Kettle Reboiler Vapor Space Calcs 1

[thermmech]

Greetings!

We have a C3/Mixed Refrigerant Evaporator (kettle reboiler) level control issues, where "kettle" reboiler is:
- BKM type, 1600 mm (4200 mm) x 8100 mm
- not conventional kettle, bundle is stabbed into the horizontal vessel, presumably due to economy
- demister pad at ~850 mm above tube bundle
- LLL = 50 mm above - ||-
- NLL = 100 mm -||-
- HLL = 150 mm -||-

We have found out that level control within specified values level values is not possible. The bundle top is run dry from time to time.

1. Considering that normally shell metal temperature is -36°C, and mean tube wall metal temperature is -32°C, is there a danger to the integrity of the exchanger, being fixed tubesheet type - if we keep it running while the bundle top is cyclically run dry? I understand that shellside vapor will have way lower heat transfer coefficient and that will lead to top tubes' metal temperature to be a few degrees warmer than the rest of the submerged bundle. My cursory calcs show that this translates in ~1000 psi of additional cyclic longitudinal stress.

2. My HTRI rating run shows that the kettle nominal diameter should be ~ 4200mm as it actually is. However, having demister pad half-way in the vapor space increases horizontal velocity of the vapor. Does anyone have criteria for selection of horizontal vapor velocity in the vapor space so I could double-check the adequacy of the existing space? It is critical to minimize the entrainment as vapor goes to the compressor.

Thank you.

 

[srfish]

Most kettle sizing equations are based on calculating the limiting vertical vapor velocity. In order to evaluate the maximum horizontal vapor velocity you can use a vapor/liquid seperator equation like a Souders-Brown type. One source for this is on page 7-7 of the GPSA Engineering Data Book.

 

[thermmech]

Thanks!

 

[thermmech]

For info, I have also found an article by Ben Tammami "Simplifying Reboiler Entrainment Calculations".

Cheers,

 

[buffbull11]

If your chiller is single pass (and I presume it is since it is a fixed tubesheet), you must also consider that the heat flux and vapor generation rate at the hot tubesheet may be much higher than at the cold end. This will further increase your actual horizontal vapor velocity, as it would not be uncommen in those situations to have most of the total amount of vapor generated in the first quarter of tube bundle length. Most kettle sizing methods will presume that the vaopr is uniformly generated over the length of the tube bundle; this is a reasonable assumption for multiple tube passes, but not for single pass units.

If your temperature difference between the shell and the tubes is only 4 degrees C, thermal stress is probably of no worries. A simple mechanical program can check the axial stress in the tubes, shell, and tube-to-tubesheet joints. Remember as well that this is a secondary stress.

 

[srfish]

buffbull11 makes a good point about the single tube pass and unequal vapor generation. Years ago when I worked as a thermal rater for a heat exchanger company, in this situation we used a 60/40 split instead of a 50/50 split of the vapor generated.

 

[TXiceman]

HTRI does a bad job of calculating vapor separation velocities and disengagement space. Stokes Eq does a good job.

Ken
KE5DFR