Total current calculated through flux integration not equaling input boundary condition
Posted 10 févr. 2024, 13:29 UTC−5 General, Results & Visualization, Studies & Solvers Version 6.0 5 Replies
Please login with a confirmed email address before reporting spam
Hi,
I have a electric current simulation setup, where my active terminal is an internal boundary and I assign it a current of 1A. My ground is the outer boundary of my model, which is built as a cylinder.
I am trying to numerically verify the total current coming out of the input terminal surface, by integrating the expression: nxec.Jx+nyec.Jy+nz*ec.Jz - which gives me -0.14A.
However integrating ec.normJ over the same input terminal surface gives me: 0.25A and integrating ec.nJ over the same input terminal gives me: 0.5A
I expect all these values to be equal to 1A (or close to 1A), but they are way off except for ec.nJ. (I realize these values are subject to change as I refine the mesh, cannot post that here because of the file size limitation, but they are never close to 1. For a finer mesh, ec.nJ is about 0.8 )
I would appreciate any inputs to help me understand where I am going wrong. May be I am incorrect in my understanding of the terms involved.
I also want to add that when I integrate nxec.Jx+nyec.Jy+nzec.Jz or ec.normJ or ec.nJ over the external boundary ground they all turn out to be 0.98A, which basically means all my input current is going to the ground (as one would expect.)
Attachments:
Please login with a confirmed email address before reporting spam
I took a look at your model. I only work on similar models now and then, but it seems to me you are doing something very strange here. You have two domains (the little cube and the big cylinder around it) both with finite conductivities that only differ by a modest ratio. Both are included in your computational space. What you seem to be trying to do (please correct me if I am mistaken) is to force a particular uniform current density (1A/mm^2,) thru one face of the little cube, outward into the rest of the cylinder. I don't think this is what a "terminal" boundary condition is for. As the Help sytem explains, "The Terminal node provides a boundary or domain condition for connection to external circuits, to transmission lines, or with a specified voltage or charge." You don't have any external circuit, right? You seem to be trying to simply specify a current density on an internal surface (generated presumably by some non-modeled invisible battery, chemical reaction, or perhaps something like the photoelectric effect??) and then force the rest of the physics to solve. Well, I believe Comsol will let you do that, but I don't think a terminal is the way to go. Instead, I suggest you use a "Boundary Current Source." Again, per the Help sytem, "The Boundary Current Source node adds a current source Q j on the boundary. It is applicable to interior boundaries that represent either a source or a sink of current." I tried that and it seemed (in just a quick check) to do what I think (not 100% sure) you wanted/expected. So please try that and let us all know if you are happy with it. Good luck.
-------------------Scientific Applications & Research Associates (SARA) Inc.
www.comsol.com/partners-consultants/certified-consultants/sara
Please login with a confirmed email address before reporting spam
Hi Robert,
Thank you for taking the time look at the model and offer your suggestions.
You are more or less correct about my intentions with the model, I apologize for not being clear enough.
However, I am trying to specify a current value, not a current density value, on a surface of the cube, then let the physics solve itself. I expect the current to primarily flow inwards first(into the cube), because the inner material is more conductive and then spread out to the ground (I am not sure if there is a way with the physics module to totally inhibit the initial outward flow of current). I am not modeling the source here, but let us assume that is connected to the surface of the cube by a wire (also not modelled here) and lets me control the input current(1A here).
I realize I am applying the terminal condition to an internal boundary, which isn't normally done, which is the reason I was trying to verify the physics by integrating the currents, but that's not adding up.
Let me know if this makes sense at all.
Please login with a confirmed email address before reporting spam
Ok, first, think about where on the surface you are imagining that this unmodeled wire is actually connected. Now think about how that location impacts the spatial distribution of currents. Now, ask yourself, if you may be specifying a problem for which there is not a unique answer. 2nd, in regard to the other, more mathematical question of specifying the surface integral as a boundary constraint, I'll defer to others here who may wish to comment on how to do that.
-------------------Scientific Applications & Research Associates (SARA) Inc.
www.comsol.com/partners-consultants/certified-consultants/sara
Please login with a confirmed email address before reporting spam
Thank you for your input.
I just want to mention that the file I have put up here is kind of an approximation of what I am actally want to model, since I am unable to share the actual model because of file size.
The active terminal surface of the cube is basically an active contact of an electrode, that is modelled in side the cylinder.
I tried replacing my terminal boundary condition with the Boundary current source as per your suggestion (converted 1A to xx A/m^2) , however, the electric field seems to be much weaker. Therefore, these two do not seem be providing the same input, which I find confusing.
I would still appreciate if someone could help me understand the flux calculations. If I am integrating the total current over my input terminal and my ground, I expect both of them to be equal to the value I assing (1A in this case). However, they are not. I understand numerical differences due to meshing, but refining the mesh doesn't improve the scenario either.
Please login with a confirmed email address before reporting spam
It isn't obvious to me that, taken at the abstract level you are describing, that you are specifying a current source that actually obeys the laws of physics. That said, I suspect that the physics of your situation would actually be straightforward to manage, but speaking for myself, I still don't understand what you are really trying to model. If possible, I suggest posting detailed drawings (i.e., from several views), along with labels, showing what you are trying to model. By that, I mean not what you think should be specified in Comsol Multiphysics, but what you think matters in the real world. E.g., "This is a very thin flat battery (like a watch battery)" (with an arrow pointing to it). "This is a resistor" (with an arrow pointing to it). "This is a copper wire" (with an arrow pointing to it). "This is a sealed metal box, with no openings." and/or "This volume is filled with water." Etc. Those are just examples. But I suspect that if you can describe your problem in terms like that, then somebody here (not necessarily me) will be able to give you useful advice about how to model it and how to interpret the computed currents, voltages, etc.
-------------------Scientific Applications & Research Associates (SARA) Inc.
www.comsol.com/partners-consultants/certified-consultants/sara
Reply
Please read the discussion forum rules before posting.
Please log in to post a reply.