21 replies to this topic

[sheiko]

Hello

The performance of one of our shell and tube heat exchanger has changed in a significant manner.

The service is crude oil in the tubes and diesel in the shell.
It has 1 shell pass and 4 tubes passes.
Tubes are equipped with Turbotal inserts.
Surface is 197 m2.

Before a shut down, performance was:
LMTD = 70C
F = 0.88
Duty = 1550 Gcal/h (heat balance closure = 1%)
Temperature cross = +14C (no cross).

During the shutdown, tube bundle have been replaced by a new one. New Turbotal as well.

After shutdown, in service, performances are:
LMTD = 50C
F = 0.34
Duty = 1800 Gcal/h (heat balance closure = 5%)
Temperature cross = -21C

Additionnal data:
C = crude
D = diesel
I = in
O = out
M = mass flowrate
Cp = specific heat

Before shutdown:
Tci = 152C
Tco = 205C
Tdi = 280C
Tdo = 219C
Mc = 47.5 t/h
Md = 37.9 t/h
CPc = 0.62
CPd = 0.68

After shutdown:
Tci = 143C
Tco = 207C
Tdi = 272C
Tdo = 186C
Mc = 45.7 t/h
Md = 32.8 t/h
CPc = 0.62
CPd = 0.66

What does this mean in your opinion?

Edited by sheiko, 12 November 2014 - 10:30 PM.

[srfish]

Wow!!. To achieve a real LMTD correction factor "after shutdown" of only 0.34 is amazing. Normally thermal design engineers do not use a LMTD correction factor less than 0.75 to 0.80. In my experience, the lowest real life correction factor I have seen is 0.66.

[sheiko]

Ok thanks
So maybe some TI measurements are wrong...I will check.
Maybe heat losses on some thermowells...?

Edited by sheiko, 24 October 2014 - 02:55 PM.

[Bobby Strain]

You should calculate the U value for before and after. That is the true comparison of performance.

 

Bobby

[sheiko]

Another question please:
Is it possible that the modelization be wrong?

I mean, if I use the F-formula for a 2 shell passes - 4 tube passe heat exchanger, F increases.

In other words, can we model a real 1 shell pass - 4 tube passes heat exchanger, by 2 1 shell pass -2 tube passes heat echangers in series?

Edited by sheiko, 12 November 2014 - 10:32 PM.

[Bobby Strain]

So, why don't you look at the exchanger drawings? And the original datasheet?

 

Bobby

[breizh]

Good day Sheiko ,

My calculation is showing different values for F : before #0.9 and after the Change #0.8 leading to heat transfer coefficient increasing from 125 kcal/h m2C to 219 kacl/h m2 C .

Hope this helps

Breizh

revised #14

Edited by breizh, 26 October 2014 - 04:40 AM.

[PingPong]

In other words, can we model a real 1 shell pass - 4 tube passes heat exchanger, by 2 1 shell pass -2 tube passes heat echangers?

No.

 

You have to find out what kind of exchanger you really have. What TEMA type? How many shells? et cetera.

 

In any case: the new tubes with new Turbotals seem to have improved the performance (U value) of the exchanger considerably.

[Bobby Strain]

Breizh,

      I get some different result for the clean performance. Is my calculation for F not correct?

Bobby

Attached Files

•  HX Performance.rtf   3.3MB   68 downloads

[srfish]

Bobby,

 

I didn't check all of your F factors but the 1st two are correct.

[PingPong]

Using my own (decades old and proven) spreadsheet I find exactly the same values for all F's as Bobby.

[sheiko]

In any case: the new tubes with new Turbotals seem to have improved the performance (U value) of the exchanger considerably.

Hello PingPong,

I have quickly calculated the thermal effectiveness on both sides, and below is what I have found:

- Crude side (tubes side): effectiveness increases from 41% to 50% (+9%).

- Diesel side (shell side): effectiveness increases from 48% to 67% (+19%).

So I believe the new Turbotal inserts (inside tubes) don't totally explain the apparent improvement.

These quick calculations seem to indicate that there is something unusual on the shell (diesel) side (Diesel outlet temperature namely)...

Edited by sheiko, 24 November 2014 - 08:54 AM.

[breizh]

Bobby ,Sheiko
Yes you are right , doing calculation instead of using a chart ( wrong one I guess) I found similar result .

F before: 0.8816
F after : 0.3589

Note : I've attached a paper about Turbotal , you may find your answer .

Hope this helps


Breizh

Edited by breizh, 26 October 2014 - 05:17 AM.

[PingPong]

These quick calculations seem to indicate that there is something unusual on the diesel side (Diesel outlet temperature namely)...

I don't think  there is something wrong with the diesel outlet temperature. If there were, then the calculation would not show that diesel and crude duties are practically the same.

Diesel effectiveness increased more than that of crude because diesel/crude flowrate ratio is now lower than in the past.

 

Replacing the old fouled and oxidised/corroded tubes in the exchanger by clean new tubes is probably the main cause of the increase in performance, due to increase in U.

Edited by PingPong, 26 October 2014 - 05:08 AM.

[Bobby Strain]

Would anyone like to have my handy little calculator? If so, you can download it here.

 

http://www.mediafire...fectiveness.exe

 

Bobby

Edited by Bobby Strain, 26 October 2014 - 10:34 AM.

[sheiko]

I understand from what is said here that the lower the factor "F", the better the performances of an existing exchanger (which sound strange to me by the way)...

Can one trend the "F" to follow the performances of an existing heat exchanger?

Edited by sheiko, 22 November 2014 - 11:14 PM.

[Bobby Strain]

Best to track the overall heat transfer coefficient. Or UA so you need not bother about the area. Calculating this value requires calculating F. You should also track flow and duty because U will vary with flow. And, you should find some good reference on heat exchangers.

 

Bobby

[sheiko]

Thanks a lot.

I do understand the increase of U (better performance due to new tubes bundle).

But I still don't understand the decrease of the LMTD correction factor (F).

Could excess surface be the explanation?
(ie. excess surface => high cold outlet temperature => low approach => low LMTD correction factor)

Edited by sheiko, 24 November 2014 - 08:52 AM.

[PingPong]

F decreases because U increased at similar flowrates and inlet temperatures.

Value of F is the result of the exchanger performance.

 

With the new U the exchanger is oversized, but that was not so for the old U.

[Bobby Strain]

No, F is a function of the design. And, of course, the required surface area for a design duty is inversely proportional to F. Stop worrying about it. Just apply it as a rule. Just like every day has 24 hours.

 

Bobby

[PingPong]

It is common practice that a new exchanger is designed such that F > 0.8 as designing for a lower F means that a lot of extra area would be needed for little extra duty.

 

For an existing exchanger F will be whatever it will become in the given situation.

If all of a sudden the U increases a lot, then the LMTD and F will decrease a lot, and the duty increases somewhat. In this particular case that is nothing to worry about.

[sheiko]

Thanks Bobby and PingPong.