Mechanical seal leaking on BJ Deepwell pumps 1

[nauticalmike]

I have been having problems with the mechanical seals on B.J. deepwell pumps leaking, often within only a few hours following their installation. The fluid being pumped is black oil, (ie. VGO, cutter stock, bunker fuels, waxes etc.) The seals we are using currently using are BW seals having silicon carbide and tungsten carbide faces and viton u-cup seals and o-rings. the seal faces look just as good when they come out as they do going in to the naked eye at least. Maximum discharge pressure of pump is around 130 PSI. The temperature of the oil is anywhere from 70 F up to around 200 F. I guess that my question is whether or not anyone else out there has experienced similiar problems with these pumps, and what they have tried/done to correct it. Also does anyone know the characteristics of viton over this temperature range?

 

[acp]

nauticalmike,

Viton's temperature limits are -15 deg.F to 400 deg.F.

I do not see any problem with Viton.Could you please let me know the type BW seals you are using. It seems there is a problem of application engg.

Regards.

acp

 

[nauticalmike]

BW 2625. We have made various modifications such as reducing the contact area of the stationary seal face and installing stiffer springs trying to eliminate problem with little success.

 

[TopFuel]

Hi Mike,

I've had similar experiences with submersible pumps in somewhat similar applications. In my instance (which may be similar to yours), the pumpage (diesel, oil...anything lubricating) actually coats the faces of the seal and let some pumpage slide by. The heat caused by the seal faces caused the pumpage to "slip" by and leak.

As double seal could work..as it did in our application. We used a barrier fluid, but you need to be sure it's compatible with the pumpage too. It took 2 tries, but we got it working. Depending on the pump type (horizontal, vertical, submersible) you might also try a "raised face" type seal..John Crane has some...called a HydroSeal. The faces are raised in spots to make the pumpage wick away. It almost acts like a luby seal in principal.

The stiffer spring may work, depending on the characteristics of the fluid being pumped. It might solve the problem, but if the faces get hot again, you'll have some leakage...just not as much.

Hope that helps some!
Jack

 

[longeron]

Just a few questions:

Is this a new application? If not, how long have these failures been going on?
Has anything changed in the process around these pumps?
What flush plan is used on these pumps? I’m assuming a plan 13.
What is the suction pressure, and how many stages is this pump?

Jack- hydropads maybe?

 

[nauticalmike]

Nothing new here except changes to sealing components. I do not know what you mean by "flush plan." suction presssure varies as they are cargo pumps for a petroleum barge. The pumps have three stages.

 

[JJPellin]

We have almost 50 Byron Jackson deep-well pumps. They run in services from propane to water to crude oil. The crude oil service is probably closest to your application. We run Borg Warner U-cup seals in several of these. We run
Silicon-Carbide versus Silicon Carbide with Viton elastomers. As noted above, Viton should work well in the temperature range you are dealing with. There are a few things to be concerned about with this type of seal in this type of pump. The seals is under full discharge pressure. The seal is at the high point in the liquid stream. The shaft is only supported by bushings rather than bearings. This leads to the three possible causes for premature failures that would not include obvious face damage. Since the seals are under discharge pressure, you can't flush the seals with a stream of product off the pump discharge. No pressure differential to drive it. Any gas that might exist in the system would tend to rise to the top, right up to the seal which can cause the seal to run dry. A good solution to both of those problems is to use a Plan 13 flush. This refers to the API piping plan. A plan 13 uses a line from the seal back to the suction (or supply vessel). This provides a flow of product past the seal to keep it cool and would tend to vent off any vapor that accumulates around the seal. This seal is a dynamic pusher seal. The rotating face seals to the shaft sleeve with viton U-cup that must move to keep the faces in contact. If there is any run-out on the shaft or the face of the seal chamber where the seal gland bolts up, this U-cup has to move a lot with each rotation of the shaft. This will very quickly cause fretting on the sleeve under the U-cup. If the run-out is excessive, the U-cup can't move fast enough and the faces open up. This can occur in this pump if the bushings in the pump are worn out or if the ridgid coupling above the seal is damaged. I would assemble the coupling without the seal. Use a dial indicator on the shaft to see if it has run-out (bent shaft). Then mount the indicator on the shaft and sweep the seal chamber face to see if it is square to the shaft. We would make repairs if either reading showed run-out greater than 0.003". If I had to guess, I would think you have worn out bushings or a damaged coupling. Increasing the face loading with stiffer springs and a narrower face wouldn't help this. If you see fretting damage on the sleeve under the U-cup, this would be a strong suggestion of a run-out problem.

 

[acp]

jjpellin,

yes,precise trouble shooting and the best solution.

Congrats,

acp

 

[nauticalmike]

jjpelin,
I thank you very much for your insight. I have a couple of additional questions for you, if you don't mind. First the coupling is a threaded coupling that threads onto both shafts (6"long with internal threads 2 1/8" - 6tpi LH). How could this coupling be bad? I am assuming of course that both shafts thread into it by hand? Also I am assuming that the two shafts must be butting up aginst each other inside the coupling because to disassemble them one needs to use a come along on the end of a 36" pipe wrench. What is the allowable clearance between the shaft and the bushings in the pump? What is the purpose of the micarta bushing in the seal cover flange if it isn't there to prevent the sleeve from hitting and breaking the stationary seal face in the event the shaft runnout becomes excessive. Why would indicator readings taken with the seal removed be any different than readings taken with the seal installed, assuming of course that the shaft deflection is not so great that the sleeve hits the micarta ring located in the seal cover flange. If the problem with excessive shaft deflection is that the u cup can't move fast enough and that this inability to move fast enough to hold the faces together, wouldn't the seals begin to leak immediately upon bringing the pump up to some critical speed that exceeded the maximum travel speed of the u cup rather than beginning to leak after a variable length of time? The more I read your post the more questions I have...

 

[JJPellin]

Based on your comments, this pump is a different configuration that I thought. The motor must be a hollow shaft with an adjustment on the top of the motor to set the impeller lift off the bottom. The only problem you could have with this style of coupler is if the ends of the two shafts are not square. The shaft do butt up against each other. If the ends are not square, the shaft will "bend" when the coupler is made up tight. The reason to check run-out without the seal it so that you can check the seal chamber as well as the shaft. If the shaft is not passing squarely thru the center of the seal chamber, the seal will have to flex with each rotating even if the shaft is perfectly straight. Leakage from high run-out could occur immediately on start-up if it is bad enough. Otherwise it might take a while. As the seal runs, the u-cup would start to frett on the sleeve and the leakage could be comming under the u-cup and not across the faces at all.

 

[JJPellin]

The gland bushing helps to protect the seal faces from damage during installation. But the main purpose is to restrict the leakage in the event of a catastrophic seal failure. We call it a disaster bushing. The bushing clearance in the pump depends on shaft size and bushing material. Consult the pump manufacturers manual. For a B-J VLT 1500 with a 2-1/8" shaft in oil service, we would use the factory original carbon-graphite composite (Graphalloy) bushings with 0.005" to 0.007" diametral clearance or bronze with 0.006" to 0.009" clearance. If you install the motor and couple up the shaft to check run-out, you can also check for excessive clearance. Put an indicator on the shaft and pull and push it to get a feel for how much clearance you have in your bushings. Not a very precise check. But if you get 0.025", you know the bushings are bad. If you get 0.010" you are probably fine. Anything in between is a judgement call.

 

[nauticalmike]

Actually my aplication is driven by a detroit 12V-71 through an angle drive.

I usually check the shaft deflection by pushing and pulling as you suggested, except I do it uncoupled, moving only the pump shaft. However the readings I usually get are anywhere from .060" to .070". The lowest reading that I have ever gotten was .055". These values are about the same for at least 7 of the pumps that I have checked using that method. These high reading are consistently this large even immediately after the pumps were sent out to a machine shop and supposedly overhauled. I have always thought that this much play in the pump bushings was excessive, but as I said before, I've never had the opportunity to read any of the literature on these units, so I didn't question it. Also these values have been reported to the customer numerous times, and they never expressed any concern about it. Since they do have all of the manuals etc, I've always assumed that they would know what the readings should be. Especially since they were the ones that requested that I check the shaft deflection in this manner every time I replaced one of their seals to begin with. It looks like I should find myself a copy of the manual you suggested. I thank you once again for your insight, you have been very helpful.

 

[GLC1]

Mike, When you fit the couplings are you locking the seal to the shaft before or after you set the impeller lift. If you do it before it could simply be that the seals are not in the correct possition. Agree about the runouts on vertical pumps they always cause problems. Most arrangements of seal fitted on vertical pumps experiance venting problems as most of the time the seal is drawn horizontally so its not concidered.

 

[nauticalmike]

I always set the seal clearance to 0.1875" after the impelller lift has been set. If I'm not mistaken the impeller clearance is grater than the seal clearance, so to lock the seal to the shaft prior to setting the impeller clearance would result in a seal clearance that was excessive.

 

[nauticalmike]

It was suggested that vent lines be installed between the seal flange and the pump well, but I'm wondering if anyone has any experience with this when the product being pumped solidifies at ambient temperatures. I'm concerned that any vent line I might install would probably work for the first discharge, and after that the product would cool and plug up the line. Once plugged it would be unable to vent the seal cavity, at least until it was warmed enough to be blown out. It seems to me that venting the seal cavity would be mostly beneficial when the pump is first started, and then again at the end when the tanks are stripped. They sstrip until the pump loses suction and then the pumps are secured. If I am correct in my assumptions, at start up the vent would not work. If anyone has any experience with a similiar application I would appreciate any insight they have. Not that installing vent lines would be very difficult, I just don't want to suggest a modification if it won't help.

 

[longeron]

Venting the seal chamber is really only half the idea. The API Plan 13 is used to ensure that any vapor that does get into the stuffing box during operation is vented back to suction, and to induce flow around the mechanical seal to help lubricate the cool the seal faces. This second function is just as important as venting the stuffing box. If you are worried about the product solidifying in the tubing I suggest using a high-point-bleed in the flush tubing. With this particular pump, the folks I’ve talked to say that it’s best to install the tubing so that it extends a foot above the flush connection on the gland, then back down to the suction or to the bottom of the barge. At the highest point of the tubing install a tee and a valve off the tee so that you can vent this line to atmosphere to ensure that the line is liquid full at start-up. When the pump is turned off, open the line to the atmosphere to drain both legs of the flush tubing.

I thought you were using a Flowserve 2562 DIN cartridge seal. There shouldn’t be any question about the seal set point with a cartridge seal. The seal is installed on the shaft and bolted to the backplate without tightening the set screws, then the impeller is installed and clearances are set, and lastly the seal set screws are tightened and the setting clips are removed. If you are not using a cartridge I suggest talking to your vendor about finding one for this application.

Your question about the gland bushings is a good one. The bushing installed on the outboard end of the seal is there to protect the shaft from the gland plate if deflection is excessive. It can also be used to help contain a quench such as water, steam, grease, or nitrogen to keep the fluid that crosses the seal faces to the atmosphere from crystallizing and keeping the springs from working or wearing the shaft. You also brought up another good point about u-cups. This is why I’m not a proponent of their usage. U cups can cause a hang-up condition easier than an O-ring. O-rings are more flexible. If you are truly suspicious of this problem you may want to look into an elastomer bellows seal instead of the u-cup. These seals usually use one large spring instead of several smaller ones which solids can interfere and cause hang-up. Bearings can also be installed in the seals to help control shaft deflection.

 

[nauticalmike]

Longeron, These seals do utilize only one spring.

jjpelin, would it be possible for you to send me a copy of the manual for one of these pumps, because I am having a hard time finding one?