[paolofug87]

Hi all!
I'm performing a simulation of a fire as a fluid volume source of heat and mass.
My energy source term is time dependent (stepwise).
Could anybody provide me the simple udf to accomplish this task?

P.S. who can suggest a simple method for accounting for smoke radiation?
I have introduced a carbon source term (inserting it in the mixture) but I don't know how to set absorption coefficient only for this material (I can only set an absorption coefficient value for the whole mixture). Is the DPM better?

Thank you in andvance

Sent from my MotoG3 using CFD Online Forum mobile app

[elmagnifico]

Hi Paolo

A little more information would be very useful.

Are you modeling chemical kinetics using a combustion model, or at least something purely mixing based like the Eddy Breakup model? What is the eventual goal of your modeling?

I'm not sure why you want to set the absorption coefficient of only carbon. Unless I'm missing something, there's no problem in setting the absorption coefficient of the entire gas mixture to a higher value. There are relations out there based on the temperature and soot volume fraction, here is a presentation that has a couple of relations along with the references (Pg 17):

http://web.student.chalmers.se/group...idesOFW5-2.pdf

[paolofug87]

Hi elmagnifico,

Thank you for the usefully material.

My goal is to describe with fluent a fire in a simple way because of large domain.
I don't use any reaction, just a mixture made up of oxygen nitrogen carbon monoxide carbon dioxide water vapour and carbon (representing soot). I set a fluid volume with an heat and mass source according to my combustion calculation (oxygen sink, product emission, convective and radiative release).
I just want to take into account the effect of combustion product on radiation.
What is the best radiation model (I'm using P1 model because it allows my to set the radiative contribution of the fire as source term)?
Following your suggestion, if I well understood, I just need to write an udf for the absorption coefficient that is dependent to carbon fraction, h2o fraction, CO2 fraction and temperature.
If it is the solution, could you please help me to write this udf? I was able to write the one for a time dependent source term but this sounds a little bit harder.

Thank you very much, I think this discussion could be very useful for many fire safety engineer, not much material is available for this methodology.

Sent from my MotoG3 using CFD Online Forum mobile app

[elmagnifico]

Paolo,

The thing is, the contribution of CO2, water vapor, etc to the absorption coefficient is way smaller than the contribution due to soot. So your absorption coefficient needs to be dependent on the soot volume fraction, and you can pretty much ignore the rest of the species. Here are a couple of UDF's I have written for the absorption coefficient. These are for a steady state simulation, but I think these should work for your transient simulation as well. You'll have to change the sootfV term though, to whatever you are using as a soot volume fraction surrogate (also remove the C_UDSI line, that is for a user defined scalar).

#include "udf.h"

DEFINE_PROPERTY(abs_coeff_Widmann,c,t)
{
real cellTemp = C_T(c,t);
real cellRho = C_R(c,t);
real cellM1 = C_UDSI(c,t,1);
real sootfv = cellM1*mC*cellRho/rhoSoot*normParameter;

real absCoeff = 2370*cellTemp*sootfv;
return absCoeff;
}

DEFINE_PROPERTY(abs_coeff_Sazhin,c,t)
{
real cellTemp = C_T(c,t);
real cellRho = C_R(c,t);
real cellM1 = C_UDSI(c,t,1);
real sootfv = cellM1*mC*cellRho/rhoSoot*normParameter;

real absCoeff = 1232.4*rhoSoot*sootfv*(1+4.8e-4*(cellTemp-2000));
return absCoeff;
}

I'm not sure you should use P1 in this case. The P1 model works well when your optical thickness is large; this might not be the case depending on what kind of combustion your are modeling. In any case, Discrete Ordinates is the superior model, and is much more comprehensive than P1. I would suggest you use that instead.

In general, I'm not really sure about your modeling approach. For one thing, how do you decide how much carbon to consider for soot? And how do you find/justify it's distribution in the domain? How do you find the temperature distribution in the flame?

If I were doing this, I would assume some representative soot volume fraction (varies from 10^-8 to 10^-5 in hydrocarbon flames), and just plug that number into the absorption coefficient UDF as the sootfv. I'd also use a really simple combustion model (like Eddy Breakup) to predict the temperature distribution. I think this approach (though far from perfect, or even just OK), might give your better results than your approach.

[paolofug87]

Hi elmagnifico,

thank you for the UDF examples.
I agree with you when you say that a simple fluid volume source of heat and mass does not properly reproduce the fire but normally in my application (and in fire safety application in general) it's a common practice to do this (see for example http://www.ansys.com/-/media/Ansys/c.../zitron-ja.pdf). Expecially because I want that the yield of production of co2 co and soot is exactly the value I have calculated with the tool you can find here http://www.thunderheadeng.com/pyrosim/resources/ "combustion calculator". This tool contains the model used by the NIST software FDS, typically I don't know what is burning so I use a default material (polypropylene) as FDS does.
The aim is to setup a method that applies to the most generic application (typically huge buildings, stations or tunnels).
To sum up I'm not interested in the fire near field but in the calculation of the tenability limits related to the smoke spread.
I tried also to reproduce the same reaction of the excel "combustion calculator" using a volumetric reaction (with the eddy dissipation model) and the coal calculator but I was not able to return the same stoichiometric coefficients.

In other words I would like to build a method that, from the maximum heat release rate, I can calculate the yields of co co2 and soot via the excel tool and than properly reproduce the diffusion of the products in the environment as well as the temperature and the radiation. Probably there are many approaches, I started with the simplest one but If you could suggest a more reliable method (with some explaination please ;-)) I will appreciate it so much!

In the meanwhile thank you again for your time

[sewgolam]

Dear friends,

I have also a question regarding the udf that you gys suggest for the absorptioncoefficient.

i'm injecting carbon particle into my domain and wat to change the absorption coefficient of thesis particles. For that i want to used a udf and change the absorption coefficient.

do you guys know how to do that i'm new in fleunt so i would appraited your help. and also when i you fluent the value for the absorption coefficient how could i able to check in fluent that the right value is set for the particles.

thank you very much.

[paolofug87]

Hi guys,

I'm wondering about model setup for simulating smoke, here some question:

1- My smoke is made up of CO CO2 and a scalar (soot), which is the typical value for UDS diffusivity? I found a CFX tutorial that says 1e-5 m^2/s with a soot density of 2000 kg/m3 but fluent asks for a value in kg/(m*s). I need to multiply for that density or just the density of the mixture (approximately the air density at that temperature)?

2- What is a the best setup for DO radiation model, and typical values for gas absorption coefficient: the material mentioned above says that the following formula can be used:
0.01 + 1.5*density*soot mass fraction/(smoke density * smoke diameter) [1/m]
and provides values for:
smoke density = 2000 kg/m3
smoke diameter 1e-6
By the way, the soot mass fraction is the one I can check when plotting the scalar contour?

Thank you in advance

Paolo

[kumphy]

how to give heat releasing rate to the source domain. is it really give smoke on the domain.how to give boundary condition.