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 dealii+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
