TEMA Type BFU Shell and Tube Heat Exchanger

[uglyspec]

I understand that a F shell TEMA type heat exchanger is specific to thermal designs where there is a temperature cross. Is it possibly that the temperature difference between the shell side inlet and outlet temperature be to great for this type of design? For example: Lets say my fluid temperature on the shell side inlet is 100 deg. F and on the outlet it is 900 deg F. With fluids and temperatures separated by the longitudinal baffle in the shell wouldn't that cause the upper half of the shell to thermally grow at a different rate than the lower half thus creating a moment on the shell flange that connects to the tubesheet?

The moment created at this bolted connection could potentially cause a leak at that joint.

Thanks!

 

[EmmanuelTop]

See article:

https://s3-ap-southeast-1.amazonaws.com/erbuc/files/5634_7dfaae9c-d0a1-4130-82ee-0750ef45da26.pdf

It discusses some of the merits of F-shell exchanger.

Your application does look extreme, with 800 deg (F?) temperature difference of the shell-side fluid. Typical temperature difference limitations I have come across are in the range 300-450 degF but thermal stress analysis should confirm the right figures for this particular application. There are some tips from Gulley's:

http://www.gulleyassociates.com/news2002.htm#Dec02

http://www.gulleyassociates.com/news.htm#Dec96

I doubt one single exchanger can be designed to manage this Delta Temperature.

Dejan IVANOVIC
Process Engineer, MSChE

 

[racookpe1978]

Agree with the above, but for a different reason than just the delta t.

Physically, the kind of design you seem to be thinking of would be a sheel and tube, U-bend tubes since you seem to be thinking of the differential expansion between top and bottom of the tubesheet stressed at the (horizontal) dividing plate. That can be avoided by a once-through HX so the colder outlet end is at the far end from the (very hot!) inlet end.

But if it were a u-bend tube, or a turn-around head at the far end, why would either of those "help" the original design intent? Your second process fluid - it can't be cooling water since you want to increase shell side temperature from 100 inlet to 900 outlet, but might be steam, right? But them you'd get a real nasty transition points inside the tubes as the water condenses at various points along various tubes randomly. Very, very hard to control for your process, very hard to control for the steam flow too. Seems like your end head would be mixing some tubes with steam, some with water mixes, some with all water and all flashing back and forth. And U-tubes would be even worse. A non-condensing liquid in the tubes might work.

Regardless, the extreme temperature difference across the shell side requires baffles or multi-shell side entry points to "even out" the delta T between both fluids. Try using two or three heat exchangers in series instead.

 

[racookpe1978]

This, for example, (from google images for shell and tube HX's) shows a gas-to-gas double-ended HX... NOT a steam system! Just nice, hot gasses on both sides. NO divider plate on the single end.

https://www.google.com/search?hl=en...iVEyHM:;6AYlyNauFKMOwM:&imgrc=7v4q4H5RiVEyHM:

http://www.essentialchemicalindustry.org/processes/chemical-reactors.html