| Issue 888: | Problem with wide band model | |
| 1 person starred this issue and may be notified of changes. | Back to list |
Application Version: Version 5.4.0 Serial SVN Revision Number: 0 Compile Date: I compile the sources (without modification) by myself Operating System: Windows XP I work on evaluating the capability of FDS to simulate the attenuation of radiation through a water mist. For that, I have compared the results of attenuation obtained with two numerical tools, FDS and a laboratory code. Their radiative models are totally different. In the second one, a Monte- Carlo method is combined with a fine spectral approach (367 bands) and a C- K model provides the characteristics of the gas taking into account H2O and CO2 contributions. I have first check the spray dynamics to be sure that the data could be compared. That’s the case. Secondly, I have run two simulations (one with and one without spray) with each numerical tool. The comparison on attenuation shows that the results obtained with FDS (with the gray gas assumption) and those obtained with the second code were very close (in terms of evolution and value). My problem is that the attenuation calculated by FDS with the wide band model is totally different in terms of evolution and value. I have tried to create a very simple case that clearly demonstrates the problem. This case (enclosed) looks like Simo Hostikka's in "Numerical modeling of radiative heat transfer in water sprays" (Fire Safety Journal 41-2006). I first thought this difference is due to the computation of the scattering and the absorption coefficients by the droplets in a cell. I have also programmed a code which calculates the scattering and the absorption coefficients of a cell using the FDS model. The input parameters are: the median volumetric diameter, the parameter for size distribution, the mass per volume of droplets, the assumed radiative source temperature, the number of polar angles and the non or gray gas assumption. The results show that these coefficients are similar between WBM and gray gas assumption what invalidates my assumption. That's why I think that this difference can be explained by the absorption of a gas but I am note sure and I am not able to extract the RADCAL model. Thank you for your help, Elizabeth
|
|||
,
Oct 07, 2009
I will ask Simo Hostikka to take a look at the case.
Owner: shostikk
|
|||
,
Oct 07, 2009
Thanks Elizabeth, I'll take a look at this when I get back home - hopefully next week. |
|||
,
Nov 18, 2009
Hello Simo, have you had time to take a look at my case? Best regards, Elizabeth |
|||
,
Nov 18, 2009
I forgot this one. I'm sorry. Trying my best... |
|||
,
Nov 18, 2009
Elizabeth, could you document what is the "difference in terms of evolution and value" you mention. The cases are very slow to start (actually to initialize) as you have set NMIEANG = 100. I don't know what to do with ATTENUATION.lay and .tec files. Based on first runs with a couple of modifications, I did not find any problem. The modifications were: 1. changed the MESH XB=0.,4.0 ,-0.5,0.5 ,0.0,2.0 2. open YMIN and YMAX instead of MIRROR. I'm not comfortable with mirrors. 3. Increased the flow rate to 24 to really see the effect. 4. Reduced NMIEANG to 20, radiation angles to 300 5. set H_FIXED=0 on MISC (we are not interested in buoyancy flow here). 6. Added 'INTEGRATED INTENSITY' slice. 7. Reduced DT_INSERT = 0.01 Also you should look at the following parameters: OFFSET=0. looks like trouble, MINIMUM_DIAMETER=0 the same. Also, add some DEVC with quantity 'RADIOMETER' to see the quantitative effect.
Status: MoreInfo
|
|||
,
Dec 15 (3 days ago)
Any success here? |
|||
,
Dec 15 (3 days ago)
hello Simo, I am sorry, I forgot to answer... I will try to answer as clearly as I can to your questions. I said "difference in terms of evolution and value" because normaly, in this configuration, the attenuation depends on the vertical location of the measurement point on the target panel : the closest to the floor the point is, the higher the attenuation is. With FDS code and the gray gas assumption, we got this evolution wheras with the wide band model, the attenuation decreases near the floor. More, the laboratory code gives an attenuation values range from 15 to 23 % on the target panel. The range obtained with FDS (gray gas assumption) is from 13 to 20 % and with FDS (WBM) is from 5 to 9 %. And I do not understand these two differences. I know the calculation is slow to start. I did not change this parameter for you. This input file has been used in a parametric study... The flow rate is very low because it is a real laboratory head. My calculation do not aim at evaluating the effect of the flow rate on the attenuation. Best regards, Elizabeth |
|||
,
Dec 15 (3 days ago)
I made the simulations with the above changes, and measured the 'RADIOMETER' quantity at three heights (1.5,1.0 and 0.5 m). See the attached pictures and input file. According to my results, the wide band model gives initially slightly lower radiative fluxes than the gray gas model (see the absolute values). When normalizing the each curve with its maximum value, we find that both models give attenuation of about 30 %. The attenuation is slightly higher close to the floor than higher up. This demonstrates that the result may indeed be due to the gas absorption. You could try playing with the initial relative humidity and the EVAPORATE=.FALSE. option of the PART type, if you want to study this further. Do you think there is still a problem with the code performance, or can I close the issue? Simo |
|||
 Sign in to add a comment
|
