Practical use of an overall heat transfer coefficient?

[MysticDream]

Say I have a real counter flow heat exchanger using air and water. I run a test so I know every parameter; mass flow rate of both fluids, surface area, and in and out temps.

This is great. Now I can calculate the heat transfer rate, my log mean temperature difference, and overall heat transfer coefficient using standard formulas. I can also calculate the effectiveness of my heat exchanger and the number of transfer units using the NTU method.

Now I realize I need different results. I may need to change the parameters of my heat exchanger, but I don't want to waste time and money building and testing different setups to get the desired results.

I realize I will need to change the input temp of one of my fluids and/or the surface area of my exchanger (linearly). I know I can use the NTU method to calculate the heat transfer rate and outlet temps IF I know my overall heat transfer coefficient. I can calculate my overall heat transfer coefficient from my first test using:

My question is , if I change the values of one (or both) of my inlet temps OR I change the surface area value OR I change the mass flow rate, can I use the SAME overall heat transfer coefficient from the above formula throughout all temps and areas and still obtain an accurate result using the NTU method?

I was surprised to find out how difficult it is to get a clear answer to this question through internet searches. There are many descriptions and examples of what an overall heat transfer coefficient is and how to use it in one specific case, however, I was not able to find any information about it's practical use in calculating results by changing parameters. Of what use would an overall heat transfer coefficient be if it is not a constant for a particular heat exchanger that could allow you to change parameters to predict the results?

 

[BvU]

through internet searches

yeah, well...

There is a magnificent book called the Wärmeatlas where you can find how $U$ depends on flowrates.

Depending on how "accurate" is defined for you, the formula allows a reasonable range of temperatures.

$\$

 

[MysticDream]

yeah, well...

There is a magnificent book called the Wärmeatlas where you can find how $U$ depends on flowrates.

Depending on how "accurate" is defined for you, the formula allows a reasonable range of temperatures.

$\$

Very interesting. Thanks.

 

[Lnewqban]

....
My question is , if I change the values of one (or both) of my inlet temps OR I change the surface area value OR I change the mass flow rate, can I use the SAME overall heat transfer coefficient from the above formula throughout all temps and areas and still obtain an accurate result using the NTU method?

The coefficient should change is the flow velocity or diameter of the pipes are changed.

 

[MysticDream]

The coefficient should change is the flow velocity or diameter of the pipes are changed.

So the same coefficient can be used for different inlet temps and different surface areas, only if the areas are changed by length (not diameter) and only if the mass flow rates (of both fluids) stay the same?

 

[Lnewqban]

So the same coefficient can be used for different inlet temps and different surface areas, only if the areas are changed by length (not diameter) and only if the mass flow rates (of both fluids) stay the same?

I would say yes.
The coefficient modifies the product of area and temperature difference to produce a value of thermal energy flow.
That flow rate is greatly affected by the conditions of the film of fluid located closest to the walls, like velocity, turbulence, viscosity, hard deposits on the walls, etc.

 

[MysticDream]

I would say yes.
The coefficient modifies the product of area and temperature difference to produce a value of thermal energy flow.
That flow rate is greatly affected by the conditions of the film of fluid located closest to the walls, like velocity, turbulence, viscosity, hard deposits on the walls, etc.

Thanks.

 

[Chestermiller]

Do you know how to model a counterflow heat exchanger (e.g., the one you are working with), including prediction of the internal tube heat transfer coefficient, external (shell side) heat transfer coefficient, and, from these the overall heat transfer coefficient U? If so, why are you not doing modeling calculations for various cases to answer this question you have been asking about U? We can wave our hands about this all day long, but that wastes everyone's time.

Please. show us how you would go about calculating the internal tube heat transfer coefficient.

 

[MysticDream]

Do you know how to model a counterflow heat exchanger (e.g., the one you are working with), including prediction of the internal tube heat transfer coefficient, external (shell side) heat transfer coefficient, and, from these the overall heat transfer coefficient U? If so, why are you not doing modeling calculations for various cases to answer this question you have been asking about U? We can wave our hands about this all day long, but that wastes everyone's time.

Please. show us how you would go about calculating the internal tube heat transfer coefficient.

Did I not show this in the OP? If I know the heat transfer rate, area, and the log mean temperature difference, I can calculate the overall heat transfer coefficient.

I can calculate the internal or external using mass flow rate, heat capacity of the fluid, temperature difference, and area.

 

[Chestermiller]

Did I not show this in the OP? If I know the heat transfer rate, area, and the log mean temperature difference, I can calculate the overall heat transfer coefficient.

I'm not talking about determining U experimentally on an existing system. I'm talking about designing a counterflow heat exchanger from scratch.

I can calculate the internal or external using mass flow rate, heat capacity of the fluid, temperature difference, and area.

Let's see your basic design equations for determining the internal heat transfer coefficient (Nusselt number).

 

[MysticDream]

I'm not talking about determining U experimentally on an existing system. I'm talking about designing a counterflow heat exchanger from scratch.

Let's see your basic design equations for determining the internal heat transfer coefficient (Nusselt number).

I’m aware that the coefficients can be approximated using other formulas that use the nusselt number but I haven’t tried that yet. I already had a heat exchanger and figured empirical data would be most reliable.

 

[Chestermiller]

I’m aware that the coefficients can be approximated using other formulas that use the nusselt number but I haven’t tried that yet. I already had a heat exchanger and figured empirical data would be most reliable.

Even if the relationships I'm referring to would definitively enable you to answer the questions you've been asking (and without any "hand waving")? Do you want an answer you can rely on or don't you? See Chapter 14 of Transport Phenomena by Bird, Stewart, and Lightfoot, particularly Section 14.3.

 

[MysticDream]

Even if the relationships I'm referring to would definitively enable you to answer the questions you've been asking (and without any "hand waving")? Do you want an answer you can rely on or don't you? See Chapter 14 of Transport Phenomena by Bird, Stewart, and Lightfoot, particularly Section 14.3.

Actually, yes. That’s why I’m here. Thanks, much appreciated.

 

[gmax137]

I think the other responses probably will help you out.

As to this question:

Of what use would an overall heat transfer coefficient be if it is not a constant for a particular heat exchanger that could allow you to change parameters to predict the results?

Think about it. If you have some Hx with known conditions and known overall HTC U, and you could "keep" the same U even if the Hx parameters change, then wouldn't all Hx have the same U? You could change one parameter at a time until the Hx is quite different, why would you expect the U to be unchanged?

 

[MysticDream]

I think the other responses probably will help you out.

As to this question:
Think about it. If you have some Hx with known conditions and known overall HTC U, and you could "keep" the same U even if the Hx parameters change, then wouldn't all Hx have the same U? You could change one parameter at a time until the Hx is quite different, why would you expect the U to be unchanged?

Not if U was dependent on fluid type and Hx geometry. I would expect it to remain unchanged so one can design a Hx with desired temps and heat transfer rate without building prototypes. If I'm just building prototypes to get parameter info, why would I even need a U? My prototype could tell me everything I need to know as far as in and out temps, heat transfer rate, and mass flow rates. It's the U that can allow me to predict the outcomes without doing a physical test. U is dependent on mass flow rates and Hx geometry. Knowing that, I can find my U with a test, or look it up from another source, or figure it out with formulas that Chestermiller suggested; then I can expand my Hx linearly and/or change in or out temps and use the same U to determine everything else. If one or both of my mass flow rates change, or I change the geometry significantly, I will need a new U to be able to predict the parameters.