Estimating Heat Transfer Coefficient

[EngAP]

With refernce to my previous thread (Wind Adjusted Heat Transfer Coefficient) I need to estimate the heat transfer coefficient on the inside surface of a sealed enclosure. The air in the enclosure is 20degC hotter than the ambient (0degC with ambient at -20degC). There is a heater and fan inside the enclosure.

I have run an experiment with this setup and estimated the combined inner and outer surface heat transfer coefficient to be 7W/m^2-K. However I expect the internal h to be greater than the external h due to forced (probably mixed) convection internally and natural convection externally. I would like to apply the experimental results to different exterior flow conditions so need to know the interior heat teansfer coefficient as the internal flow will be the same for all external conditions.

I do not have access to thermal simulation software. My thoughts are to repeat the experiment with external forced convection but the air speed would need to be extremely fast to be confident that the external heat transfer coefficient is in a stable (i.e. will not increase with increased air speed) condition and has reached an asymptote.

The internal geometry is reasonably complicated so that standard and published empirical heat transfer equation do not apply. I also do not know where on the fan curve the fan is operating so cannot estimate internal flow velocity.

Another way is to assume the external heat transfer coefficient was 4W/m^2-K during the experiment (natural convection), giving me an internal value of 10W/m^-K. However this is a big assumption, especially if I want to predict the heat transfer if there is wind/ice/rain externally as using h=10 or h=11 internally gives a substantial difference in heat transfer.

Is there a way of estimating the internal heat transfer coefficient?

 

[corus]

The heat transfer coefficient for steady state one dimensional heat flow is simply derived from -kdT/dx=h(T-Ta). If you can measure the surface temperatures in a region away from other features then you'll know dT/dx, and k for the material. Ta is just the ambient temperature.

corus

 

[EngAP]

Corus - thanks for the response.

I have been using the following equation to estimate the average values for h for the internal and external boundary layers:

Q = h.A.(T-Ta)

I had been working with results from a previous experiment that gave Ti (internal air temperature) and Ta. It was not designed for this problem hence the wall temperature (Tw) was not measured. The heat input (Q) was controlled and the surface area (A) was known. I had assumed h(internal) and h(external) were the same as I had no way of differentiating bewteen the two, thus eliminting one variable and giving an answer of 7W/m^2-K for h(internal) and h(external).

However I am now designing an experiment that will also give Tw (enclosure wall temperatures). Therefore I will know Ti, Tw (inner and outer - although as it is thin walled aluminium I expect Tw to be constant through the wall thickness and provide negligible thermal resistance), Ta, Q and A.

If I then take:

h(internal) = Q / A.(Ti - Tw)

and:

h(external) = Q / A.(Tw - Ta)

This will give me the average values for h for the inernal and exteral boundary layers.

Is this correct thinking?

Thanks.