Condensation algorithm and SETCOM results

[ESett5555]

Hi everyone,

I am currently studying the condensation algorithm in FDS, specifically the wall condensation algorithm, which is based on a wall function. This wall function is described in the Technical Reference Guide as follows :

I have two questions about this :

1. References for the wall function
   It appears that this wall function originates from a 2016 conference (CFD4NRS-6). However, I have not been able to find the conference proceedings or any other articles presenting this wall function. Does anyone know how I could access the original article or any related literature ?

2. SETCOM Experimental Data
   I am also looking into the SETCOM Validation Case for FDS. I want to compare the results given by FDS (using the files located in Validation/Juelich_SETCOM/FDS_Input_Files) with a simple Sherwood-based analytical calculation, as well as with the raw experimental data from SETCOM.
   In the Validation Guide, there is only a scatterplot comparing the predicted heat flux to the measured heat flux, and I wanted to also compare the condensed liquid water mass. Does someone know where i could find SETCOM experimental data ?

Thank you for your answers and for your time !

[rmcdermo]

Jason will probably have to answer the questions about the wall function. But have you looked at the heat flux data here?

https://github.com/firemodels/exp/tree/master/Juelich_SETCOM

If you need more than this, I can put you in touch with Alex Belt from Juelich.

[ESett5555]

Hello @rmcdermo,

Thank you for your answer.

I had not seen this data before, at least not in table format.

I assume that this is the data used to generate "Figure 12.142: Summary of SETCOM results in the Validation Guide", but could you clarify which heat flux it represents? Is it the total experimental heat flux, or specifically the condensation experimental heat flux?

If it is the condensation heat flux, I think that I would be able to calculate the condensed liquid water mass flux by just dividing by the heat of vaporization, correct?

[drjfloyd]

The heat flux is derived from TC measurements made inside the cooled plate (q = k dT/dx). It would be a total heat flux; however, most of that flux would be due to condensation.

[ESett5555]

Okay, I will see if I can work with that

[drjfloyd]

On the paper I can ask Alex if we can get permission to attach it here. I forget where I downloaded it from to be able to give you a wayback machine link for it.

[ESett5555]

Hello @drjfloyd,

If you could attach it here, that would be great !

Thank you !

[drjfloyd]

The author's gave permission to post a copy of the paper:

PaperCFD4NRS_2016.pdf

[ESett5555]

This will help me a lot, thank you !

[ESett5555]

Hi again,

I took the time to review the paper you posted two weeks ago and conducted several tests using FDS condensation algorithm.

It appears that FDS uses Kader's Wall Function as presented in equations 7 to 11. If I understand correctly, this article proposes two improvements to this wall function:

1. First, an adapted beta function is introduced, using beta = 1.1 instead of the formulation in equation 9
2. Second, it uses a turbulence model developed by Popovac and Hanjalic to better calculate Ys,turb+ (equations 12 to 18, specifically through equations 13, 17 and 18) and to model the 'suction effect'

Given that the article concludes the final version (called Improved wall function + adapted beta) is the best of the three, I was wondering why this specific version is not the one that was implemented in FDS. Is it due to the added complexity and higher computational cost ?

I also ran some tests on the condensation algorithm using SETCOM data alongside other data from a thesis on condensation. I found that FDS nearly systematically underestimates the condensation mass flux and the condensed water mass by 20 to 50 % compared to experimental data. I tried varying y+ and the cell size, but FDS stills underestimate the condensation mass flux.

Furthermore, since the CFD4NRS article dates back to 2016, I might be able to find more recent literature proposing an even better wall function based on this work. Would you be interested in a better wall function for wall condensation ?

Thank you again for your time !

[drjfloyd]

This was added to FDS 7 yrs ago. I don't recall all the details of the development process from back then. Probably a combination of cost and that for the larger grid cells typical of FDS simulations used in engineering practice the derivatives would not have been good for use.

This isn't something that has seen a lot of use since putting it in. If you are aware of a better wall function that is inexpensive and sutiable for use across a very wide range of grid sizes, we can look at implementing it.