Odd heat exchanger performance

[rjw57]

Hi everyone. I am in the midst of testing a tube and shell gas cooler which is giving me (as noted in the subject) odd performance. Background - unit is liquid (unnamed) on shell side, gas on tube side. Gas on tube side is about 80degC dewpoint and arrives at the cooler inlet at 270degC and leaves at a variety of temperatures, some of which make no sense. Coolant is at 65-70 degC (below gas dewpoint). Coolant boiling point is ~230degC at operating pressure. The unit is a cocurrent setup and the cooler is unbaffled. The outlet gas temperature will rise as the coolant flow is increased (odd!!!) if the coolant flow was very low to start with. When increasing th flowrate thru the shell side, a certain flowrate is reached where the gas temperature will dive. After this point is reached, a decrease in flow rate will increase the outlet gas temperature. The coolant temperature (bulk at outlet of cooler) does not vary by more than 20degC above the inlet of 65-70. Can anyone dream up a reason for the "hysterisis" of the system. I have some odd theories, but none are satisfying. Has anyone ever seen this type of phenomena?

 

[235zzzo]

Hello RJW

Indeed, an interesting case!
In order to help you, i need to go deeper. You have to provide more technical and testing information, such as:

- What is the gas and liquid in presence?. Its thermodynamic properties, where to find?
- Need to know the table/graphic with the temperature behaviour in correlation to the input temperatures and flowrates.
- Any other information, which you think, can have some role/importance.

My guess, at this very preliminar point, is related with the gas and how it performs the gas-liquid phase change, as well the conductance coefficient behaviour experienced at the heat-exchanger in those particular conditions you have.
Hope this can be of some help.
(i'll be absent of the forum for two weeks.)
zzzo

 

[butelja]

Is the liquid condensate from the gas viscous in nature? I've seen similar performance with coolers handling viscous products. They will "film up" and performance will fall way off if the cooling media is too cold. If this is the case, a circulating pump can be used to create a "tempered water" cooling system.

 

[rjw57]

235zzzo and butelja -

After running some further tests, I've discovered that my problem was tube side condensation. Not a viscous fluid as you thought butelja. It is water on stainless tubes, in laminar flow. Raised coolant temp and control issues disappeared. Not sure though if the new coolant temp is acceptable, but if not, not certain how to lick this problem. Thanks for your help...
rjw57

 

[ChasBean1]

rjw: Liquid coolant flow increases. Temperature of the liquid is below the dewpoint of the gas. With the increased liquid flow, condensation within the tubes happens to a point where the it blocks some of the 270°C gas from establishing direct contact with the heat exchange surfaces due to liquid film coating within the tubes. Water acts as an insulator, coating a large portion of the inside of the HX tubes at higher liquid flows.

This is the reason, but you've solved your problem (congrats!). Avoid this in the future by lowering the dewpoint of the gas or raising the temperature of the liquid coolant. -CB

 

[rjw57]

ChasBean1 -
At the time, the problem did not seem quite so simple. The fact that there was water condensing tube side seemed strange since the outlet gas temperatures were well above the dewpoint temperature (~20-40 degC). I realize that the boundary layer temperature is lower than the bulk stream temperature and suspected condensation from the beginning, but had a hard time imagining steady state tube side condensation with these higher than dewpoint bulk stream HEX outlet temperatures. Also, the gas dewpoint was theoretically only 2-5 degC lower than the coolant temps we were using (never actually measured - using Aspen data). The fact is that we attempted to vary both shell and tube side flow rates during these tests (with the original coolant temp) in order to attempt to correct the problem, so that at some points I am certain that shell side boiling was occurring (low shell side flows) when we were attempting to obtain higher outlet gas temperatures. This further clouded the issue. BTW, the temperature of coolant was supposedly fixed in stone, but now is being rethought because of the controllability we were able to obtain. I will now forever have coolant versus gas dewpoint temperature forever etched in my mind...

 

[ChasBean1]

rjw, not a no-brainer. Excellent post & interesting problem. Good luck, CB