HX in series (AES vs AFS) - configuration

[02164]

This is related to possible arrangement of exchangers for series configuration, their pros and cons.

I am designing a Feed bottom exchanger with cold side approach of 40°F.
Temperature cross exists and hence i opted 2 shells in series .
Initially i designed AES (E shell) and it took a considerable amount of time, since i faced a reverse heat flow in one of the shells. Finally i managed to get rid of it by increasing the tube passes and tuning relevant parameters.

But later , i tried AFS in series and found that there was no problem of reverse heat flow. Furthermore i was able to reduce required heat transfer by about 25% compared to AES in series.

My question :
Most reference show AES in series, and i havent come across AFS in series.
Can somebody advise me if AFS in series is acceptable.
Does this configuration have any kind of drawback, or problems.

Regards,

[cdehn]

The first thing you do is to determine the number of shell passes required for the servive using a temperature step diagram (Q on the X axis and Temperature on the Y axis). This heat load diagram must also be done on a weighted temperature difference basis as opposed to using the cold and hot end approach temperatures (not all heating and cooling curves are linear as one may have condensation or vaporization). You step off on the diagram to determine the number of shell passes (sometimes called the number of HTUs or Heat Transfer Units) required for the 4 temperatures you have for a specific duty (like a Mcabe Thiele diagram for determining the number of theoretical trays in distillation). An AFS shell has two shell passes and equates to the same as installing two separate shells in series (hence one AFS shell is similar to having two AES shells in series). The F correction factor will always be 1 for an AFS shell (true countercurrent flow). What limits you in an F shell in a lot of cases is the pressure differential allowance across the longitudinal baffle. One should limit the pressure drop for an AFS shell to a maximum of 10 psi. If you have a high pressure drop, one will get vibration on the longitudinal baffle. The next thing that you need to take care of when using an AFS shell is thermal leakage through the longitudinal baffle (from one shell pass to the other by conduction - it is not small). This is dealt with in reality by using two steel longitudinal baffles with an insulating layer in between (it obviously is fully welded up around the periphery). The next thing to take care of is to provide seal strips i.e. lamiflex or equivalent. One must put these on the longitudinal baffle (see wall wiper type).

The major disadvantage with AES shells is that you are limited by the F correction factor (the F correction factor is always 1 for an AFS shell as you have true counter current flow) and one looses a portion of the shell internal cross-sectional area so as to accommodate the internal floating head that is bolted on the opposed tube sheet. One can loose in the order of 3 inches around the bundle (clearance distance requirement so as to enable the bundle to be pulled out). This hence results in a significant reduction in the number of tubes that one can install. For the equivalent AFS shell diameter, one can install a lot more tubes and the $/ft2 heat transfer area cost will be much lower than the AES shell.