Hi.

I'm stuggling to get correct heat transfer using the k-e model.

I'm trying to model heat transfer from fluid to a wall, using the k-e model, for a cylinder in a cross-flow. I've got y+ values in the range 30-300 as the good book suggests, but still the Nussult numbers are ten times higher than those suggested by an experimental correlation (Churchill & Bernstein, J Heat Trans. 99:300). Is there another criterion that I have to meet to make sure heat transfer is coorect? Anyone had similar experience? HELP!!!

Thanks, Hugo.

Dear friend,

I think, as per my knowledge what ever are u getting is right. Wall function approach predictions for heat transfer is very poor though the fluid dynamic solution is very good. I have expierienced the same, when i applied for the offjet flows. particularly, in recirulating and strong curvature in strain rate. I suggest u to go for the non-equillibrium wall functions or RNG modelling. I got very poor results in tne recirculating region even qualitatively.

Hi Vishnu,

In fact I carried on and found that the Nusselt numbers Fluent was giving were rubbish, but the heat transfer was quite reasonable - if i calulated my own Nu numbers they were is very good agreement with the experimental correlation, as long as 30 < y+ <300, which is the general recommendation.

I'll follow your suggestion and look at the non-equilib wall functions, but I want to avoid the RNG k-e model cos in my experience it didn't give good results for buoyancy driven flows, which is another part of the model.

Cheers, Hugh.

Hi Hugo,

you could try the SST Model, too. It should be pretty usefull for buoyancy driven flows.

Greetings! Fab