Nice. Here's a page from 1999 about TEMA:
http://www.chemicalprocessing.com/web_first/cp.nsf/Contents/8625688C005A24978625\
690B0075362F?OpenDocument
(How to Compare Heat exchangers).
Just a minor observation: some process fluids are to be cooled-down. In those cases, contraction, not expansion is the problem.
Paul
> Mohamed,
>
> I have has to make a few assumptions about the type of exchanger you
are
> talking about. I hope I got it right.
>
> The heat exchanger is generally a long vessel, and the tubes and the
shell
> will have different temperatures (otherwise it would not be a heat
> exchanger) so obviously if you clamp the shell and the tubes together
at
> both ends they will put each other under strain. The hot component
(shell
> or tubes) will go into compression and the cold one into tension.
That is
> particularly bad for the tubes because they are often held in place
by
> frictional grip alone. There is no sense in putting those joints
under
> longitudinal stress. There is also no sense in putting the shell or
the
> tubes under unnecessary stress. That is solved by not connecting the
tube
> bundle to the shell at both ends. One end of the tube bundle is
allowed
> to move freely in the longitudinal direction within the shell. They
> normally call that end the FLOATING end. The other end is just about
> always fixed and I think it is this fixed end you are talking about..
This
> arrangement still has to be made to work however in that it must be
> possible to connect a pipe to the inlet of the tube side and also to
the
> outlet of the tube side. This can not be done on the floating end
because
> it is floating around within the shell. This is solved by dividing the
> tube bundle into two distinct sections and making the two pipe
connections
> on the fixed side of the tube bundle only. The fluid enters one half
of
> the tubes on the fixed side, flows to the floating end where it is
routed
> into the second half of the tubes. It then flows back to the fixed end
> where there is a second nozzle to discharge it. The fixed end is
normally
> called the channel because in section it has the shape of a channel.
This
> channel has the triple purpose of 1) clamping the tubeplate of the
tube
> bundle to the shell (thus fixing it) 2) dividing the tube bundle into
two
> so that the fluid can flow to and from the floating ends in separate
> sections. That is done by installing a divider plate in the channel.
3)
> providing a fixed place to install the inlet and outlet nozzles to the
> tube bundle. The inlet to the bundle is located on one side of the
divider
> plate, the outlet is on the other side.
> So you can see the channel end is the end where all the action takes
> place. Where the tube side nozzles are located, where the tube bundle
is
> fixed and obviously where the tube bundle can be removed. That is
simply
> how the exchanger is designed.
> Other contributors have mentioned the fixing of the exchanger to the
> foundation. That is also an issue and I would normally expect the
> exchanger to be fixed to the foundation at the channel end as well.
As
> you undoubtedly know, pipes expand when they get hot and one needs to
> control the movement of the points where they terminate. Generally it
is
> better for the end points to be fixed and not moving around stretching
and
> compressing the connected pipes. The channel has two nozzle attached
and
> at least one nozzle on the shell side will be close to the channel.
It is
> generally a good idea to fix that end of the shell to the foundation
and
> to allow the other end of the shell to slide over the foundation to
> accommodate thermal expansion. That arrangement normally places
minimum
> strain on three of the four pipes attached to the exchanger.
> The above is a description of a floating end shell and tube exchanger.
> There are a number of other arrangements that are common. If you can
lay
> hands on a copy of the TEMA standard (of the Tubular Exchanger
Manufacturers Association ) you will find details of just about every
arrangement in common
> commercial use.
Received on Wed Mar 24 19:41:00 2004
This archive was generated by hypermail 2.1.8 : Mon Oct 27 2008 - 20:24:03 EDT