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.

why COMSOL is so difficult?

Please login with a confirmed email address before reporting spam

I'm really angry with COMSOL! 6 months ago I heard about COMSOL in Spectrum IEEE, i told myself: COMSOL is the best software that would help me in my thesis, it seems simple, its user friend and... then i started to learn this software, but i found that its documents cant really help, every time i started to design a new model i encountered a lot of problems! in documents you can not find many things about basic concepts of this software! what is moving mesh ALE! ( not it's mathematical concepts but about how it works in COMSOL, how you can use it in a general form...) ? how many boundary conditions you can use and what is the meaning of of them!? how you can calculate force! charge!!! and many other concepts that you can't find them in texts and if you find you can't understand a clear relationship between them and the way comsol use them!
by using Model library document you just follow a procedure without any understanding of what really is happening! like a blind!
I don't know these problems are only my problems or not! I'm really disappointed about learning this software!
may you guide me? what should i do?

9 Replies Last Post 25 ago 2010, 09:37 GMT-4

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 22 ago 2010, 05:08 GMT-4
Well, I have to admit that I passed through most of those paths you mentioned, and still struggling with some new ones.

I think it is more than the software; it is about the approach that is complex. I was a confused at the start when comsol solved a laminar flow in an eye blink, against to all the extensive courses I had in bachelor about the flow. Nevertheless, the physics itself is still very complex, not to mention when you want to couple them with each other. The solver, the engine, has many options and configurations. I don't think a software can really have so extensive help files anyway. It is the multiphysics that makes working with comsol so damn confusing. e.g. recently I learned the hardway that a default checkbox in a physics can cost me 3 months waste of time (and experience of course). But it is not Comsol's fault, it was me who didn't understand what it exactly means. As my prof. says, when you are on the frontline, it is the price you pay.

I am trying to read a lot and stay on these forums as much as possible. You can really learn some tricks and understand the concepts from classic books. There are not very much textbooks about ALE or FSI, that's true, but there are many nice papers that each tells something about boundary conditions or special usage of ALE. The authors are usually kind to reply a few line back when I asked them a clear question too.

Sorry for the long text, but I wanted you to know you're not alone in that boat.

Best (va ba arezuye movaffaghiyat),
Danial
Well, I have to admit that I passed through most of those paths you mentioned, and still struggling with some new ones. I think it is more than the software; it is about the approach that is complex. I was a confused at the start when comsol solved a laminar flow in an eye blink, against to all the extensive courses I had in bachelor about the flow. Nevertheless, the physics itself is still very complex, not to mention when you want to couple them with each other. The solver, the engine, has many options and configurations. I don't think a software can really have so extensive help files anyway. It is the multiphysics that makes working with comsol so damn confusing. e.g. recently I learned the hardway that a default checkbox in a physics can cost me 3 months waste of time (and experience of course). But it is not Comsol's fault, it was me who didn't understand what it exactly means. As my prof. says, when you are on the frontline, it is the price you pay. I am trying to read a lot and stay on these forums as much as possible. You can really learn some tricks and understand the concepts from classic books. There are not very much textbooks about ALE or FSI, that's true, but there are many nice papers that each tells something about boundary conditions or special usage of ALE. The authors are usually kind to reply a few line back when I asked them a clear question too. Sorry for the long text, but I wanted you to know you're not alone in that boat. Best (va ba arezuye movaffaghiyat), Danial

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 22 ago 2010, 11:18 GMT-4
"There's no royal way" said Euclides to Alexander the Great when he asked him once to teach him geometry quickly. :-)

cheers
Andras
"There's no royal way" said Euclides to Alexander the Great when he asked him once to teach him geometry quickly. :-) cheers Andras

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 22 ago 2010, 12:12 GMT-4
Dear Danial
salam
really thank you for your attention and kindness, I understand most of you say, Comsol in some ways is like Microsoft Visual Studio .Net! you may be really expert in programming but you can't work with the software itself! i found some books and ebooks which can be helpful in Multiphysics modeling and finite element analysis learning. i think i have to read a lot of books and papers to become expert in this software. although i believe Comsol documents could be more better than these documents, or at least developers could introduce some useful source for beginners and students!!.
glad to meet you here Danial. (va ba arezuye movafaghiat vase shoma, khoshal misham bishtar ashna beshim...)
Dear Danial salam really thank you for your attention and kindness, I understand most of you say, Comsol in some ways is like Microsoft Visual Studio .Net! you may be really expert in programming but you can't work with the software itself! i found some books and ebooks which can be helpful in Multiphysics modeling and finite element analysis learning. i think i have to read a lot of books and papers to become expert in this software. although i believe Comsol documents could be more better than these documents, or at least developers could introduce some useful source for beginners and students!!. glad to meet you here Danial. (va ba arezuye movafaghiat vase shoma, khoshal misham bishtar ashna beshim...)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 22 ago 2010, 12:20 GMT-4
Thank you Andras, it's true! there is no Royal way toward COMSOL. :)
Thank you Andras, it's true! there is no Royal way toward COMSOL. :)

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 22 ago 2010, 14:13 GMT-4
Hi

I'm afraid that I can only confirm: it takes time (I counted about mosts nights of 2 years to get comfortable with COMSOL), and still there are many traps I end up in.
I agree that the doc could be different, but now in the state of starting to teach Comsol to others I end up in saying how to rewrite it ? Without a clear answer.
You need to learn the methodology of FEM, the art of meshing, the maths of PDE and BVP but still you need the knowledge of all the physics from the different fields and this is the first time I see so many physics in the same tool, each of us have very different habits, default units, norms, all must be unified somehow to fit. The result is that nobody really finds his usual way. Normally you need a few years at any university to cover all that, then to unify it in one tool, coherently ...

Sure the V4 revolution is not making things easier for us, but that one I appreciate overall, it makes the programme more streamlined against the physics, even if I feel I still do not master the solver section yet.

But I see that several academics are not happy with the units of COMSOL, they want the "unitless" version. As engineer first of all, I love the units, having burnt my finger too often on unit errors (of myself and others), not to say when dealing with the "imperial units" people ;)

An advice, try to get onto a few of the COMSOL courses, they help for a good kick-start

Finally, I would say COMSOL is not that difficult, but you must start "simple", it's our true-life models that are in fact astonishingly complex, when all comes to all

--
Good luck
Ivar
Hi I'm afraid that I can only confirm: it takes time (I counted about mosts nights of 2 years to get comfortable with COMSOL), and still there are many traps I end up in. I agree that the doc could be different, but now in the state of starting to teach Comsol to others I end up in saying how to rewrite it ? Without a clear answer. You need to learn the methodology of FEM, the art of meshing, the maths of PDE and BVP but still you need the knowledge of all the physics from the different fields and this is the first time I see so many physics in the same tool, each of us have very different habits, default units, norms, all must be unified somehow to fit. The result is that nobody really finds his usual way. Normally you need a few years at any university to cover all that, then to unify it in one tool, coherently ... Sure the V4 revolution is not making things easier for us, but that one I appreciate overall, it makes the programme more streamlined against the physics, even if I feel I still do not master the solver section yet. But I see that several academics are not happy with the units of COMSOL, they want the "unitless" version. As engineer first of all, I love the units, having burnt my finger too often on unit errors (of myself and others), not to say when dealing with the "imperial units" people ;) An advice, try to get onto a few of the COMSOL courses, they help for a good kick-start Finally, I would say COMSOL is not that difficult, but you must start "simple", it's our true-life models that are in fact astonishingly complex, when all comes to all -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 23 ago 2010, 14:27 GMT-4
Dear Ivar
Hi. I read your Comment 3 times, really thank you for help and attention and your useful guidelines. fortunately I'm student of mechatronic and i have to learn knowledge of all the physics from the different fields. when it takes 2 years for you to become expert in COMSOL, it must take more time for me to learn this software. of course i believe COMSOL is easier and more user friend and more general than ANSYS and other FEM softwares. and I prefer COMSOL to any other software. unfortunately there is no course for this software here in our country and we have to learn this software by ourselves. when i become expert in this software i will start to write a book for this software ;-)
thank you again...
Dear Ivar Hi. I read your Comment 3 times, really thank you for help and attention and your useful guidelines. fortunately I'm student of mechatronic and i have to learn knowledge of all the physics from the different fields. when it takes 2 years for you to become expert in COMSOL, it must take more time for me to learn this software. of course i believe COMSOL is easier and more user friend and more general than ANSYS and other FEM softwares. and I prefer COMSOL to any other software. unfortunately there is no course for this software here in our country and we have to learn this software by ourselves. when i become expert in this software i will start to write a book for this software ;-) thank you again...

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 23 ago 2010, 15:38 GMT-4
Hi

Good to hear that you take it positively, and that you see the advantage of playing with physics and with COMSOL (thats also why I'm fan of it). Writing a book could be a good idea, but it takes time, documenting well the steps you are going through helps too. Do not forget the different presentations on the web site, and the webcasts from time to time, these are 50% marketing but still 50% technical and gives you some interesting hints now and then.

Concerning COMSOL courses, well I would say try to find other users in your country / region and propose a course split betwen you all, I'm sure COMSOL will consider it if there is enough people, discuss it with your local rep and perhaps a local university or company(ies) ? It's worth a try, no ?

Then take the 3.5a doc (or the 4.0a) and go systematically through the examples, and why not try out some of the examples on the Forum, I have lerned a lot by searching for explanations to others here.
Mechatronics, means (for me) first of all structural, and this is well documented in books and other software, so you have matter to compare. Then often some ACDC magnetism or electrodynamics, perhaps PZT, but do not attack all in one go, try it out one by one, then mix them. Fluidics I'll say at last (depends on your previous knowledge) but that is only because I believe it requires some further brain stretching to cath it correctly, and model convergence is trickier

--
In any case Good luck, and have fun Comsoling
Ivar
Hi Good to hear that you take it positively, and that you see the advantage of playing with physics and with COMSOL (thats also why I'm fan of it). Writing a book could be a good idea, but it takes time, documenting well the steps you are going through helps too. Do not forget the different presentations on the web site, and the webcasts from time to time, these are 50% marketing but still 50% technical and gives you some interesting hints now and then. Concerning COMSOL courses, well I would say try to find other users in your country / region and propose a course split betwen you all, I'm sure COMSOL will consider it if there is enough people, discuss it with your local rep and perhaps a local university or company(ies) ? It's worth a try, no ? Then take the 3.5a doc (or the 4.0a) and go systematically through the examples, and why not try out some of the examples on the Forum, I have lerned a lot by searching for explanations to others here. Mechatronics, means (for me) first of all structural, and this is well documented in books and other software, so you have matter to compare. Then often some ACDC magnetism or electrodynamics, perhaps PZT, but do not attack all in one go, try it out one by one, then mix them. Fluidics I'll say at last (depends on your previous knowledge) but that is only because I believe it requires some further brain stretching to cath it correctly, and model convergence is trickier -- In any case Good luck, and have fun Comsoling Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 25 ago 2010, 07:48 GMT-4
Hello Ivar,

I am following your inputs to the discussion forum and find them very helpful, mostly because they do not highly depend on theory but more on the application and methodology.


I agree that the doc could be different, but now in the state of starting to teach Comsol to others I end up in saying how to rewrite it ? Without a clear answer.


I am refreshing myself with my C coding skills and cleaned the dust off from my old copy of "The C Programming Language" by Brian Kernighan and Dennis Ritchie. In that book the authors say that, the only way to learn a new programming language is by writing programs in it, on page 5 before giving their famous "hello, world" example.

I see no difference here. There can be heaps of theory of Finite Element Method and Comsol can reveal as many things as they want about the "behind the curtain" of the software in their documentation but I know from my teaching experience in the past that, giving one example is far more better than tens of pages of describing how something can be done.

I also paid attention that, most of the questions here is about how to describe a physical problem to Comsol. These questions are not coming from the theory, they are coming from user-software interaction. Therefore for your question above, my answer would be, examples, examples and more examples.

Regards,
Onur
Hello Ivar, I am following your inputs to the discussion forum and find them very helpful, mostly because they do not highly depend on theory but more on the application and methodology. [QUOTE] I agree that the doc could be different, but now in the state of starting to teach Comsol to others I end up in saying how to rewrite it ? Without a clear answer. [/QUOTE] I am refreshing myself with my C coding skills and cleaned the dust off from my old copy of "The C Programming Language" by Brian Kernighan and Dennis Ritchie. In that book the authors say that, the only way to learn a new programming language is by writing programs in it, on page 5 before giving their famous "hello, world" example. I see no difference here. There can be heaps of theory of Finite Element Method and Comsol can reveal as many things as they want about the "behind the curtain" of the software in their documentation but I know from my teaching experience in the past that, giving one example is far more better than tens of pages of describing how something can be done. I also paid attention that, most of the questions here is about how to describe a physical problem to Comsol. These questions are not coming from the theory, they are coming from user-software interaction. Therefore for your question above, my answer would be, examples, examples and more examples. Regards, Onur

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 25 ago 2010, 09:37 GMT-4
Hi

for the examples, I fully agree wit you.
That is why I, in fact, I pass quite som time here on the Forum.

The only thing is that today, nobody has time ... or will give us any time.
So I end up stealing some time, here and there, to play with phsysics, and Comsol ;)

--
Have fun Comsoling
Ivar
Hi for the examples, I fully agree wit you. That is why I, in fact, I pass quite som time here on the Forum. The only thing is that today, nobody has time ... or will give us any time. So I end up stealing some time, here and there, to play with phsysics, and Comsol ;) -- Have fun Comsoling Ivar

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.