CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   FLUENT (https://www.cfd-online.com/Forums/fluent/)

 pchoopanya March 28, 2013 19:14

Current density visualisation (PEM fuel cell add-on module)

5 Attachment(s)
Hello folks,

I am using a PEMFC add-on module to simulate a simple pemfc with a serpentine flow field. I believe I have achieved a converged solution and want to visualise the local current density distribution.

What really confuses me is that, every single published paper refers to different location (surface) when they show the predicted local current density distribution. Where (which surface) exactly should I be looking at?

I have investigated all these surfaces (interfaces);

1. mid-plane of the membrane

2. anode CL/membrane interface
3. cathode CL/membrane interface

4. anode CL/anode GDL interface
5. cathode CL/cathode GDL interface

6. anode GDL/flow field interface (no current collector)
7. cathode GDL/flow field interface (no current collector)

All I can see so far is that the current (I am using User Defined Memory, UDM3: Current Flux Density Magnitude [A/m^2, or current per area] as the variable of interest) at the anode side is a lot less than that at the cathode side, so regardless of which cathode-side interfaces, we should be interested in only the current density at the cathode side right? Please correct me if I am wrong. And then, which interface should I display the contour of local current density distribution exactly? (Interfaces 3 VS 5 VS 7?)

My another question is regarding the variable of interest; the CURRENT DENSITY as referred to as the performance parameter of the PEMFC in the open literature. Which UDS, UDM do we mean by this performance parameter CURRENT DENSITY?

I am using UDM3 as the unit is A/m^2 which matches the unit of current density in the published papers. What about UDM13: Transfer Current [A/m^3, or current per volume]? Can someone point out the difference between these two and where and when shall we look at each of these parameters?

Additionally, I have found a bizarre contour as a result from using UDM3 VS UDM13; I picked the interface between cathode CL/membrane and plotted these contours;

Please note that the oxygen inlet is located at the bottom-left and exits at the top right of the cell.

pic1: contour of UDM3: Current Flux Density Magnitude
Attachment 20233
pic2: contour of UDM13: Transfer Current
Attachment 20234
pic3: contour of mass fraction of oxygen
Attachment 20235
pic4: contour of mass fraction of water VAPOUR
Attachment 20236
pic5: contour of volume fraction of LIQUID water (UDS2: liquid saturation, s)
Attachment 20237

Many papers suggest that their current density (here I do not know which variable to use yet) follows the distribution of oxygen and opposes the distribution of liquid water. This means at higher current density will occur near the inlet area where oxygen is present and less current density will occur at the exit (area with high water saturation which is the indication of water flooding).

From the discussion above, using UDM13 gives a sensible result and agrees with the contour of oxygen, water vapour, and liquid water. The only frustrating issue is its unit of current per volume, not current per area.

On the other hand, UDM3 contour seems to be regardless of species distribution. What happens here?

One more question about the liquid/vapour water. From my understanding, the contour of water vapour in pic4 is not the indication of water flooding in the cell, as the name implies, it only shows the region (in red) where the mass fraction of water is very high due to the fact that oxygen gas is consumed in the upstream region. Is this right? Can someone assure me that UDS2: saturation is equivalent to the volume fraction of liquid water, s in multiphase model calculation?

Another question which is more generic. Tutorial given from ANSYS suggest a way to check for a good converged solution by checking mass flux balance through the REPORT > REPORT FLUX tool. When I calculate the net mass flux, the net flux is in the same order of magnitude of the inlet flux (ANSYS suggest net flux should be 0.1% of the inlet flux to ensure good convergence). However, the point monitors throughout the domain show that the variables have converged though. What is happening here? What is wrong here? Is this because the simulation involves species consumption?

When the programs iterates, it returns the value of averaged current density for both anode and cathode (why both? and what is the physical meaning of this averaged current density?), how does it calculate this value? I guess it must be something about surface integral over a certain surface, which interface is this? How exactly do we obtain this averaged current density?

Last question is that if the multiphase option is activated (under PEMFC model sub-tick box, not a separate multiphase model), if my reactant gases are both fully humidified. What is the correct boundary condition at the inlet for water saturation under UDS tab? Should this be "specified value" or "specified flux" and what value?

Please feel free to ask if you need some clarification on my questions. Thank you very much.

Pattarapong

 A CFD free user March 29, 2013 08:43

Hi
I'm working on PEM fuel cell as well. you asked a couple of questions. Regarding exact place for displaying contours, I suppose that most of the papered used membrane-catalyst interface for contour. You also talked about current density, when you start iteration, current density is reported in console window. you can also get the same value for current density from report-surface- integral type and The surface you need to select is wall terminal whether in anode or cathode not important.

 P.Charoen-amornkitt January 20, 2014 06:14

Quote:
 Originally Posted by A CFD free user (Post 417176) Hi I'm working on PEM fuel cell as well. you asked a couple of questions. Regarding exact place for displaying contours, I suppose that most of the papered used membrane-catalyst interface for contour. You also talked about current density, when you start iteration, current density is reported in console window. you can also get the same value for current density from report-surface- integral type and The surface you need to select is wall terminal whether in anode or cathode not important.
I'm currently working on PEMFCs also and the things that I don't understand is why the magnitude of current density from surface integral of anode (unit : A) is very low as compare to cathode (Should the value on both side be the same? Is there anything wrong?)
To plot the IV curve, can I just use the current density (unit : A/cm^2) that show in every iteration so that no need to divide the value from surface integral function by the area of the cell. and which one that i can use? cathode or anode? because the value is not the same.

and the final question, the massage "reversed flow" at outlet is shown and I have tried many ways to fix this but it still occurs. What does it mean? something flow back from outlet? how to fix it? (this thing has occurred since high voltage (1.0 0.9 V) and I haven't tried any low voltage yet) Is it mean that the solution has not converged yet?

Any comments would be highly appreciated
Thank you very much

Best Regards,
Patcharawat

 SAM_R February 5, 2015 13:26

Hi friends, Divergence in AMG Solver

I'm using PEMFC addon module of fluent and model of single pem fuel cell channel was made in Gambit 2.4 according to below open online tutorial file:
at last when I start the calculations following errors occurs:
Divergence detected in AMG solver species-2
Divergence detected in AMG solver uds-02
Divergence detected in AMG solver temperature

I'm trying to know what this means...
I've modified Under Relaxation Factors many times.
But Prblem is still happening,

-Saumy.

 manikanta March 24, 2015 07:58

divergence error in species

just try to change the v cycle to f cycle and keep BGS tab as method in controls>advanced control>AMG solver

 brahimchoice September 16, 2015 11:47

PEMFC tutorial model issues

Hi, I follow the tutorial provided by fluent for the PEM single channel, Still I have issues. Could you please send me a working model (case data or a valid gambit mesh for the pemfc). Thanks

first issue when I upload the mesh from gambit to fluent
WARNING: The mesh contains high aspect ratio quadrilateral,
hexahedral, or polyhedral cells.
The default algorithm used to compute the wall
distance required by the turbulence models might
produce wrong results in these cells.
Please inspect the wall distance by displaying the
contours of the 'Cell Wall Distance' at the
boundaries. If you observe any irregularities we
recommend the use of an alternative algorithm to
correct the wall distance.