Vacuum Tower Overflash pumps

[Falcon03]

Hello everybody,

We are having some difficulties in controlling the overflash flow back to the vacuum tower. Our pumps can’t deliver the required flow rate to maintain the level in the overflash draw pan. So we took the LIC out of control and we are controlling the level manually.

Based on past experience these pumps did fail several time, two month ago we had a seal leak in one of the pump which caused a small fire in the area.

Currently, we are controlling the flow by pinching down the discharge flow to these pumps and not to have big temperature swings in the flash zone.

Does any one have an experience in such system like ours? What is the best control system to maintain the liquid level in the overflash zone ?

I would like to add one more point that the metal contents in the FCC feed (MVGO and HVGO) has increased which mainly due to improper O/F flow.

 

[TD2K]

I'm a little confused. If the pumps don't have enough head to get the flow you need, how does putting the LIC in manual help you?

When you do run this in manual, do you get enough liquid back to control the metals or are they still running high?

 

[Falcon03]

Thanks TD2K for your response. To answer your question. Taken the CV back on manual beside pinching down the discharge flow valve will steady out the process. In addition to that when we pull the strainer we will find it full of coke. So I'm note sure if we do have a pump problem (NPSH) or our control system is not appropriate for this service.

To give you a full picture about our system: We are pulling O/F liquid and the discharge stream is split into two streams. One goes to Crude tower bottom this stream controls the overflash level draw pan and the other stream will be sent back to Vacuum tower directly above the O/F draw.

 

[lgmarti]

Some refineries have discontinued the use of the pump in this service. Instead, they gravity feed the overflash to the vacuum tower bottoms. Of course, this can be done only if there are no light ends in it. In our refinery we pump it to the crude tower btms on flow control. We have been fortunate, the level is everywhere but the maintenance in the pumps is not overwhelming.

I can't believe you split this flow! If you can't deliver enough flow to maintain level why would you split this flow and return it to the vacuum tower?

I do not think that flow alone is a good way to prevent coking in the flash zone. You should monitor the temperature in the stream. It should be at least 35 degF lower than the flash zone temp. If the dT is too low, there is not enough overflash.

Do you sample the overflash? If asphaltenes are present entrainment is the cause of your high metals. If not I would look at flash zone T and Pressure. Exchangers leaks?

 

[qazx]

Falcon03,
Refer to your question,I am intrested to transfer some operating data with you becuase we have also some problems about the overflash and some equipments on it.If it is possible please contact me (asdfqazx@yahoo.com).

 

[krsmani77]

Dear Falcon3,
I will answer this question from my refinery experience. Points to note and absorb are:
1. In a vacuum column, the role of overflash ( sometimes called vacuum slop or demister wash ) is to control the quality of fcc feed stock/heavy neutral (lube distillate) which is the cut drawn above flash zone.
2. The quantity of overflash is a function of coil outlet temperature, Bottom stripping steam( partial pressure effect), Flash zone vacuum and the net draw of distillate cuts above flash zone in addition to the feed quality ( Quality of Reduced crude).
3. In view of the above point 2, there are great many variables working on the available quantity of overflash for automatic control. All other variables being fixed, you can change the overflash by changing COT of Vacuum heaters or dropping some or drawing more of FCC feed.
4. As the quantity of this product is small compared to all other streams, but temperature being one of the highest, the heat exchanger/cooler tubes normally fail and leak. So avoid heat exchange with this stream as much as possible. Normally it is to be re-evaporated and hence can be directly routed to crude bottoms going to the Vacuum heaters.
5. The demister wash, which is either vacuum slop or a part of FCC feed/Heavy neutral cut is given on top of wash deck plates/ demisters, is to improve the quality of FCC feed w.r.t carbon residue/color. If there is no wash, the quality will suffer.
6. There is a chance that the plates/decks/draw off pan may have been disturbed and you are not getting the over-flash at the draw-off pan.So even if you dont get the overflash at the pump, it does not matter except that you may not be getting the demister wash flow. If I were in your shoes, I will direct all the available overfllash liquid to the demister wash and forget about routing to the crude tower.
But in case your short residue quality is important, then you will have increase your stripping steam and COT to reevaporate the overflash.
7. Pumps in this service should not have self-flushing mech seals, as it is liable to fail. Have an externally flushed seal. A failed seal is liable to leak and make the pump lose suction aggravating the problem further.
Think over the whole thing, You will surely get the right solution.
Best of luck

 

[25362]

It might seem obvious but sometimes people locate valves and restrictions on the horizontal draw-off lines -before their vertical descent- causing draw off or trap out sumps to overflow, and pumps to lose suction. Restrictions should be located after the vertical pipe drop.