Heat transfer coefficient 3

[jproj]

I am currently working on some insulation thickness calculations (trying to develop a spreadsheet) and need some help in determining a heat transfer coefficient. The tank (horizontal carbon steel) is a little more than half full of saturated liquid with saturated steam in the head space (51 psia @ steady state). I have the thermal conductivity of steel and the insulation, but I am not sure how to calculate the heat transfer coefficient from the liquid & steam to the shell. I really haven't looked into it yet, but I would assume that I need the same type of help calculating the heat transfer coefficient from the outside insulation surface to the ambient air.

My thinking is that from the liquid to the shell is conduction, while the steam is convection (and maybe radiation?). If someone could point me in the right direction, it would be much appreciated.... heat transfer was definately not my forte in college!

jproj

 

[jproj]

One more question.... if neither the steam or water (liquid) are moving is conduction the only heat transfer mechanism?

 

[IRstuff]

Assuming that the liquid has the highest temperature, the convective heat transfer at the surface supplies new energy into the gas, which in turn convects heat to the tank wall.

I'm a little hazy on whether this always happens, but if the liquid is watery, there should be some convection within the liquid itself. It's the same effect that's visible when heating water in a clear vessel.

TTFN

 

[tickle]

Rockwool produce a book giving numerous tables for economic insulation thicknesses for different requirements (heat conservation, personnel protection, freeze protection). I find it very useful and find that it correlates well with the figures that larger organisations use in their insulation specifications.

(Sorry to Rockwool if you get an influx of callers asking for copies of your book - but I find it very useful and use it regularly)

 

[poetix99]

IRStuff:
When liquid and vapor - whether water or some other substance - are in equilibrium, the temperature is the same in both phases. Even with convective flows established, as I describe below, there is no significant convective h.t. directly from the liquid surface to the vapor.

jproj:
Both the liquid and the vapor will have convective "cells" where fluid near the tank wall that has been cooled will sink, with warmer fluid rising in the center. The liquid phase will, of course, also have conductive h.t., but it is not intuitively obvious which mechanism will dominate, since it depends upon the actual size of the tank (among other things) regarding convective h.t. rates.

The vapor will also having condensing h.t. taking place along the walls. Chances are good that this will dominate the vapor side h.t., but you'll have to determine some h.t. coefficients from a good graduate level text; the condensing h.t. rates could depend on the vapor velocites established by convective circulation. (It's a little bit "circular", to make a bad pun.)

If heat is not added to you tank, it won't stay at 51 psia for very long: because the volume is fixed, the temperature (and pressure) will steadily drop, even though it will remain in a saturated state, with liquid & vapor.

 

[1969grad]

For air side film coefficients see the ASHRAE Handbook, Fundamentals volume.

 

[Haf]

jproj,

"if neither the steam or water (liquid) are moving is conduction the only heat transfer mechanism?"

Poetix basically answered this, but let me throw a couple terms out that may jog your memory. If gaseous material is in contact with solid or liquid material and the two materials are at different temperatures (away from the interface), there is always some convection. If the gas has some velocity, you have forced convection. If the gas is stationary, you have free convection (based on thermally induced flow, as poetix pointed out).

As far as calculating convection coefficients: in general, they are difficult to calculate (especially free convection coefficients), but here are the basics. Generally, to calculate the convection coefficient, you must calculate the Nusselt number, which is a function of Reynolds number and Prandtl number (for local coefficients, it is also a function of distance). This means you need to know the conduction coefficient, thermal diffusivity, fluid velocity, fluid dynamic and/or kinematic viscosity, fluid density, and fluid characteristic length. If conditions change significantly as you travel across your surface, you must use local coefficients (based on local conditions) instead of average coefficients.

If you have material condensing at the interface, that makes this calculation even uglier.

In practice, convection coefficients are often determined experimentally.

Hope that helped, and good luck.

Haf

 

[jproj]

Thank you all for your advice. I now see why I couldn't find anything online! While I am interested in calculating the heat transfer coefficients, I think that the easiest (and most conservative) approach would be to just assume the inside surface temperature is equal to the operating temperature and go from there. If anyone has some good references (books, journals, etc.) regarding "free convection" heat transfer coefficients and their calculation, I'd be interested to hear about them.

Haf:
Is calculating the transfer coefficient from the outside of the insulation to the air as difficult? We are really just interested in the surface temperature of the insulation (personal protection), but I am curious as to how it's calculated.

Thanks everyone!!

jproj

 

[TD2K]

The outside heat transfer coefficient is critical to estimating the insulation surface temperature. The heat transfer across the insulation from the inside temperature to the outside surface temperature equals the heat transfer to the ambient air. The lower this coefficient, the higher the skin temperature is required.

Perry's Handbook for Chemical Engineers has a table showing the heat loss from surfaces to air, you could take a look at those. Barring that, I've typically used a ho of 2 BTU/hr/ft2/F for those calculations corresponding to a larger surface in still air conditions.

 

[Haf]

The heat and mass transfer book I used in undergrad is: "Fundamentals of Heat and Mass Transfer" by Incropera and DeWitt, published by John Wiley and Sons. I have the Fourth Edition. I think this is an excellent text.

Chapter 9 focuses on free convection. Section 9.6.3 (page 501 in my edition) gives a relatively brief treatment of the case of free convection with a long horizontal cylinder (there are also a couple of references to journal articles at the end of the chapter). There is even an example problem (example 9.4, page 502 in my edition) that considers free convection of a 0.1 m outside diameter horizontal, high pressure steam pipe.

Haf

 

[Haf]

Oops, I just noticed you're interested in a tank, not a pipe. That can be handled as well, provided you are comfortable with modeling your tank as, you guessed it, a sphere. See Section 9.6.4.

Haf

 

[speco]

I have a suggestion which may simplify your problem. Since your tank is to be insulated, the controlling resistance in the heat loss is due to the insulation. That is, the resistance of the insulation is far greater than the metal resistance of the tank wall or any inside convective components.

Here are some ideas to help justify this. In a plate-coil application, where a plate coil is submerged in a watery solution, it is customary to use an overall coefficient of about 140 (BTU/hr-ft^2-deg F) between steam and a watery solution. Normal steam condensing coefficents (for the steam film condensing only) for the steam portion of your tank would run in the 1000-1500 range.

In your case, I would therefore suggest that you consider the inside wall of the tank to be at the saturation temperature of the steam (about 282F), and only condider the plate, insulation, and air film coefficients (or resistances) in your calculation. This should give you a conservative design and save you a bit of tail-chasing.

 

[25362]

To Haf, when estimating the HTC (or the Nu) for natural (free) convection the Re number is usually kept out.

 

[jproj]

Ok, from what I've read regarding natural convection from heated surfaces cooling to air, the heat transfer coefficient is a function of the Grashof number and the Prandtl number. My old Heat transfer book has used experimental data to correlate the them to the Nusselt number, which can be used to calculate the heat transfer coefficient. Does that sound right?

 

[Haf]

That sounds right.

Keep in mind that you'll be working with an estimate, at best.

 

[quark]

This link could be of help for insulation design and calculating outer skin temperatures.

http://www.cheresources.com/insulationzz.shtml

Regards,

 

[1969grad]

As I stated before see the ASHRAE Fundametals Volume for the outside (film) heat transfer coefficient. They have a bunch of equations some simplified for air. They also have all the equations for calculating the overall U value for an insulated pipe. (Q=UA*TEMP DIFF)

You may find that the inside film coefficient to be insignificant and can be ignored. Fact is that for an insulated vessel the conduction factor for the vessel wall is not a big influence. Major factors are the insulation K and the outside film coefficient. You will see what I mean if you check the above reference for calculating the overall U and plug in some representative values. Most all heat transfer books show you how to calculate the overall U for a insulated pipe (a vessel is just a big diameter pipe.)

If the vessel is inside then you will have natural convection on the air side. If it is outdoors then you will have forced convection and should assume a wind speed of 15 mph.

For large diameter vessels assume a vertical flat plate. You can also assume that the vessel heads are a vertical flat plate.

For burn protection I use an outside surface temp of 150 F. Some use 140 F, some use 160 F, etc. As far as I know there are no OSHA or other standards for the surface temp.

ASTM has a standard which includes a Fortran program to calculate the heat loss and surface temp for insulated pipe. Included is an approximation for the outside film coefficient which is sufficient for insulated piping and vessels. It is not accurate enough for a bare vessel.

 

[IRstuff]

The reason there is so much variation is that the degree of injury depends on time of contact and intimacy of contact. Nonetheless, as the following indicates, a 3 second immersion in 140°F water will produce a 3-rd degree burn.

TTFN

 

[IRstuff]

Sorry, left the link off, plus can't find the original, but here's another one with similar numbers 5 seconds at 140°F for 3rd degree burn:

http://www.burnfund.org/library_burnpreventiontips.html

TTFN

 

[jehar]

You can find calculation methods for heat loss through insulation at

http://www.HRSGdesign.com

under "Other". And, you can find the publication "Determination of Heat Gain or Loss and the Surface Temperatures of Insulated Pipe and Equipment Systems by the Use of a Computer Program, ©1995 ASTM" under "publications".
Jack Hardie