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.

PDE and any physics together

Please login with a confirmed email address before reporting spam

I am trying to add a source in the Heat Transfer (HT) module by applying a formulation which is a function of temperature. I add a PDE coefficient form. Then I enter "T" for the temperature expecting to take the temperature from HT module. The source term is a rate and it is in derivative form (only time derivative). Since it has only time derivative, an initial cond. is enough I thing. When I solve the program, I see the results is not convenient with exact solution that I made by hand. For example, When I plot the figures of each variable (T and source term) I see they are different. I should also note that HT module asks unit for source in W/m^3 but I have no any unit in PDE form. I neglect the red colour unit error. Thanks...

9 Replies Last Post 4 gen 2011, 05:42 GMT-5

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 3 gen 2011, 06:21 GMT-5
I calculate the formulation in PDE form by taking the variables from plots for certain node and certain time and formulation does not give a equivelance.
I calculate the formulation in PDE form by taking the variables from plots for certain node and certain time and formulation does not give a equivelance.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 3 gen 2011, 06:25 GMT-5
The unit is important or not? If important, How can I enter a unit for a variable in PDE form?
The unit is important or not? If important, How can I enter a unit for a variable in PDE form?

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 3 gen 2011, 07:22 GMT-5
Hi

Units are important for us users to check that we are putting up correctly the physics, by the way "unitless" or unit [1] is just as imporant, it's just a particular case. But COMSOL only Warns you about unit errors, so it will calculate with the values you give, assuming they are consistent in any case, so it's you the programmer/end use's responsability to know what you/we are doing ;)

It iscorrect that even in v4.1 units int he global expressions or user defined PDE's are not fully implemented so there you must be carefull to add up consistent elments. I mostly multiply my variables by

MyPDE_variable*1[my_desired_units]

in the equation to help stay consistent

--
Good luck
Ivar
Hi Units are important for us users to check that we are putting up correctly the physics, by the way "unitless" or unit [1] is just as imporant, it's just a particular case. But COMSOL only Warns you about unit errors, so it will calculate with the values you give, assuming they are consistent in any case, so it's you the programmer/end use's responsability to know what you/we are doing ;) It iscorrect that even in v4.1 units int he global expressions or user defined PDE's are not fully implemented so there you must be carefull to add up consistent elments. I mostly multiply my variables by MyPDE_variable*1[my_desired_units] in the equation to help stay consistent -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 3 gen 2011, 08:41 GMT-5
MyPDE_variable*1[my_desired_units] did not work.

Also I recognized that all boundary conditions should Neuman type, not Drichlet, although any BC is not required for a time derivative only. Even that case results are sometimes so different from exact values, I could not understand why?
MyPDE_variable*1[my_desired_units] did not work. Also I recognized that all boundary conditions should Neuman type, not Drichlet, although any BC is not required for a time derivative only. Even that case results are sometimes so different from exact values, I could not understand why?

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 3 gen 2011, 08:49 GMT-5
I named the variable R for PDE and I wrote R*1[W/m^3] in source term but it does not work. I also tried it in global expressions but it also did not work. The other desadvantage of writing the expression in global expressions, I cannot plot it at global variables plot --> predefined quantities in postprocessing, It writes ... is not a valid global expression.
I named the variable R for PDE and I wrote R*1[W/m^3] in source term but it does not work. I also tried it in global expressions but it also did not work. The other desadvantage of writing the expression in global expressions, I cannot plot it at global variables plot --> predefined quantities in postprocessing, It writes ... is not a valid global expression.

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 3 gen 2011, 09:32 GMT-5
Hi

I should have been claeare ;)
you cannot define the variable with the units in COMSOL (in the name field),
but I use the variables in the equations by multiplying them by *1[my_units]

Sorry for that (note this will certainly change in v4.2 or 4.3)

--
Good luck
Ivar
Hi I should have been claeare ;) you cannot define the variable with the units in COMSOL (in the name field), but I use the variables in the equations by multiplying them by *1[my_units] Sorry for that (note this will certainly change in v4.2 or 4.3) -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 3 gen 2011, 14:46 GMT-5
Dear Ivar

What about the errors in calculating a simple example such as u=T^2 where u is the variable for PDE and T is temperature from HT module. In this example, the results are not same with the exact soln. mostly. I chose the Nueman condition for each boundary.
Dear Ivar What about the errors in calculating a simple example such as u=T^2 where u is the variable for PDE and T is temperature from HT module. In this example, the results are not same with the exact soln. mostly. I chose the Nueman condition for each boundary.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 3 gen 2011, 17:20 GMT-5
I do not understand why the boundary conditions should be Nueman type for this manner (for only time derivative eqn. in order to put the source or reaction rate into the HT or Mass physics) . If I use spatial derivatives in PDE form, I should enter a boundary condition for that derivatives, rigth?
I do not understand why the boundary conditions should be Nueman type for this manner (for only time derivative eqn. in order to put the source or reaction rate into the HT or Mass physics) . If I use spatial derivatives in PDE form, I should enter a boundary condition for that derivatives, rigth?

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 4 gen 2011, 05:42 GMT-5
Hi

what about uploading your "simple model" it's easier for us out here to talk about it then, or explain step by step the way you build it. Such exercices also often helps to find the solution ;)

--
Good luck
Ivar
Hi what about uploading your "simple model" it's easier for us out here to talk about it then, or explain step by step the way you build it. Such exercices also often helps to find the solution ;) -- Good luck 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.