louieho
Mechanical
- Mar 27, 2001
- 4
Hi guys,
I am having difficulty co-relating a mathematical model to an actual experiment. And I’d wondering if someone can help. Please pardon my ignorance to the subject as I had minimal exposure to heat transfer. I am relying heavily on a heat transfer textbook by Incropera and DeWitt.
I have an anodized chamber with diffuser at the inlet and outlet. The chamber is formed by placing a piece of glass over a gasket that has the hexagon shape (the diffusers form the two ends of the hexagon). Below is roughly the shape:
/------------------------------Inlet__/ \__Outlet
\ /
\------------------------------/
The depth of the pocket is 0.5mm deep only while the distance between the inlet and outlet is about 75mm. I was interested in the temperature profile of the region where the two edges are parallel and the distance between them are about 15mm.
The aluminum side to the chamber is maintained at a high temperature and the water supplied is about lower than room temp.
I have ensured that the Reynolds number is very low, i.e. the flow should be laminar.
In order to account for the cross sectional change in the diffuser, I programmed my model to do small finite steps calculation. And obtain the Nusselt number for fully developed laminar flow based on the ratio of the depth to width.
I used the formula for a internal flow system with constant surface temperature to do my calculations: dTo/dTi = exp[-PLh/(mass flow rate * specfic heat capacity)]. However, noting that we have essentially one heating surface, instead of using the PL, I used bL, where b is the width of the chamber, i.e. I neglected the convective heat transfer from the other 3 faces.
Is what I did right?
The problem I have is that my model is more efficient than experimental data, despite my attempts to make it conservative like counting one-side heating surface and slightly smaller width, which will increase the flow velocity for the same flow rate.
My calculated temperature is about 3-5 degrees C higher when I set the surface temperature to 40C and in addition, the slope is much higher that that of the experimental data.
Someone suggested that the anodized surface might provide thermal resistance. I tried to verify the thermal resistance of the 2mil think coating and find that the temperature gradient across the coating is insignificant. I can hardly measure the difference.
So what else can go wrong with my calculations?
Thank you.
Louie
I am having difficulty co-relating a mathematical model to an actual experiment. And I’d wondering if someone can help. Please pardon my ignorance to the subject as I had minimal exposure to heat transfer. I am relying heavily on a heat transfer textbook by Incropera and DeWitt.
I have an anodized chamber with diffuser at the inlet and outlet. The chamber is formed by placing a piece of glass over a gasket that has the hexagon shape (the diffusers form the two ends of the hexagon). Below is roughly the shape:
/------------------------------Inlet__/ \__Outlet
\ /
\------------------------------/
The depth of the pocket is 0.5mm deep only while the distance between the inlet and outlet is about 75mm. I was interested in the temperature profile of the region where the two edges are parallel and the distance between them are about 15mm.
The aluminum side to the chamber is maintained at a high temperature and the water supplied is about lower than room temp.
I have ensured that the Reynolds number is very low, i.e. the flow should be laminar.
In order to account for the cross sectional change in the diffuser, I programmed my model to do small finite steps calculation. And obtain the Nusselt number for fully developed laminar flow based on the ratio of the depth to width.
I used the formula for a internal flow system with constant surface temperature to do my calculations: dTo/dTi = exp[-PLh/(mass flow rate * specfic heat capacity)]. However, noting that we have essentially one heating surface, instead of using the PL, I used bL, where b is the width of the chamber, i.e. I neglected the convective heat transfer from the other 3 faces.
Is what I did right?
The problem I have is that my model is more efficient than experimental data, despite my attempts to make it conservative like counting one-side heating surface and slightly smaller width, which will increase the flow velocity for the same flow rate.
My calculated temperature is about 3-5 degrees C higher when I set the surface temperature to 40C and in addition, the slope is much higher that that of the experimental data.
Someone suggested that the anodized surface might provide thermal resistance. I tried to verify the thermal resistance of the 2mil think coating and find that the temperature gradient across the coating is insignificant. I can hardly measure the difference.
So what else can go wrong with my calculations?
Thank you.
Louie
