28 replies to this topic

[processeng05]

Hey, i am working on a spreadsheet to calculate heat loss from a hot emulsion of water and bitumen flowing through an insulated pipe. We have hot emulsion (water and bitumen) at 176 C entering an 8" pipeline, 1000m long with 1" thick fibre glass insulation. The volumetric flowrate of emulsion is 500 cubic metres per hour. The ambient air temperature is -15 C and we want the emulsion to be cooled to 120 degrees when it exits the pipeline. The wind speed is 5m/s. I want to calculate the total heat  loss through the system and also the time that it would take for the emulsion to cool down to this desired temperature while exiting the pipeline. I have created this spreadsheet to calculate the heatloss. I calculated the heat transfer coefficient for convection from insulation to air and heat transfer coefficient from emulsion to pipe. Then I calculated the total resistance of the system to get the total heat loss. I have attached my work with this reply. Can someone be kind enough to check my work for me and tell me if there is something wrong? I have been stuck on this for a while now. Thanks a lot
 Che resources.xlsx   495.28KB   1689 downloads

[Steve Hall]

Are you sure about the thermal conductivity of the fiberglass insulation? Should it be 0.035 W/m-K, not 0.0035 W/m-K?

 

What is the viscosity of your emulstion?

[Raj Mehta]

Yes Mr. Steve is right the thermal conductivity of the fiberglass insulation is 0.035 W/m-K (engineeringtoolbox.com gives it as 0.04 )

 

By the way I am unable to download the spredsheet due to reasons unknown to me. I will try to download it and get back to you.

 

Best Regards,

Raj

[Steve Hall]

Using for the emulsion, viscosity = 5 cP and thermal conductivity = 0.64 W/m-K, I calculate the exit temperature from the pipe as 174.1 C (uninsulated) or 175.9 C (1-inch insulation). I used my newest version of PIPESIZE which calculates inside and outside heat transfer coefficients (program in Beta, not released yet).

 

Note that the velocity is 4.3 m/s, so the residence time in the pipe is less than 4 minutes.

 

Check me on this: 176 C - [214 W/m x 1000 m / (500,000 kg/h x 4.18 kJ/kg-C)] = 175.9 C

Edited by Steve Hall, 11 April 2013 - 05:38 AM.

[processeng05]

Steve, thanks. You are right with the exit temperature of 175.9 C but like I mentioned we want the emulsion to be exiting at a temperature of 120C. So is there anything that we can change in our process conditions to change our exit temperature? Also I changed the numbers in my spreadsheet with data for water because i wanted to compare the results from my spreadsheet to this hysys simulation that I did but my spreadsheet results seem to be way off. The heat loss that hysys gives me for water with the exact same conditions gives me heat loss of 3.5*10^4 kW. I am so lost -__-

[Steve Hall]

I'm nowhere near 35,000 kW = 35,000,000 W either.

 

For uninsulated pipe, 1000 m long, I get 4,280,000 W total heat loss.

With 1" insulation, I get 215,000 W.

 

Inside heat transfer coefficient 2,150 W/m2-C

Outside heat transfer coefficient 32 W/m2-C

 

Overall U 22.4 (uninsulated), 1.33 (1-inch insulation) W/m2-C

 

Temperature at end of the pipeline:

174C uninsulated

176C insulated

Edited by Steve Hall, 26 March 2013 - 11:44 AM.

[Steve Hall]

I've quadruple-checked my calculations and think my original posts were correct. I revised Post #6 to remove the ambiguity.

 

I see no way to cool the hot emulsion to < 120C just by changing process conditions - even reducing the flow rate to 300 kg/min results in an exit temperature of about 130C. There just isn't enough surface area on the pipe, and outside film coefficient in air is extremely low.

[Dacs]

Your heat capacity is a bit off. Is it really 4180 kJ/kg-K (or 4.18 kJ/kg-K)?

 

For what it's worth, I did my own calculation (using 4.18 kJ/kg-K) and I'm getting around 175.1°C exit temp (insulated). This might be less accurate since I only considered HTC at the air side.

[processeng05]

Steve you are right, really appreciate your input. Thanks

[Steve Hall]

Breizh,

 

Thanks for the spreadsheet. I have a couple of questions:

 

1) Is your Reynolds number calculation correct? I don't see pipe diameter in your equation (F45 on your insulated pipe worksheet)

 

2) Doesn't Prandtl number require the kinematic viscosity (= dynamic viscosity divided by density)?

 

3) Did you mean for the viscosity in F35 to be 0.005 Pa-s, or is that supposed to be mPa-s (= centiPoise)?

 

Regards,

Steve

[breizh]

Thanks Steve,

 

Becoming rusty , need to review it.

 

Pr = cp*mu/ lambda  with cp ; heat capacity ; mu :dynamic viscosity and lambda : fluid thermal conductivity .

 

For dynamic viscosity  5 cpo = 5 e-03 Pa .s 

 

Breizh

Edited by breizh, 03 April 2013 - 02:03 AM.

[Steve Hall]

Got it, for Pr and viscosity. Thanks.

 

Re = density x velocity x diameter / dynamic viscosity

or

Re = velocity x diameter / kinematic viscosity

 

So, instead of Re=8.6e5, get 1.7e5

 

Please recheck your non-insulated pipe calculation for U. I don't see the error (if there is one), but 4.6 W/m2-K when hi = 37 and ho > 3000 seems too low. Should be around 22 to 25, I think? 

 

By the way, I'm using ID = 202.7 mm for Schedule 40 8" pipe (after changing ID, Re and U referenced above are calculated in your spreadsheet).

[processeng05]

This is my work, can someone double check it please. My overall heat transfer coefficient is completely different than Breizh's, not sure why though. I compared my results to a hysys simulation as well and the results seem to match this time.

Attached Files

•  Che resources.xlsx   593.42KB   1246 downloads

[Steve Hall]

I think your calculation for U has an error.

 

My calculations agree with yours for hi and ho (well, fairly close - I'm at 3577 for hi), but my U is 1.3, giving an exit temperature of 175.6.

[Pingue2008]

Dear All,

 

I have not dealt with this problem before. it looks interesting could anyone of you point me to a good reference where I can read some more about this type of problem and hopefully understands the formulas that you have used.

 

Thank you,

Allen

[breizh]

Pingue 2008 ,

let you consider the document attached , and any good book about heat transfer . Another resource could be

"Handbook of chemical engineering calculations-3rd Edition ,NP Chopey - Chapter 7

 

Note : As pointed out by Steve , there is a mistake for the calculation of U ( global heat transfer coefficient)  in my previous excel sheet for non insulated pipe . My apologies

 

 

Breizh

Edited by breizh, 07 April 2013 - 08:37 AM.

[processeng05]

Steve yeah i have checked my calculations like a gazillion times but for some reason, I just can't get it to match with your's or Breizh's. Well I figured that there is some difference in our ID and OD, that could probably be the reason for slight differences but 0.33 and 1.2 is quite a big difference for value of U if you think about it. 

[Steve Hall]

Dear ProcessEng05 and everyone following this thread,

 

It's clear that the debates seen above are somewhat contradictory, and the topic is not fully settled. With Breizh's collaboration (off-line), I put together what I think is a definitive and fully articulated spreadsheet that illustrates how to calculate the heat loss from a pipe, insulated or not.

 

Rather than have separate worksheets for the insulated or not-insulated cases, this spreadsheet combines all into one. Simply put the insulation thickness = 0 to calculate for a bare pipe.

 

Steve

Attached Files

•  Heat_Loss_from_Pipe.xls   88KB   2836 downloads

[Art Montemayor]

Steve:

 

I am one of probably many other members who have been following this thread - mainly because of the basic, and important  heat transfer topic that it deals with.

 

I want to be the first in giving you my profound thanks for the great input you have contributed to this thread - together with a professional, polished, and efficient workbook that allows everyone to not only resolve a pipe heat transfer problem, but also stands as an example of real, professional engineering work product.

 

I would hope that every student and young engineering graduate realizes what a great gift you have put at their disposal and not only downloads the workbook, but spends all the time needed to carefully study your calculation methodology and profressional presentation.

 

Once again, many thanks and Kudos.

[Chris Haslego]

Quick note from the admin: Yep....this will be stored in the File Repository very soon. Great work Steve. Thanks for taking the time. And...thanks to Art for pointing me to this post.

[sang]

Hi Steve

 

This is indeed a fantastic spreadsheet. Do you have one for pipes where the ambient fluid is water, or can this be easily modified to use water properties?

 

Really appreciate your contribution and look forward to hearing from you.

 

Regards

Sanjay

[Steve Hall]

Sanjay,

 

Thanks so much for your kind appreciation.

 

Modifications are needed to calculate for water as the ambient fluid. There are, again, two cases: natural and forced convection. And the radiation component doesn't apply. I could do this for you, but for a small fee. Message me privately if you are interested.

 

I'm about to release Version 3.1 of my spreadsheet program, PIPESIZE (available on this website or from my site). The new release includes the heat gain/loss calculations described by the spreadsheet included in Post #18 above, but they are implemented in a VBA macro. PIPESIZE uses pull-down selections for different pipe and insulation materials, has a built-in user-extendable database with temperature-dependent properties for more than 100 fluids, and many more features. PIPESIZE calculates pressure drop due to friction for liquids and gases, can be configured for gaseous isothermal or adiabatic solutions, and allows you to toggle between customary U.S. and SI units of measure. The new version adds a function to save and restore calculations to a line list with the click of a button. All of my spreadsheets are protected from accidental changes to formulas, but I don't use passwords so users can freely modify them to meet their specific needs.

[breizh]

Sanjay ,

Take this opportunity to learn ,all the relationships are there to calculate the Global heat transfer coefficient , just to consider no radiation and to calculate Nusselt number and ho based on free convection or forced convection .For hi , the work is already done.

 

For the benefit of others , shows us your spreadsheet .

 

Breizh

[sang]

Hi Breizh
 
Yes, I'll do it myself and update the spreadsheet soon and share with everyone.
 
Meanwhile I had a question which might appear silly,, but it is claimed to be correct by a reputed pipeline software developer company, so need to treat them seriously.
 
The question is: For a bare pipe like above, which is exposed to air, the Ho, should it be arrived at by summing the free convection H and forced convection H (assuming  wind v > 0.5 m/s)? I am under the impression that it's either natural or forced convection that should be considered, not both together? I have pored through several books and trawled the net but haven't found a source which says that the natural and forced H should be added together. Please help, all you heat transfer gurus.
 
Even Steve's s'sheet considers the maximum of both the terms but calls them both 'no wind' (see cells D103 and D104). I'm sure it's just a typo.
 
I'm unable to convince the tech support people that what they are doing is not quite correct - or are they?
 
Thanks in Advance
Sanjay

Edited by sang, 18 April 2013 - 01:58 AM.

[Steve Hall]

Yes, my sheet uses the greater of the wind or no wind calculation. With natural convection, the heat transfer changes the temperature of the air in the immediate vicinity of the pipe, and the consequent change in density causes the air to "naturally" move up or down relative to the pipe. There are different correlations for the cases of horizontal or vertical pipes since with a horizontal pipe the heat is whisked away by the air movement but with a vertical pipe the heat continues to travel along the pipe surface. With forced convection, the actual wind moves the air across the pipe and there is no additional heat transfer from natural movement (which is small relative to a measurable wind velocity). I use the greater of wind or no wind because I always want to be on the conservative side so I'll choose the stronger of the two correlations. From my perspective, the forced convection correlation must lose accuracy near the boundary point (no wind) and I'm happy to abandon it there in favor of the natural convection estimate.

 

There is no reason to sum the two results. The correlations already account for the convective forces.