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.

Solving coupled PDE's in COMSOL 5.2

Please login with a confirmed email address before reporting spam

Hi!

Is there any way to solve coupled PDE's without using the Coefficient form PDE? I am not able to fit in my equations using that module.
Also, How can i custom add boundary conditions(I am not able to add my boundary conditions using the inbuilt formats)?
I have attached few equations for example.
It will be very helpful if someone can provide step by step procedure roughly.

Thanks in advance!


5 Replies Last Post 10 giu 2016, 01:35 GMT-4

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 3 giu 2016, 08:10 GMT-4
Hi

I see a diffusion problem (Transport of diluted species) in spherical symmetry. The 2nd equation is No flux boundary condition, but if r = 0 the boundary is reduced to a single point - not very physical. The 3rd equation is Flux boundary condition - you may write anything you like to J (flux).

You do not need to convert the (x,y,z) coordinates to sperical ones, just draw axisymmetric domains, and make, e.g. a radially symmetrical mapped mesh.

Wish this helps

BR
Lasse
Hi I see a diffusion problem (Transport of diluted species) in spherical symmetry. The 2nd equation is No flux boundary condition, but if r = 0 the boundary is reduced to a single point - not very physical. The 3rd equation is Flux boundary condition - you may write anything you like to J (flux). You do not need to convert the (x,y,z) coordinates to sperical ones, just draw axisymmetric domains, and make, e.g. a radially symmetrical mapped mesh. Wish this helps BR Lasse

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 6 giu 2016, 03:59 GMT-4
Hi Lasse!

Thanks for your reply.

I want to use mathematical model instead of Inbuilt Transport in diluted species module. I have few custom equations more that are not present in the module.

Is there anyway that I can add custom boundary conditions atleast, I will modify my equation to fit in the Coefficient form PDE.

Thanks in advance!
Hi Lasse! Thanks for your reply. I want to use mathematical model instead of Inbuilt Transport in diluted species module. I have few custom equations more that are not present in the module. Is there anyway that I can add custom boundary conditions atleast, I will modify my equation to fit in the Coefficient form PDE. Thanks in advance!

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 9 giu 2016, 05:18 GMT-4
Hi Lasse!

I have been trying lot of options but most of them have been of less use.
Can you help me out in inserting this equation(attached) in COMSOL in coefficient PDE form.
Here phi1,phi2,Rsei,iloc,i2 all are variables and functions of (x,t).
phi1 and phi2 are different potential functions for two electrodes and electrolyte.
iloc and i2 are current densities.
Rsei is the resistance of a layer.

I am not able to decide which variable should be selected as dependent variables, as they have different boundary conditions interlinked with each other.

Thanks in advance!
Hi Lasse! I have been trying lot of options but most of them have been of less use. Can you help me out in inserting this equation(attached) in COMSOL in coefficient PDE form. Here phi1,phi2,Rsei,iloc,i2 all are variables and functions of (x,t). phi1 and phi2 are different potential functions for two electrodes and electrolyte. iloc and i2 are current densities. Rsei is the resistance of a layer. I am not able to decide which variable should be selected as dependent variables, as they have different boundary conditions interlinked with each other. Thanks in advance!


Please login with a confirmed email address before reporting spam

Posted: 8 years ago 9 giu 2016, 15:12 GMT-4
Hi

My guess is that U is the total voltage of the cell and you are subtracting the electrode potentials fii_1 and fii_2 from it, as well as the ohmic drop R_sei*i_loc. That leaves you the Galvani potential in the solution (?) phase; a_s is the specific area and C the specific capacitance (F/cm^2)?

The dimensions do not quite add up because (d i_2/dx) is in A/cm^3, a_s*i_loc is in A and the time derivative also in A. But that is only a technical detail, more serious is the term (d i_2/dx) because if it deviates from zero there is a current source, i.e. excess charge in the system; in an electroneutral system the divergence of current is always zero.

You need as many PDE physics nodes as you have variables as well as pertinent boundary conditions. The equation you sent does not make very much sense to me at least. Then you couple the equations in an appropriate way which I cannot possibly tell you, but I have a feeling that fii_1 and fii_2 cannot be free variables but boundary conditions of your problem. Currents neither are variables because they must be constant across the system (unless you have, indeed, a current source).

BR
Lasse
Hi My guess is that U is the total voltage of the cell and you are subtracting the electrode potentials fii_1 and fii_2 from it, as well as the ohmic drop R_sei*i_loc. That leaves you the Galvani potential in the solution (?) phase; a_s is the specific area and C the specific capacitance (F/cm^2)? The dimensions do not quite add up because (d i_2/dx) is in A/cm^3, a_s*i_loc is in A and the time derivative also in A. But that is only a technical detail, more serious is the term (d i_2/dx) because if it deviates from zero there is a current source, i.e. excess charge in the system; in an electroneutral system the divergence of current is always zero. You need as many PDE physics nodes as you have variables as well as pertinent boundary conditions. The equation you sent does not make very much sense to me at least. Then you couple the equations in an appropriate way which I cannot possibly tell you, but I have a feeling that fii_1 and fii_2 cannot be free variables but boundary conditions of your problem. Currents neither are variables because they must be constant across the system (unless you have, indeed, a current source). BR Lasse

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 10 giu 2016, 01:35 GMT-4
Hi Lasse!

I am sorry I did not give you a complete picture of what I am trying to do.

I am trying to simulate Li-ion fuel cell incorporating few additional terms to incorporate Solid Electrolyte interphase layer. In am using a 2-D axial model to simulate diffusion in electrode and 1-D model to simulate the entire cell(2 Electrodes, electrolyte)

Actually "a_s" is specific surface area (m^2/m^3), iloc is intercalation current and i2 is the current in the electrolyte phase, C was capacitance, phi1,phi2 were solid and liquid potentials. U is open circuit potential and Rsei was resistance of SEI Layer.

I have boundary conditions on phi1,phi2. And yes, I have an external current source(for simulating charging and discharging).

1) Is it sufficient that i write all equations involving time derivatives in Coefficient PDE form and rest in Component->Definitions-> variables?
2) And is there a way i can extract variables from one component(2-D Axial) to 1-D Component? I have been trying using Linear extrusion coupling but I seem to be missing something, I have matched the boundary vertices in 2-D to boundary point in 1-D but it is still showing unknown variable(Because the concentration variable in 2-D has same value across a boundary)
3) Can we use the same variable in two domains in one component? Like lets say phi1(Solid potential) (has a similar equation in anode and cathode except the constants i.e conductivities etc) is to be used. In that way i will be able to define a single variable phi1 to both solid phases.

Thank you in advance!

Vema Sundeep


Hi Lasse! I am sorry I did not give you a complete picture of what I am trying to do. I am trying to simulate Li-ion fuel cell incorporating few additional terms to incorporate Solid Electrolyte interphase layer. In am using a 2-D axial model to simulate diffusion in electrode and 1-D model to simulate the entire cell(2 Electrodes, electrolyte) Actually "a_s" is specific surface area (m^2/m^3), iloc is intercalation current and i2 is the current in the electrolyte phase, C was capacitance, phi1,phi2 were solid and liquid potentials. U is open circuit potential and Rsei was resistance of SEI Layer. I have boundary conditions on phi1,phi2. And yes, I have an external current source(for simulating charging and discharging). 1) Is it sufficient that i write all equations involving time derivatives in Coefficient PDE form and rest in Component->Definitions-> variables? 2) And is there a way i can extract variables from one component(2-D Axial) to 1-D Component? I have been trying using Linear extrusion coupling but I seem to be missing something, I have matched the boundary vertices in 2-D to boundary point in 1-D but it is still showing unknown variable(Because the concentration variable in 2-D has same value across a boundary) 3) Can we use the same variable in two domains in one component? Like lets say phi1(Solid potential) (has a similar equation in anode and cathode except the constants i.e conductivities etc) is to be used. In that way i will be able to define a single variable phi1 to both solid phases. Thank you in advance! Vema Sundeep

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.