Seal and Bearing Failure approx. every 6 months 5

[mikeyd2]

Hello All,

I am working in a Pulp Mill and have been given the task of troubleshooting a pump in which mechanical seal failure is causing water contamination and failure of the front bearing. We are having to replace the bearing approximately every 6 months. The front feet of the pump is mounted to channel which is welded to the base. The welds have continously broken over time and have been rewelded. The channel has also cracked and been re-welded.

The pump is running on a fairly high efficiency part of its curve based on flow data taken from the DCS, so run out or low flow condition is not the problem. The fluid is pulp stock and I am not certain of the temperature (probably around 50C). The ambient temperature is ~8C (winter) and ~23C (Summer).

Some think that the problem is from entrained air due to the agitator that is installed on the same tank 5 feet away from the pump suction connection. This is a concrete, tile lined tank. I have been informed by the pump manufacturer that placing the agitator close to the pump suction is a common practice and diverts the flow around the circumference of the tank and into the suction.

There is excessive vibration when this pump is operating. I do not think cavitation is the culprit as there is no pitting or corrosion on the impeller or casing when it is re-built.

The pump suction is horizontal and is a straight run of 24" diameter SS pipe (~4 ft long) from the tank followed by a 24x14 exccentric reducer into the pump suction. Could thermal expansion in the suction be causing this problem? It seems like a short run but the tank it is connected to will not flex like a thinner walled metal tank.

Another reason why I think piping loads may be to blame is the arrangement of the discharge piping. The pump discharges vertically up about 9 ft before turning horizontally. Directly after it turns horizontal there is a guided support which only allows axial growth. This support looks like it is restraining the piping from growing up from the pump and putting the thermal expansion loads back into the pump.

Does it sound reasonable that piping loads may be causing this problem? Does anyone have any other ideas at what may be causing this?

Thanks.

 

[MikeHalloran]

Check that the pump actually has the special impeller designed to handle paper stock.

Free up that discharge pipe a bit. Letting it apply pipe loads to the pump body is just asking for trouble.




Mike Halloran
Pembroke Pines, FL, USA

 

[mikeyd2]

Thanks for the reply.

I should have mentioned that the consistency is fairly low. It is around 3.5% and that this pump is an old Alice Chalmers PWO pump which was used a lot for low consistency stock applications.

Cheers!

 

[Artisi]

Sounds like an installation problem rather than a pump problem, especially as the baseplate is fracturing.

3.5% stock is no problem at all for a PWO assuming it has been sized correctly for the duty.

What size PWO, is it a USA or a Canadian manufactured pump?

A photo of the installation might help.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[JJPellin]

The first thing I would do is perform a piping alignment / pipe strain check. Uncouple the pump from the driver. Place dial indicators across the coupling gap in the vertical and the horizontal position. Lock any spring supports in position. Unbolt the suction flange and then unbolt the discharge flange. Note the movement of each pipe and note the reading on each indicator. Our policy places the following requirements on these parameters:

• Neither dial indicator should register movement greater than 0.002”.
• The piping alignment across each flange should be such that you can insert all of the bolts by hand with no more force applied to the pipe than you can apply with your bare hands.
• The gap between the flange faces should be equal to the gasket thickness plus or minus 1/16”.

If it fails this test, then the cold pipe strain / pipe alignment is unacceptable. The next test is harder to quantify. Each line needs to be supported in such a way that any additional forces or movement will not be imposed on the pump. Each line should be fixed near the pump so that thermal growth is directed away from the pump. Appropriate spring cans should be used to support the weight of the fluid in addition to the weight of the pipe.

A thorough analysis using available software should provide the final piece of the puzzle. The software should confirm that the loads imposed under running conditions will not exceed the design capabilities of the pump nozzles. However, in my experience, the result of this computer analysis is often not worth the paper it is printed on. An experienced piping designer can spot a design problem that the program cannot.

Inspect the coupling using a strobe light with the pump running. Can any signs of severe misalignment be seen?

Coming straight off of a tank seems like a bad idea. This line would normally have an expansion joint to accommodate movement of the tank. How big is the tank? How is it fixed? What temperature variations can the tank experience?

A support on the discharge line that forced thermal growth toward the pump seems like a bad idea.

If it is not pipe strain, then I would suspect acoustic resonance. I have read about this problem in pulp and paper services. The paper that I recall most clearly mentioned failures of piping supports near elbows. But failures of the pump could also be explained by this circumstance. I would search for technical papers presented at the Pump User’s Symposium between about 1995 and 2005. There was a particular paper about an acoustic resonance problem in a pulp and paper service.


Johnny Pellin

 

[electricpete]

Out of curiosity, what kind of sign of misalignment are you expecting to detect visually during strobe inspection with machine running?

=====================================
(2B)+(2B)' ?

 

[spciesla]

Do you have any additional details on the high vibration, like frequency content and amplitudes? This might help us point you in the right direction as far as a resonance, misalignment, or pump issue.

 

[TenPenny]

That service should be a piece of cake for that old PWO, those things are rugged.

I recommend following the advice of JJPellin - what the mills here do with such pumps is (obivously when shut down) to loosen all of the bolts on the suction and discharge piping - if you can't wiggle the bolts by hand, or if there is visible movement of the pipe when bolts are loosened, the piping needs to be redone. After that is fixed or eliminated as a cause, then the other factors can be investigated.

 

[mikeyd2]

Thanks for all of the messages everybody. I'm going to recommend a few of these items that you have shared be done on the next shut (or bearing failure). I will also get a vibration analysis done soon and keep you up to date.

Thanks for the help!

 

[stanier]

I had experience of AC pumps where the baseplate was so flimsy the alignment could not be achieved. When pipe loads were applied to the pump the alignment would change again.

"Sharing knowledge is the way to immortality"
His Holiness the Dalai Lama.

http://waterhammer.hopout.com.au/

 

[JJPellin]

Our standard coupling is a disk-pack. A visual inspection using a strobe light can show coupling spacing problems (axial misalignment), uneven distortion (horizontal or vertical misalignment) or overload. If the condition is extreme enough to crack welds and fail bearings, there is a good chance that it would be visible on a disk-pack coupling. For other styles of couplings, it might not be possible to see these affects.

Johnny Pellin

 

[PumpSmart]

mikeyd2,
Please do:
-Take laser aligment readings.
-if base is light duty you will made stronger.
-Take analysis vibration FFT readings with agitator on.
-Take analysis vibration FFT readings with agitator off.
-Made JJPellin recommendations.
-Made a piping flexibility analysis if you does not have a software made a hot pipe strain / pipe aligment.
-Made hot pipe strain / pipe alignment, so when piping is cold, put 02 dial indicators on each pump flange with "0" readings(suction and discharge). Then star pump a see dial indicator value when pump is hot. If reading are more than 0.002", pump is dinamically loaded and it must be corrected.

Please send your findings and send photos or pump piping installation.
PWO was made by Allis Chalmer in Canada.
What material of the baseplate? If it is cast iron, welding is not good idea.
PWO usually have open impeller on Duplex material, if so, you will no able to see cavitation.

I hope this help.

 

[Artisi]

PumpSmart

For Info', the PWO started life in the USA evolving from the PW into the PWO a first cousin of the CW . The Canadian PWO is a redesign and a cost reduction of the US pump using identical hydraulics it was also built in Australian and UK.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[mikeyd2]

Hello All,

Thanks again for all of the great tips. I've attached a picture of the installation for those that asked for it. I've embedded photos of the tank connection and cracked channel in the upper right corner. I've also compiled a checklist of items to check for our next scheduled outage. Vibration analysis is ongoing as I type this.

Cheers

 

[Artisi]

Looks like 12x10x17 PWO(although hard to be sure of size looking at pic) Canadian designed unit, oil lub. bearings and packed gland.

1. is the inlet pipe well supported and not being carried by the pump inlet flange. Suggest the inlet is supported close to the inlet flange. Likewise for the discharge pipe work.
2.the tapered reducer could be a problem if the consistency gets above the 3.5% you have advised - 3.5 % is really the upper limit for this type of configuration - higher consistency can result in "dewatering" of the stock as it gets "squeezed" in the reducer causing the flow to chock further increasing the consistency at the inlet – this can cause flow surges and if the dewatering is sufficient flow can cease altogether.
3. entrained air in the stock could be a problem but this usually puts the pump “on /off- prime” causing surges and not steady vibration.

Not sure what the vibration an analysis will come up with but would suggest that it is an installation problem either piping load or misalignment of the drive or else the approach pipework / inlet.
There are hundreds if not thousands of these pumps installed in paper plants throughout Canada, UK, Australia and New Zealand operating under some terrible conditions with actual pump problems being extremely rare, installation problems are usually narrowed down to pipework alignment /loads, pump driver/driver alignment or inlet conditions.

If the vibration analysis and pipework come up ok then you need to look at the inlet condition together with the position of the agitator. I can advise more on the inlet configuration later if this looks to be the problem.


It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[Artisi]

One further thought, the impeller could be pre-rotating the flow which will give noise and possibly vibration.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[electricpete]

JJPellin - thanks. I agree shim pack coupling can crack under misalignment (look for ejected shim pieces underneath coupling guard) or show axial misalignment. If strobe is not available, camera with a flash provides quick easy way to inspect rotating coupling.

mikey/Artisi - I'm not used to seeing a reducer on the pipe suction. Is there supposed to be a minimum distance between reducer and pump?

=====================================
(2B)+(2B)' ?

 

[electricpete]

camera with a flash provides quick easy way to inspect rotating coupling.

actually, you have to go to some contortions to photograph up from below the coupling guard, and safety of course always a concern.



=====================================
(2B)+(2B)' ?

 

[mikeyd2]

Thanks for the response Artisi,

1.) The suction piping is not supported at all except for at the tank and from the pump suction flange. Similar installations in the mill have either a weight support or full anchor near the pump flange. This will definately need to be looked at.

2.) The consistency is actually upwards of 3.85%. Over the last 10 days it seems to hover between 3.8% and 3.9%. At what point do you want to go to sudden (flat plate) reducers? Could this cause the vibration we are seeing?

We ran the vibration analysis yesterday and are getting values of 0.921 (1st driven bearing) and 0.666 in/s (2nd bearing) with the agitator on and 0.635 and 0.636 in/s with the agitator off. The support channel is cracked or the weld is broken off at 3 out of 4 corners where it attaches to the base.

I am going to order a repair of the base as soon as possible and re-evaluate vibration at that time. If it is still vibrating at that time we will look into nozzle loading and some of the other suggestions that have been put forth from all of the helpful people on this forum.

Thanks Again.

 

[mikeyd2]

One more thing I should mention is that the dilution water at the reducer is cavitating through its upstream control valve.