Hi all,

I'm modelling a mixed natural / forced convection problem, which is driven in part by a number of industrial electric fan heaters. The heaters are a very standard type of design, with a fan blowing air over electric elements. Power is ~9kW each, air throughput is ~1000 m^3/hour each (see http://www.tombling.com/heaters/portable-heaters.htm for a typical example of the type of heater).

I want to model in these as sources in my CFX model. I've included sources for the momentum and energy equations, which give me the correct heat input and flow rate. However, I'm not 100% sure about how to calculate appropriate turbulence model sources. I'm using the SST turbulence model in CFX. I will perform a sensitivity analysis to the turbulence source terms, but I would really appreciate any help you can give in establishing a sensible "best guess" on which to base my model.

What recommendations / best practice guidelines would you suggest for modelling the turbulence generated by a fan heater? A percentage of the fan motor power perhaps as a KE source?

Many thanks for your replies. Andy

(PS: Experimental assessments will not be possible in this case.)

Hi,

As a starting point I would guess that a crude fan would have 10% turbulence and a more refined fan say 1% turbulence. If anybody knows any better guesses let us know! This is k as percentage of the flow kinetic energy.

A sensitivity check is wise, if it is sensitive to it then you will be forced to get accurate turbulence figures.

Glenn Horrocks

try knowing about the velocity components, so you will able to calculate the Reynolds number. After that, you can choose the level of turbulence.

Many thanks Glenn - I'd been thinking of using 20% of flow KE - not a million miles away from your suggestion of 10%. I think these are fairly crude fans.

Any other experience / opinions about this (and particularly any references) will still be warmly welcomed.

Regards, Andy

Glenn,

Thanks again for your reply. I do have a quick question to clarify my understanding of your suggestion:

Did you envisage enforcing the "10% flow KE" value of K in the source region by applying a source to the turbulence equation of the form

S = -C*(K - Kspecified)

with some suitably large value of C, or would you just postulate a simple constant source term reflecting the volumetric production rate of K?

(Would you do something for the omega equation too, or ignore that?)

Many thanks, andy

"After that, you can choose the level of turbulence."

Rogerio,

Thanks for your reply - but the hard bit is knowing *how* to choose the level of turbulence. What rule would you suggest I use once I've computed the Re number?

Regards, andy

Take a look on papers throught the internet.

Hi,

Method of applying the turbulence source: If turbulence going into the fan is significant then you might just want to add the additional turbulence, maybe as a volumetric source term. If the turbulence going into the fan is irrelevant (this would be the norm I guess) then your S=-C*(K-Kspec) approach sounds good.

Source term for omega: Yes, you should do a source for omega as well. If you don't then the turbulence dissipation will be wrong. Again I am guessing here but if you guess a turbulent length scale of 1% of the duct diameter that can be worked back to a omega value.

Glenn Horrocks