7 replies to this topic

[fatimah]

Hi all

i have to design BEM type (880 mm x 3050 mm), which tube side is liquid phase. I have to install 6x12" nozzle at inlet & outlet channel side to cater the liquid velocity to be less than 8 ft/s. the viscosity of the liquid is less than 1 cP.

is there any other way i can do to decrease the no of nozzle & at the same time the expected tube side velocity can be met ?(the channel side is long with 3 nozzles install in series, top & bottom side)

thanks for your help.

Edited by fatimah, 06 January 2011 - 01:06 AM.

[DB Shah]

Dear Fatimah

If I understand your input- you intend to have 6 nozzles of 12" dia each and with velocity <8fps in each nozzle.
With this data roughly you are handling about 3500 M3/hr of liquid in tube side & your exchanger size is roughly 150 to 200 M2.
If I use this figure & considering even single tube pass max tubes(19.05 mm OD) in the shell will be less than 1000. Even if I consider 1000 tubes the velocity in tubes will be >12 fps and pressure drop too high.

I will request you to provide sufficient data of flow and configuration.

[Art Montemayor]

Fatimah:

It is very difficult too understand your original post. It isn’t because of your use of the English language; it is due to how you describe what you are proposing to do. A picture is worth a thousand words. I recommend you visit the thread on designing and fabricating a BEM heat exchanger in our Forums: http://www.cheresour...heat-exchanger/ .

Copy the sketch(s) I have made in the uploaded Workbook titled, Designing a BEM Heat ExchangerRev1.xls and use this as a means of describing what type of heat exchanger you propose. There can be more than one tube pass in a standard BEM. Which one do you mean? Also indicate the diameter, length, heat transfer area, and size & pitch of the condenser tubes used.

Indicate the flows and conditions within the flow arrows at each of the four nozzles. This is the normal and conventional manner of describing a heat exchanger and leaves nothing to doubt and question regarding what you are proposing.

Use the same Excel workbook as your calculation pad and show your heat balance and end temperature calculations – very much as I have done. Don’t do this because I say so; it is not my method. Don Q. Kern demonstrated this over 60 years ago and it is simple, practical, and communicates accurately.

DB Shah is correct in trying to interpret your information but it is leading him to ridiculous results. Although I don’t have my Ludwig tube layout information here with me, I don’t believe you can “push” 3,500 m³/hr through tubes that are within a shell diameter of only 880 mm – even in one pass. I also don’t think you can install 1,000 tubes in such a small diameter. And why do you limit the shell (& tube) length to only 3 meters?

Also, be specific about what “velocity” you are referring to – the nozzle velocity or the internal, average fluid velocity in the tubes? Are you concerned with tube impingement and vibration, or are you interested in a good convective heat transfer coefficient?

I think I know and can guess what you are proposing. But why should I, DB Shah, and the rest of our members have to do all the guessing, pondering, estimating, calculating, sketching, and speculating about what you mean when you can more accurately devote your effort to communicate the correct information to us so that we can be of better help to you?

I don't think a BEM type has "channels"; I believe it has "bonnets" as heads.

Await your reply.

[fatimah]

Hi DB Shah & Mr Art

i apologize as i did not provide accurate information for my problem, i work with design company & my design based on our client requirement.

below are information regarding the unit;

STHE type = BEM, i tube pass, 880 mm x 3505 mm

Shell = liquid

Tube = gas
gas MW = 19
inlet pressure = 1004 psia

Calculated HT area (HTRI) = 141.684 m2
pitch = 1.25 with 25.4 mm tube OD, A-179
(pitch, tube OD, & tube length is client requirement based on petronas standard)

my concern is i have to install 6x12" nozzle at the channel side, both at inlet & outlet to cater the tube side nozzle velocity, which is in gas phase (my apology again). as i refer to Ludwig chart for nozzle velocity, it should be below 42 ft/s. in order to meet the required velocity, i have to install 6x12" nozzle for each inlet & outlet channel side.

is there any other way i can do to decrease the no of tube side nozzle, instead of using bigger nozzle? Thanks for your help.

& for Mr Art, as from my understanding, bonnet is the channel head. nozzle will be installed at the channel pipe & about 75 mm from the nozzle, the channel head (which is bonnet) will be welded.

Attached Files

•  BEM.xlsx   11.12KB   32 downloads

Edited by fatimah, 06 January 2011 - 09:29 PM.

[DB Shah]

Dear Fatimah,

Your initial & recent post have too many contradictions( fluid allocation shell/tube, tube length 3505 - 3050 mm), but considering your latest post I will comment further-
1. Check tube side velocity in HTRI, is it too high crossing erosional velocity limits?
2. Check Uservice, with such a high velocity your Uservice is too low (<100), are you getting Area ratio very high, ideally should be around 110~ 120% meaning 20% excess area. But in your case it seems to be too high.
3. If you optimize above values you may be able to design with a single 30" tube inlet nozzle(parallel to tube orientation) instead of 6 x 12" nozzles.

If you can upload tema sheet of HTRI, may be we can guide better.

Art-
Yes, 1000 is too high a tube count for 800 shell ID (almost double), I intentionly took the count to emphasis that even with 1000 tubes she is too high on velocity. However as now the fluid allocation has changed (liq in shell) it does not hold good now.

I guess she is restricting tube length to maintain pressure drop, with the configuration given I estimate ~0.2 to 0.3 Bar of pressure drop with a velocity of 25 m/s in tube bundle, any extension of tube length will increase the pressure drop. Also she is very high in area ratio and has still lot of area available for heat transfer.

With all optimization I still expect an unsual configuration, see the heat sink flow (gas) 100 times more than the liquid flow, with heat duty of <0.3 Gcal/hr, LMTD of ~20 and velocity of 25 m/s uD should be too good > 200 kcal/hr-m2C. Area required should be in the range of 80 m2. And with such a high flow tube count will be also very high thus restricting tube length for the given duty.

[Art Montemayor]

Fatimah:

Please refer to my comments and notes in the attached revised workbook.

I hope my comments help out.

Attached Files

•  BEM-Rev1.xlsx   23.36KB   34 downloads

[DB Shah]

Art-
I referred your comments, I will like to further add
Unusually high heat sink- yes that is why designer will land up in a something like near to square shaped exchanger. Tube side flow is 100 times shell side flow.

Tube length-
As I stated in my earlier reply, Fatimah cannot increase tube length due to pressure drop limitation.

Tube pass-
Again pressure drop limitation. She is at the peak of velocity, need to reduce velocity. No scope of multiple tube pass.

6"x12" Nozzle-
I think there is some misunderstanding, as I understand she needs 6 nos of nozzles of 12 inch dia each to restrict 42 fps. (Look at the flow 342 kg/s = 1231200 kg/hr ~ 23230 m3/hr (Approx density 53 kg/m3)

Exchanger adequacy
The exchanger with 142 m2 will certainly work, no doubt about it. Considering gas cp of ~ 0.5 kcal/kg-C
Q = 1231200 * 0.5 *(27.9-27.4)=0.3 gcal/hr. LMTD ~ 20 Uservice = 0.3*10^6/(142*20) = 108 kcal/m2hrC.

Uservice of 108 is very conservative fig with above given velocity. I will expect > 200 ud & hence 200/108 is ~ 100% over design.

Fatimah:
The exchanger with 142 m2 will certainly work, on the other hand I see an margin to still reduce the area by reducing tube length.
You have raised vel & rohv2 issue in your ealier posts. I will not bring it in again. Just chk your velocity & increse shell dia/reduce tube length, provide adequate tube side inlet axial noz as drawn in Arts sheet. I expect >32" nozzle.

[fatimah]

Hi DB & Mr Art

For this unit, the calculated tube side velocity is 26 m/s. Tube side erosional velocity limit is always respected (HTRI will provide warning message if the velocity exceeded)
over design for this case is 17%
tubecount = 631
allowable pressure drop (tube side) = less than 1 bar, no decrease for shell ID can be done further.

i apologize as no upload from HTRI is allowed for this case.
i tried to change the nozzle orientation into axial, & 1 x 28" nozzle fits the design. and for the flow distribution, HTRI only provide shell fluid flow distribution, anyway your point is good if we can highlight this to our support.

You also stated about checking the tube side velocity which can form centralized flow. At design stage, Is there any where we can check this condition from being happened? As from Mr Art, he proposed internal flow distributor which I believe same to spray nozzle. The bigger the nozzle, there will be less chance of centralized flow occurs, but do we have any rule of thumbs how small a nozzle can bring this problem?