API plan for sour water pump

[sinectica]

Hello,
We are haing problems with single stage pumps that work with sour water (from crude oil desalators). Temperature: 120ºC
These pumps have single seal with flush plan 41, with cyclone separator and heat exchanger and the seals doesn't work properly as they leaking after few months. The flushing circulating doesn't seem to be good, as the temperatures doesn't follow a rational sequence.
I would be very grateful, if you had experience with this refinery service, if you could tell wich flush plan you use and your feelings about it.
I feel that plan 41 is not good and I would avouid the cyclone separator, as I do not have very good experience with this filtration element.

Any help will be greatly appreciated,
Marc

 

[JJPellin]

The service you refer to sounds like what we call mud wash on our crude oil desalters. We have 5 desalters in three crude units and they use two different flush arrangements. All of our crude units have vertical pumps type OHV. These are vertical pumps with bearing housings and flexible couplings. We flush the seals on these pumps with an external flush (Plan 32). In our older crude units, the flush is a cleaner stream of sour water from another part of the unit. Our newer crude unit uses boiler feed water flush (also Plan 32). But, since the BFW is too hot, we cool it with a water cooled exchanger just before it is injected into the seal. In this service, I would probably not use a cyclone separator. And any flush that comes from discharge pressure with a pressure drop back to the seal can have a problem. Because of the service, the pump product is saturated with gas and light ends. At any pressure drop (needle valve, orifice, etc.) some of the gas will flash out of solution. This can result in gas bubbles being injected into the seal with the flush flow. The affect can be minimized if you use a very tight bushing at the bottom of the stuffing box. A floating carbon bushing works well. Then, you can increase the size of the orifice or eliminate it entirely. Since the pressure drop takes place primarily through the throat bushing, the gas should be more likely to stay in solution until it has passed the seal. But, I would still prefer an outside flush. Depending on what you do with this sour water (sour water stripper, or straight to your waste water treatment plant) you may not be able to use any water with hardness as flush. The use of boiler feed water, cold condensate or demineralized water can be expensive. A stream of stripped sour water would be less costly, but may not be available at a high enough pressure. If I misunderstood and this is not a mud wash service, please describe what it is and I may be able to comment further.

Johnny Pellin

 

[longeron]

I think JJ's got the right idea. We're using a 32 as well and getting excellent service.

I've not had a whole lot of experience with cyclone separators. They will wear out over time and they may need to be "tuned" so that the DP is enough to drive adequate flush to the seal. I do believe that they have a place in the world- especially in some older pumps that will not allow for larger cross section seals designed for slurry service.

JJ, 3 Crude Units? How many BPD?

I'm contracted to a refinery rated for 190 barrels/day.

 

[JJPellin]

We are currently rated for about 275M BPD. With projects they already have on the books, we will be up to 320M BPD within about 3 years. We are located in Minnesota and run mostly sour Canadian crude.

Johnny Pellin

 

[sinectica]

Hello both, I am sorry because of the delay in my answer. I am grateful of your kind help.
As you have said, the best solution would be to use a clean external source of flushing but at this moment we can't assume this modification.
What I've seen in this sela as good as other ones that work with an API plan 41 is that the flush doesn't flow in the right direction, sometimes going straight from the discharge to the suction and from the seal to the suction as well(passing through the exchanger and the cyclone). I feel that sometimes don't help and it's more a problem than a useful item.
I am not sure if it's possible to dismantle the cyclone in this service, because I feel that it's better to have fresh flush with enough flow rather than to have a very small flow not fresh(through the bushing).

Thank you so much for your considerations,
Marc

 

[longeron]

Perhaps there are too many restrictions between the top of the cyclone and your seal chamber. You might try to increase the diameter of the tubing between the cyclone and your seal, and/or increase the size of the drill-through of your gland plate flush port- talk to your vendor about this improvement for your next repair. You might also want to check your seal chamber pressure.

There are other options available depending upon the solids content of the process fluid and the cross section of the seal chamber in your pump.

A rotating metal bellows can be used in some concentrations of solids. The bellows convolutions will "cut" through the solids and sling them away from the seal. There are also pusher seals available for higher solids concentrations. (I'd not suggest a single spring elastomer bellows in sour service- most of the elastomer bellows are offered in buna or fluoroelastomer- not my first choice for H2S) Several seal manufacturers have designs that will work in more than 20% solids- I'd suggest a tandem seal for process isolation and to help cool the inboard seal. Again, I suggest talking to your favorite vendor.

JJ- How is Jose doing? I'm his counterpart contracted to a refinery in St. Charles Parish, La. running mostly Venezualean, Mexican, and other inexpensive sour.