No of Passes

[arun17]

Dear friends,
I would like to know how to decide about the no. of passes in a shell and tube type heat exchanger.
Thanks
"Knowledge is power"

 

[Scratch2002]

What do you want to know? Your number of passes is dependent on the length of pipe you need to affect your temperature difference. I can help you if you tell me a little more aout the problem.
Mass flow rate, required temperature difference, fluid?

 

[TD2K]

The number of tube passes is essentially set to keep tubeside velocities in the desired range. The design process is an interative one, estimate the area required using a starting overall heat transfer coefficent, check tube side velocities, set number of passes for the desired velocity, check effect on the temperature difference, check effect now on heat transfer coefficient, check whether dP through tubes is okay (and if not, go to the next larger tube size) and go around the loop till you get a design that works for you. You have a similar approach on the shell for baffle arrangements, cuts, etc.

This is a VERY abbreviated approach. I've got a much better writeup at home in a course I took on heat exchanger design, I'm just running off a (rusty) memory. Typically, you want tube side velocities in the 3 ft/sec to 8 ft/sec for liquids (3 is on the low side) to keep heat transfer coefficients high and to minimize fouling tendencies.

Okay, I've left myself open for all sorts of improvements in this post, fire when ready everyone ;-)

 

[abeltio]

TD2K, there is nothing wrong with your posting... and the improvements are the very reason for these fori (if I recall my latin correctly)... only egotistic maniacs would pretend to post the ultimate note on a subject...
May I add to your considerations:
1. Layout: some times your piping does need inlet/outlet on the same side for maintenance reasons...so an even number of tube passes is called for.
2. U Tubes: cannot have anything but even number of passes (duh!)
Regarding the fluid velocities inside the tubes it all depends on the materials involved... non-ferrous materials and carbon steel your numbers are ok, stainles steel you can go as high as 11 ft/sec, titanium even 20ft/sec
I do not pretend to disagree with your posting, let's consider it another point of view and as just stating what I've learned as the main reason (according to the tube manufacturers literature) to limit the velocity: the erosion caused by the flow, specially at the entrance of the tubes. Our good friend Kern (Process Heat Transfer) considers these entrance losses.
Summarizing:
Usually piping is better suited to have both nozzles on the same side of the exchanger, therefore the calculations should aim to prove that an extra tube pass to have even an even number does not adversely affect the pressure drop budget and does not fall into the non-recommended velocity ranges by the tube manufacturer.
Hth
a.

 

[arun17]

Dear friends,
Actually I am trying to estimate main parameters for a 100 kW water/water Shell & tube type heat exchanger. I am a novice in this field. I started with a single pass heat exchanger with hot water on tube side and cold water on shell side. I used 30 tubes of 3/4", 1.5m long, and calculated the Re, Nu and heat transfer coeff on tube side as well as on shell side. But it seems that the Re on tube side is less, hence Nu(Nusselt No. = 0.023*Re**.8*Pr**.33)is also less, which leads to a low overall heat transfer coeff. So the heat trnsferred through the surface available is less. I tried with a higher no. of passes with less no. of tubes per pass, to increase Re on tube side. Basically I want to know whether one should assume an overall heattransfer coeff. first and then calculate the no of tubes and passes or first select the no of tubes and passes and then calculate the Overall Heat trf. coeff. It seems the pressure drop is not much even if I go to a higher no of passes and less no of tubes, since I have a good pr. head available to me.
Hoping for the response
Arun Agarwal "Knowledge is power"

 

[TD2K]

Doesn't really make any difference where you start Arun. You need to start off with an assumption and keep working through the design until the performance of your unit matches what you need it to do within your exchanger's constraints (velocity limits, pressure drops, etc).

Those are pretty short tubes, why that length? Shell and tube exchanger tubes are usually quite a bit longer than 1.5 m. Your most efficient exchanger (minimum area) will be the one in general with the longest tubes.

 

[TD2K]

Arun, I suggest the following approach. Caution, I don't design heat exchangers. I'm a process engineer but I develop preliminary areas when the cost people need a ft2 to estimate costs and I've audited some exchanger courses and played around with some heat exchanger software (HTC-STX). So, I qualify as a dangerous individual with a little knowledge.

You have a duty so the flow rate of your hot water and its temperatures are fixed. I'm assuming you've set the cold water side temperature rise so you also know that flow rate. With the temperatures, you have your temperature difference across the exchanger. Since you don't have a configuration yet, you set the F factor to 1.0.

Select a typical overheat heat transfer coefficient for water/water S&T. Perry's Handbook of chemical engineering has them, TEMA has them as does the GPSA Engineering data books and lots of other reference material. You have a duty, a temperature difference and an assumed heat transfer coefficient so you can calculate the area.

3/4" tubes are a reasonable choice for water (so are 1/2&quot. Calculate the area for one tube (20 ft or 6.1 m is a common length). The number of tubes is determined from the total area you just calculated.

Decide if you want a single pass or 2 pass (depending on your piping arrangement) and then calculate the velocity through the tubes. Set the number of passes to give you a reasonable tube side velocity.

Do a similar exercise on the shell side to set baffle spacing.

Check the tube side and shell side pressure drops that they are reasonable and within your limits. Adjust design as needed.

Calculate the shell side and tube side heat transfer coefficients, add in your fouling factor to calculate the overall Uo. Go back through this exercise based on the new area you calculate you need. When the starting heat transfer coefficient at the start of a cycle is sufficiently close to the calculated overall heat transfer coefficient and the dPs for your tube and shell side area acceptable, you have a workable design.

 

[abeltio]

Arun,
Are you planning to design and build a shell&tube heat exchanger?

Other considerations may include: plate type HE... they are very well suited as water/water HE.

Depending on the intended service, and the requirements of the owner (e.g. an oil refinery is much more stringent than a merchant ship)

Do you have TEMA requirements? (e.g.: C vs R?...some customers will require R class regardless of the service...because they may want to re-use the HE for a different service later on).

I've custom designed and built S&T HE for quite a while and always used 5/8" tubes (for water/water)... found those tubes to have the best compromise between pressure drop and heat transfer surface.

Using the longest possible tube does have advantages, the thickness of the shell increases with diameter, and so does the number of holes on the tubesheets which may become very expensive... so you should try to have the highest pressure on the tube side, also the highest pressure should be for the fluid that cannot be contaminated.
e.g. if you have motor cooling water and river water...the pressure of the motor cooling water should be higher so in the event of a leak you leak cooling water into the river water and not the other way around...in this case, I would put the cooling water in the shell side and the river water on the tube side (dirtier fluid on the tube side)...this seems to contradict what I stated before about the highest pressure going through the tubes... but the design is a trade off... and for HE's it is better to have the dirtier fluid through the tubes, specially for low to moderate pressures (up to 150 psig)

Also, as you have water water... what I would do is select heat transfer coeff of h=1000 Btu/ft^2...etc. for each stream (shell and tube) from the TEMA stds select the appropriate fouling resistances (the bad guys in your case), tube wall resistance and estimate the total U per TEMA Stds.
do a first estimate of your surface as S = Q/(U*LMTD)

With that first estimate for your surface... select the length and number of tubes considering the available space and piping arrangement.
Check the resulting U (use Kern)
Check the pressure drop both sides (Use Process Heat Transfer by D. Kern).
On the shell side playing with the # and pitch of the baffles will allow you to get the h and the dp required.

Usually the limit for the pressure drop to is between 7 to 10 psid.

Notes
Check owner's spec for:
1. materials
2. cathodic protection
3. required tests
4. spare parts required
5. reserve surface
6. connections for instrumentation

HTH
Saludos
a.

 

[MJCronin]

arun17,

You could always be a sleazeball and let others, namely the S&T fabricators, (who have licensed software) do the work for you.....

If you have the HX duty, flowrates, materials, tube sizes, operating pressures and maximum dimensions, you can have a fabricator "do your work" for you...... sure, it is a bit sleazy...... but it has always worked for me.....

TEMA multi-pass heat exchangers are typically required when the thermal approach is too small or there are space ( read "length&quot limitations to the design.

Remember, a long thin "pencil" type S&T heat exchanger is always cheaper than one that is "fat and short" ( the reason is that the tubesheet costs strongly affect the overal cost)

Consider use of a "plate-frame" type heat exchanger if the approach is too small ( see the Alfa-Laval website for more information)

Good Luck !!!!

MJC