Need Assistance About Fire Model

[Berkan_vent]

Hi,

I want to simulate tunnel fire and i have some questions;

* For 50000000W fire i used resistance but i dont have any idea about Through-Flow K, Normal Direction 1 K, Normal Direction 2 K and Conductivity. How should i assign proper values for these properties?

* How to assign delay time for jet fans working? For example i want to activate the first group of jet fans about 30 seconds later.

* In the beginning of simulation i want some of my jet fans flow +x direction and for example after 3 minutes later i want them flow -x direction. Can i do that?

* I have fire and smoke curve ( T-squared Fire and T-Squared Smoke) values and these values depend on time. Which time curve do i need to use?

* For transient analysis, what is the best time step size and inner iterations.? I thought time step size is "10" and inner iterations is "5" is ok for my simulations?

* When i simulate my simulation in steady state i have high turbulance values and my simulations reach divergence because of high TKE and TDE values. I read Royce sayings and he said "we can use Turb/Lam ratio 1000 (max.)" if im wrong please correct me

* Also i want to simulate heat transfer between hot smoke layer and cold tunnel walls. Is it enough to activate radiation?

-For better understanding i attached .cfz file of my model 

Thanks in advance,

Berkan

[Jon.Wilde]

Hi Berkan,

To answer your questions:

* For 50000000W fire i used resistance but i dont have any idea about Through-Flow K, Normal Direction 1 K, Normal Direction 2 K and Conductivity. How should i assign proper values for these properties?

Your setup here is fine.

* How to assign delay time for jet fans working? For example i want to activate the first group of jet fans about 30 seconds later.

I would stop and start the analysis to do this. Set them as air initially and then switch to a fan material after.

* In the beginning of simulation i want some of my jet fans flow +x direction and for example after 3 minutes later i want them flow -x direction. Can i do that?

Same as above, although you are going to need a small timestep to capture that change. Maybe 0.1s or less.

* I have fire and smoke curve ( T-squared Fire and T-Squared Smoke) values and these values depend on time. Which time curve do i need to use?

As we can only have steady state on the scalar, I tend to use the thermal curve. How different are they? This may be one assumption you have to make.

* For transient analysis, what is the best time step size and inner iterations.? I thought time step size is "10" and inner iterations is "5" is ok for my simulations?

Initially, use a small timestep. 0.5s is a good start, to estalbish the flowfield, you could increase it later. Try 3 inner iterations. You can tell if you have too many inner iterations as the convergence will look like a digital wave. If you have too large a main iteration, it will diverge.

* When i simulate my simulation in steady state i have high turbulance values and my simulations reach divergence because of high TKE and TDE values. I read Royce sayings and he said "we can use Turb/Lam ratio 1000 (max.)" if im wrong please correct me

Yes, you can, although I suggest this is more likely down to your timestep (test this first). Maybe the mesh too, double the mesh on the fire - note the mesh around the underside, we need to ensure we capture the flow entering correctly.

* Also i want to simulate heat transfer between hot smoke layer and cold tunnel walls. Is it enough to activate radiation?

I would have thought you will need radiation here as the temperature differentials will be huge. By default we would see the walls heat up through conduction/convection anyway. You may wish to set a film coefficient to the outside of the tunnel to allow some heat to escape.

I hope this is useful to you.

Kind regards,

Jon

[Berkan_vent]

Thank you so much for your reply Jon,

I doubled the mesh around the fire site and i put an image for fire area, i think i captured the flow entering correctly. What do you think?

I have fire and smoke curve graphic like this:

Also smoke curve graphic like fire curve graphic. My smoke graphic reaches 80 m3/second at the end of 513 second. So can i use piecewise linear time curve? i didnt figure it out,

Thank you so much,

Berkan,

[Jon.Wilde]

Hi Berkan,

Take a look at the flow beneath the fire, where it passes through the small gap - it looks a little strange to me - indicating we should have a finer mesh. Otherwise, looks good!

Yes, you can apply the curve you have there as a transient heat load, using a piecewise linear graph, that is exactly what I would suggest.

Best regards,

Jon

[Berkan_vent]

I offset my fire model approximately  30 cm from road surface. Flow is entering the beneath the fire but i will assign finer mesh at there.

But velocity values and TDE, TKE values are very high, i think it is not normal.

Any suggestion Jon?

[Berkan_vent]

By the way it seems hot smoke and gases doesnt rise to the air, i dont understand why.

*I assigned environment variable for air.

Thank you

Berkan,