Learning Forum

[Kim Nam]

Hello, I hope you are all doing well. 

I've been doing analysis of a 20MW fire in the road tunnel, with the vertical gradient of around 2.5%, expecting the bouyancy force generated by the temperature difference caused by the fire to form a certain level of airflow towards the portal of the uphill side. 

However, I am confused about Boussinesq Parameters and Variable-density parameters.  I have read the user's guide and it says for bouyancy-driven flow problems, I have to activate either boussinesq model or operating density. 

I have very limited thoerical base, so I am struggling to understand all those mathematical terms, so I looked through some papers about fire analysis in the tunnel, and I found one paper stating that the Boussinesq approximation is not suitable for tunnel fire cases as it creates large rise/ increase in temperature. So I tried activating a specified operating density and I set the density of air at the initial state(at the initialized temperature) and the result seems to make sense for me. But I've been told to several people that I should not use specified operating density options for the fire case in the tunnel. 

I wish I could study more and find out the right answer by myself but this project has to be done within this week. Anyone's advise and idea will be very much appreciated. I’m on a strict deadline and am desperate for any sort of help or direction, hopefully someone on this forum can give me some guidance or tips. Thank you. 

[Rob]

:)  

You're correct that Bousinesq isn't suitable: it's good for temperature ranges around +/-10C. 

Incompressible ideal gas may be a good choice, read up on it. Who's telling you to not use the density as you described? That's how I do it.... 

[Kim Nam]

Hello, thank you so much for the advice and recommending a new way of employing ideal gas model, I appreciate it! I will definitely look into it.

 

Apologies for the barrage of questions but could you please spare some more of your time to check if my set up for the modeling is correct?

 

It’s 500m long tunnel and I put 20mw fire in the centre of the tunnel. The Vertical gradient of the tunnel is 2.5%, so the air flow and smoke will move towards to the uphill.

I set both of the portals as pressure outlet with the temperature of 310K and put energy value to the fire. I used SIMPLE and Body weight force scheme. I initialized with the tempeature of 310K and put the air density of the domain(around 1.1kg/m^3) on the Specified Operating Density  under Variable-Density parameter  (I definitely included gravity)

 

Does it seem to be a probable model setting?

 

Moreover, if you could possibly give some information on what would be different between putting specified operating density and leave the setting as default (unchecked). Could you recommend some materials that I can refer to, apart from the user's guide?

[Rob]

Looks about right, you're probably OK with incompressible ideal gas, if the flow is fast enough to be fully compressible the fire may be the least of your worries!  Did you use ALL of the decimal places in the operating density? 

https://ansyshelp.ansys.com/account/Secured?returnurl=/Views/Secured/corp/v231/en/flu_ug/flu_ug_bcs_sec_operating.html  Focus on the text just above eqn 7-68.  

[Kim Nam]

Yes, I am using incompressible-ideal-gas, and I put the initial density on the specified operating density.

I've looked up some materials and they say if I use the incompressible ideal gas option, I should put 0 in the specified operating density. Wondering if that is correct and if it makes sense?

 

Lastly- Could you tell me what I will have if I adopt incompressible-ideal-gas and not specify the operating density? Would the bouyancy caused by the fire be generated? I am still very confused about this operating density option.. and about what it does.

[Kim Nam]

[Kim Nam]

And god... this naturial convection problem seems to be never going to be converged. everything else seems to go down but only continuity goes up. could I trust the value from this model? I've put hex map mesh quite densely and the shape of the tunnel is a lot like the square so I think it's not a problem of mesh..

 

Thanks again Rob! You have no idea how relieved I am receiving help/ advice like this from you.

[Rob]

Put some monitor points into the model, and have a careful look at the flow field. Natural convection driven flows tend to be inherently transient so the steady solver will produce convergence plots like that. 

[Kim Nam]

Rob, thanks to your help I finally finished the project. I still have some uncertainty about the setting operating density: 

1. For single phase imcompressible ideal gas, If I set the operating density to 1.00kg/m3 for example, the bouyancy induced by the density difference will be triggered if the density is changed from specified density? 

2. If I don't specify the operating density, what is the difference from the specifying the operating density? 

3. What's better option when the large density change is expected? Should I specify the operating density to the ambient air denstiy or should I leave operating density unspecitied. 

 

[Rob]

 

The density value is set to assist at the domain outer flow boundaries. If rho_operating – rho_outside is zero you can set the pressure boundary as 0 Pa. If you don’t get a zero difference you need to account for boundary height against the difference so pressure becomes a profile and varies with height. 

For a couple of metres the pressure variation on the boundary isn’t much, but in natural convection flows the driving pressure for the flow is also very small. 

Hopefully  https://courses.ansys.com/index.php/courses/heat-transfer-modeling-in-ansys-fluent/lessons/how-to-model-natural-convection-in-ansys-fluent-lesson-4/  will explain more clearly. The search tool was upgraded overnight, so we can look everywhere rather than just in the Forum.