[kostnermo]

Hi everyone,

I am trying to validate a passive scalar transport case. The idea is to model the emission of simplified by people in a room with a certain air change rate. My goal is to obtain the average concentration in ppm in that domain.

My current issue is that I cannot correctly reproduce average concentrations compared to mass balances calculated on paper.
A simple mass balance can be done like this: http://www.engineeringtoolbox.com/po...oms-d_692.html

As test cases I tried the simpleFoam solver as well as the passiveScalarsimpleFoam by Yan Wang. In order to keep it simple, lets start with simpleFoam.


---

Steps I took:

1. The domain is a cuboid with (30*10*2.5) m, which yields 750 . Both inlet and outlet spans one full boundary (25 )

2. The source term is triggered by a (volumetric) source at breathing height within the room of about 40 (10*4*1) m

3. For the source term I use a field called TS with an entry in ./0/ and dimensions of . Apart from the inlet, I set all the BCs(TS) to zeroGradient.
The inlet concentration should be 400 ppm (vol-based). yields 7.22 E-04 for the inlet.

Then, the following entry in controlDict:

TS
  {
          type scalarTransport;
          active true;
          resetOnStartUp false;
          autoSchemes false;
 
          fvOptions {
                  TS-01
                  {
                          active true;
                          type scalarSemiImplicitSource;
                          scalarSemiImplicitSourceCoeffs
                          {
                                  cellSet volumeCell;
                                  selectionMode cellSet; 
                                  volumeMode absolute; // absolute <quantity>; specific <quantity>/m^3
                                  injectionRateSuSp
                                  {
                                          TS     0.000528; //kg/s/ accounts for 64 people (number of people * density(CO2) * 15 l/h CO2 /3600s)
                                  }
                          }
                  }
          }
 
  }

4. This test case represents 64 people with an emission of 0.96 per hour in a room with an air change rate of 2.3. This should yield 1000 ppm after a certain amount of time.

5. Now, I obtain the following visual result which looks reasonable to me (inlet is on the left):



An evaluation with swak4Foam yields an average value of ~1000 ppm for those conditions which is acceptable.
Now, if I test this setup for different occupancies, the results get significantly worse (rel. error > 10 %), as can be seen in the following plots:

http://postimg.org/gallery/icavz4dy/

The only sub-case that goes along with the static calculation, coincidentally, is the 64 people emission case. On both ends of the spectrum the results deviate significantly.
In my opinion, if I change the number of people, which scales the emission linearly, the relative error should not change at all.

Interestingly enough, the option specific/absolute doesn't change the result to a large extent, which is not what I would expect. In fact, the absolute option yields better results.

---

With that being said, I have some questions regarding my method:

• Is this setup comparable with a static calculation on paper?
• If yes, am I doing something wrong with the corresponding BCs?
• How do I appropriately address the source-term (absolute vs. specific) (maybe even SemiImplicitSource vs. ExplicitSetValue)? The fluctuating relative error leads me to think that I am assuming wrong source-term emission values.
• Do I need to multiply with density in order to calculate the source emission massflow, although the solver doesn't calculate the density explicitly? If not, how do I calculate the correct mass source-term?
• Does OF take care of the dimensions regarding the TS field? In other words, how come that using absolute vs. specific doesn't make much of a difference?

Any hint is appreciated!


Kind regards,

Patrick

[Naresh yathuru]

Hi patrick,

I have the same problem and same questions. were you able to find answers for your questions?

[kostnermo]

Hi Naresh,

I did receive some remarks as PMs but none that solve the issue.
How about you share your current findings and we work it out together?

Best,

Patrick

[kerim]

Hi,
Could some of you please share the case setup for openfoam? Best regards