partial condenser outlet: gas and liquid phase temperatures different

[roland6939]

Hi,
i am a poor little french engineer lost in this country of senior process engineers !
My problem (actually not a ral problem, rather a question) is that following.
I have a partial condenser on top on a distillation column. Distillate is in gas phase. I have temperature measurement on gas oullet (distillate) and liquid outlet (reflux).
Consensation is performed on tube side. The heat exchanger is a vertical BEM. At the bottom of the condenser, after the tube plate, gas and liquid phase are separated.
The temperature of liquid phase is relatively different from the temperature in gas phase. It is obvious that thermal losses or measurement accuracy could explain this difference. But all the equipment is pretty well insulated, both measurements are pretty close and the dynamic response of both thermometers should be the same. The relative difference between the two measurements is more than 10%, which a priori is much higher than standard accuracy for these equipment. Furthermore, both thermometers have been checked recently without (accuracy <1%).
I wonder wether the liquid outlet could be undercooled which could explain the temperature difference.
Did you experience such a "problemn"? Do you think i am wrong to envisage the possibility of liquid subcooling?
Thank your for your answers!
Roland

 

[25362]

Just a thought: a vertical BEM unit may have fluid maldistribution, especially when the inlet stream (whether it contains liquid droplets or not) enters the channel head horizontally.

 

[Montemayor]

It is very possible for a vertical, one-pass downflow in tubeside, partial condenser to yield a 2-phase product that is not in equilibrium. In fact, I would state that this is more than possible if the designer, fabricator, and operator is/are not experienced in creating or handling saturated fluids. This follows the same line of thought/experience that 25362 expresses: in this type of heat transfer, the final result depends primarily on efficient fluids' distribution and handling.

The fact that the condensation is being done in a vertical unit immediately identifies the tubeside flow as ONE PASS. This, of course, means you only have one shot at condensing and establishing equilibrium. The downward tube side flow also means that there is little (if any) contact flow taking place between the condensed liquid and the exiting, uncondensed vapors. This last action is sorely needed for the establishment of a good, homogeneous equilibrium - something akin to a distillation tray - which is what is being attempted in order to fix a common temperature between exiting liquid and vapor. There are many possibilities that can be hampering the establishment of an exit equilibrium. A few are:

1) Bad fluid distribution - on both tube and shellside;
2) Thermally undersized condenser (too little heat transfer area) or physically oversized/undersized tubes (large diameter/short length);
3) Bad design of vapor/liquid separation in the bottom channel head; this is a frequent mistake because a normal channel design for a horizontal unit is used instead of designing the specific separator needed for this application.

For a resolution, this problem requires process design details as well as mechanical design drawings - both of which are beyond the scope of this Forum. Perhaps additional detailed process design info can shed more light - fluids' identies, fluids' flow rates, velocities, Reynolds Numbers, cooling media, temperatures, pressures/pressure drops. Mechanical details like tube size, tube length, shell baffle distribution, channel separator size, etc. would certainly help to analyze what is happening.

Right now, all we can do is speculate - and guess.

 

[katmar]

I believe that in a vertical, condensation-in-tube, single pass condenser the bulk liquid exiting will always be colder than the bulk gas. As the gas flows down the tubes and starts forming liquid on the inside of the tube the film of liquid will grow thicker as it goes further down.

There will be a temperature gradient through this film of liquid, i.e. it will be colder where is in contact with the tube than where it is contact with the gas. The gas might be close to equilibrium with the (hotter) surface of the liquid that it is in contact with, but once the liquid is separated from the gas and its temperature is measured as a bulk, the temperature will be somewhere between those of the two surfaces which were in contact with the tube and gas respectively. The average liquid temperature will therefore be lower than the temperature of the liquid film which the gas was in contact with.

The extent of the differences between the temperatures will of course depend on all the factors listed By Montemayor.

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[25362]

True verticality of the unit is assumed.

 

[25362]

Although Katmar's is an interesting and logical approach, it assumes segregated (or stratified) flow inside the vertical condenser tubes. The two-phase flow regime may be changing all along as more liquid is formed. I haven't seen studies on this subject to substantiate any proposition.

 

[CJKruger]

25362,

In two phase down flow there is an extremely strong tendency for annular flow. Thus, Katmar's theory seems reasonable.

I have also seen significant temperature differences between the vapor and liquid temperatures from the shell side of a horizontal condenser.

 

[Montemayor]

Every time I have applied or worked with a vertical, downflow tubeside, partial condenser I have identified a ridiculously small tubeside Reynolds Number and, as a consequence, a resulting poor & laminar film heat transfer coefficient. As Harvey suspects, this is very probably stratified. Although we can't prove this and much less see it, the Reynolds Number indicates that it is going in that direction.

Additionally, from a basic shell & tube mechanical design, one immediately is driven to suspect this when the fact that we have one, sole pass through the tubeside is confronted. I have worked with this problem before and it is a hairy one. The answer to efficient, turbulent tubeside flow is a high Reynolds (high velocity, resulting from multi-passes) Number. In order to avoid multi-passes in a 2-phase environment and all of its problems, one is driven to try smaller tubes - which often is not permitted due to standardization on 3/4" or 1". The only option left is either splitting the incoming flow prior to entering the inlet bonnet or making the exchanger smaller in diameter and longer. The latter option has given me ridiculously long lengths in the past. I've never had the nerve to split the incoming stream (& the inlet bonnet) with "equal" cross-sectional areas and wishing that multiple, parallel streams will enter their respective passes with equal flow rates. I don't think I could fool anyone into believing that it could be done - much less sustained.

I've always seen the exit liquid colder than the exit vapor in this type of condenser. I would like to hear about other person's experiences and if they were able to establish a near-equilibium condition between both exit streams.

 

[25362]

Assuming this to be a partial condenser with a large vapor volume, one can rationalize the r[&eacute;]gime is shear-dominated non-stratified (probably slug-type) rather than a gravity-dominated stratified flow. As already said, this is just speculation.

 

[moltenmetal]

Poor condensate disengagement from a cooling surface naturally tends to produce a subcooled condensate. A vertical BEM with partial condensation happening inside the tubes would seem to be a perfect example of this phenomenon.

Even though the sub-cooling and liquid film "insulation" tend to reduce the U value you observe, tubeside condensation is still usually (by far) the most economical option if you need to make the exchange surface from exotic material.

 

[AndreChE]

Please check the condenser capacity and real flows.

If you have much higher capacity in the condenser than the flowrates you are experiencing, you may have normal condensation and subcooling due to the extra capacity.

Please increase load to the column and check operation results against HE datasheet.

This is quite common in thermosyphon exchangers, mainly in column reboilers.

AndreChE