Increasing water rates to an existing Reflux Drum

[OilBoiler]

Hi everyone,

We have very bad corrosion problems at one of our fractionators. It has been decided that increasing water wash to the overhead system may further dissolve the salts and solve the problem. The new water wash rate has been determined by simulation. Our task has been to calculate the new settling rates for the new water flows at the overhead receiver that is a 3-phase horizontal separator. In GPSA, the methods for calculating settling velocities are quite vague.

To determine the droplet diameter, I have assumed a residence time of 5 minutes (per GPSA tables) and a diameter of 100 microns and then used the following formula for settling rate (in/min) = 8.3e5*[(d^2)*(SGwater-SGoil)/Viscosity)]

The 8.3e5 is the proportionality constant to get velocity in in/min. If the distance travelled by the droplet is higher than the diameter of the drum, then I use a different droplet diameter. I obtained a diameter of 100-120 microns for the residence time I specified.

Question 1: What is the droplet diameter you think is appropriate for OVHD coker fractionator streams?
Question 2: What is the usual residence time for this service?
Question 3: What method do you all use to determine settling rates?
Question 4: What is the typical range of settling velocities for these types of services? I have read that is about 10 in/min

If you have better approaches for these types of calculations, I appreciate your input.

Thanks again,

OilBoiler.

 

[25362]

I'm sorry to disappoint you, but the Stokes' law is of little practical value because the droplet size distribution is rarely known. A general ROT for designing horizontal drums for water settling through a light hydrocarbon is as follows:

4 inch per minute (ipm) or less, good to complete settling.
6-10 ipm, acceptable settling.
18 ipm +, water settling incomplete.

Usually water droplets would settle at 5 to 6 ipm.

The above ROT applies as long as there are no emulsions by surfactants or turbulence (agitation). Rather than guessing a settling rate, performing lab experiments on settling rates is sometimes recommended. Try to maximize the distance from the inlet nozzle to the three outlet nozzles: gas, naphtha and water.

To avoid agitation the inlet nozzle is generally turned 90^o towards the drum head and flared-out as with a horn.

All the above notes are taken from Process Design For Reliable Operations by Norman P. Lieberman (Gulf), ISBN 0-87201-747-8.

To avoid pumping problems one should be sure that the levels monitored and measured are correct. See, please, fig.4 + in

www.distillationgroup.com/technical/p095.htm​

good luck.

 

[25362]

See also A Working Guide to Process Equipment by Lieberman and Lieberman (McGraw-Hill) ISBN 0-07-038075-9.

 

[epoisses]

Sorry for looking back, but before you put a bigger drum in... how sure are you this is really going to solve the problem? If it's the overhead bundles that are corroding, that usually occurs at a very specific location in the bundle where the first acids condense and where the pH is extremely low locally. Not sure if adding water helps. Does anyone have good experience with a similar case?

 

[jalvarez]

OilBoiler:
Pipestill overhead corrosion is a very well know issue.
I've never experienced the corrosion you mentioned in the Delayed Coker I was involved in the past (OilBoiler, is it a DC?).
I assume (I'm a bit reckless) this isn't coming from salt content in the VPS bottoms (if the salts are chlorides, they are volatile like in the PS ovhd, or must remain in the bottoms, so they must appear in the coke. Little amounts in the Coker gas vapor stream? Perhaps, I'm not sure. Like to see other's opinion.
High sulphur content in the crude oil being processed? Did you verify the origin of the corrosion?
Now, if you have the corrosion there, I agree that maybe an injection of water could be a good solution, but now you face the fact that you need to re-design the distillate drum, as you have a bigger amount of water to be removed.
I have some formulas for designing this, but this is not a comfortable text processor.
Let me know your email and I’ll send them to you. Let me know further:
Do you have a crinkled wash mesh screen in the drum outlet?
Do you have a compressor downstream?
Do you have a 90° elbow (as 25362 mentioned), or a holed, or a slotted inlet distributor?
Have a safe day
J.Alvarez