KSB feed pump balancing disc damage 2

[rlee1234]

I am new to the forum and looking for some tips.

Application is on a 3x1 combined cycle project. Deltak HRSG's with GE 7 FA combustion turbines, D11 steam turbine.
HRSG's have duct burner.
GE 7 FA's have steam power augmentation used on a regular basis.
Have had 7 disc failures on KSB HGC-5 8 stage pumps.
Each HRSG has two 100% capacity pumps with 5KV drive motors.
Evidence on the pumps indicates "loss of water on the balancing disc". No evidence of cavitation on the impellers. The drive end bearing will indicate some wiping. Little or no bearing wear on the non-drive end bearing. Pump alignment and pipe stress has been checked and no mis-alignment has been found.
First indication any problem is a trip from the seismic vibration switch, (no continous analog X-Y vibration or axial monitoring available).
Recirculation flow in low flow conditions has been 10 to 20% above minimum required by manufacture, confirmed through DCS historian. (An analog control valve is used for recirculation control, not an arc valve.)
The NPSH has 20 feet of margin, on paper, for the worse case condition.

We plan to install temporary full time axial and X-Y vibration monitoring.

Has anyone else experienced this failure and what was the resultant findings?

 

[JJPellin]

Is this the version of the pump that has no thrust bearing? If it has a thrust bearing, what is the configuration? Has the thrust bearing been found to have failed at the time of the balance disk failure? Is this a true balance disk or a combination drum/disk? Without these details, I would suspect some of the following:
* A restriction in the balance line. Perhaps this one has a restriction orifice or venturi fitting to measure balance line flow. Could there be something lodged in this fitting?
* An assembly error. Improper set up could result in a mounting that allowed the balance disk to rub axially while the pump is running within its normal range of shaft positions. This could be an error in the mounting of the balance disk (split ring retainer, shimming) or the thrust bearing (if it has one).
* A design error. The specifications for the size and clearance of the balance disk could include an error that results in repeated failure. This would be hard to prove.

We don't have any pumps of this exact type or configuration, but I just finished overhauling a 9 stage Worthingon, stacked diffuser, high pressure barrel pump that probably has some similarities. I have never seen an axial rub in this pump at the balance disk face. A similar smaller pump in another unit used to rub the axial face of the balance disk/drum regularly. We found design problems and controls problems that were resulting in extreme variations in flow (much too high and much too low).

 

[rmw]

What type of sudden load fluctuations do you encounter?

rmw

 

[rlee1234]

JJPellin,

Thank you for your response.
These pumps utilize a "lift-off" assembly for initial start and do not have a true thrust bearing. The "lift-off" assembly, (two back to back roller bearings) has no indication of damage or wear.
These pumps have an oil bath "slinger ring" lubrication, not force lubed.
The balance system in a drum/disk assembly.
The balance line and fixed reducing multi-ring orifice have been checked and no debris or plugging was found. The reducing multi-ring orifice has been checked for correct part number but I do not know how to insure the reducing orifice is manufactured correctly.
After the pumps were rebuilt by the manufacture, the axial and radial clearances were checked and meet the manufactures clearance requirements.
There still could be a manufacturing error, but based on the thrust clearances it meets spec.
We are pulling more data from the DCS control system to further investigate "high flow" and "low flow" conditions.

rmw,
Thank you for your question.
At present we are looking into the extreme peaks and valleys the pump may be experiencing. The preliminary data does not indicate anything beyond pump guarantees. This data was a long term request, which averages the data. We are now pulling 1 second data from the areas in most obvious operation, which would be on start, shut down of the unit, and during power augmentation operation.

 

[JJPellin]

The Worthington pump that I referenced in my first reply (9 stage) had a number of failures shortly after it was installed. I was not here at the time (1984) but have read some of the notes from that time. The solution ended up being a redesign of the first stage impeller. It had been designed for relatively low suction head since operations wanted to be able to pump the tank almost to the bottom. As a result it had a high suction specific speed. The first stage impeller was failing from suction recirculation cavitation. At the time the pump did not have continuous monitoring and they were having a hard time determining what was going on. Once the pump got a full overhaul they found severe cavitation damage in the eye. The solution ended being a change to a higher NPSH required impeller (32 feet versus 28 feet) which lowered the Nss enough to stop the suction recirculation cavitation. The point of the story may be that the original design showed that the pump had a substantial NPSH margin and even at low flow through the spill-back it was running well above the manufacturer's recommended minimum flow. When you install the temporary shaft probes, check for radial vibration indicative of cavitation or at vane pass frequency and check the axial probes for axial vibration in addition to thrust position. I don't know how suction recirculation cavitation and axial vibration might affect the balance disk/drum configuration in your pump. But I would be skeptical of the manufacturer’s number for minimum flow. Excessive NPSH margin cannot suppress suction recirculation cavitation.

 

[checman]

rlee1234
Where does your balance line discharge to? We sometimes use the flow from the balance line as part of our minimum flow so we will discharge it to the DA. We will also have a bypass for additional flow to meet our minimum required flow. I have seen where customers have tied the two together in a small diameter pipe. The pump has plenty of pressure to push its minimum flow through the line. While the pump is bypassing the differential across the balance line changes and the flow through the balance line drops off. And that is the problem. I guess the question would be where does your balance line discharge and where does your bypass discharge?
checman

 

[JJPellin]

Our balance line ties directly into the suction line of the pump with no other connections. The minimum flow system is a multiple orifice assembly that drops the 3500 psi discharge pressure down to about 15 psi and this flow dumps back into the bottom of the water supply tank.

 

[JJPellin]

Routing the flow from the balance line anywhere but the suction seems risky to me. I have no experience diverting this flow but would be very concerned. We once had a situation where a unit was coming down for turnaround and the operators needed to inject a chemical into the process. There was a bleed valve on the balance line of the large multi-stage barrel charge pump and they connected their chemical injection system up to that. Shortly after this, the pump shaft fractured under the thrust disk. I was never able to prove conclusively that the extra pressure from the chemical injection was the cause of the shaft failure. But it would seem an unlikely coincidence it is wasn't. I would also not feel comfortable counting the balance line flow as part of my minimum flow. I have never heard of this. Even though the balance line flow does come from the pump discharge after passing through all stages, this was probably not intended by the pump manufacturer to be included in the minimum flow requirements.

 

[rlee1234]

JJPellin,

We have had these pumps completely disassembled on several occasions and I have not seen any evidence of suction recirculation cavitation, cavitation on the impellers, pump casing, or pump nozzles. I am not brave enough to say that some flashing is not occurring, it is just not evident except that the balance disk is wiped. We plan to monitor both radial vibration and axial movement.

JJPellin and Checman,

The "spill back" or return flow from the balance disc returns to the suction nozzle, ~30 gpm. Total pump flow is ~2,000 gpm. The return line and cavities have been check for debris or plugging, no problems have been noticed.

 

[JJPellin]

In addition to radial vibration and axial position, you can jumper your axial probes over to a vibration monitor and measure axial vibration. If the primary failure is the axial face of the disk part of the drum/disk, it could show up in the axial direction. A few more thoughts occurred to me as I read back over this chain. A problem with the coupling could be transmitting an axial load to the pump shaft. Incorrect setting of the distance between shaft ends. Failure to properly mark and measure for motor magnetic center. An under-designed coupling; a grid coupling will produce a large axial force when subjected to high torque. Failure to account for axial thermal growth by pre-stretching the coupling. A pump mounting problem could also lead to distortion that might contribute to balance disk problems. A large barrel of this type is probably keyed for case growth depending on the process temperatures. Shipping washers could be locking down a foot that is designed to slide. Alignment keys on the pedestals or the bottom of the case could be locked up or not correctly installed. We installed a large set of barrel pumps in 1993 that had this problem. The contractor didn't weld down the fixed pin on the inboard end or the key block on the outboard end and had left the shipping washers in place locking down the feet. One of these pumps had a rub of the first stage wear ring as a result of the case distortion this caused.

 

[sprintcar]

What did KSB tell you? The first step in any field problem should be Vendor Contact. They have a global field service team and should be able to assist you

http://www.ksb.com

"If A equals success, then the formula is: A = X + Y + Z, X is work. Y is play. Z is keep your mouth shut."
-- by Albert Einstein

 

[pdldavis]

Hi, I am involved in this in a peripheral way and RLee is headed back out to the job site again. KSB was first on the contact list and at this point they are pretty much scratching their heads too. There has been no solution so far.

The odd thing with this is the total trashing of the balance disk assembly, minor wiping on the drive end bearing with no observable impeller damage. The photos returned from the job site had good pictures of the pump impellers and they looked brand new. If there is any evidence of pump cavitation, we have not found it.

In the interim we have another site going into outage later on in the month that also utilizes KSB boiler feed pumps. We will be inspecting those as well.

This is a warranty issue so we are quite naturally interested in getting to the bottom of the balance disc problem and were hoping someone else out there might have experienced similar failures with KSB Boiler feed pumps.

Thanks, Dan

 

[checman]

Pdldavis

What is your minimum bypass flow? Where does it discharge? jjpellin I have trouble shot several high pressure boiler pumps. After misalignment and pipe strain, heat from the balance line and bypass flow have been common problems. On some I have changed where the balance line discharges to and on others I have increased the bypass flow to dissipate the heat from the balance line in the pumps suction. I have some good stories on how pipes and buildings shook until changes were made. At the pump company I work for we have several pumps balance lines running back to the DA. When properly done it’s a good way to dissipate heat from the pump.

Always busy and outside the box. checman

 

[JJPellin]

The main fear that I would have from diverting the balance line flow to anywhere except the pump suction would be the possibility of there being a block valve in that line. Any chance that someone might mistakenly close that valve or fail to open it after working on the system seems very risky. Our largest barrel pumps are typically in services where a small amount of heat from the balance line flow would be insignificant: Coker jet water, gas oil charge to a desulfurizer, etc. But I can easily see where some BFW applications might have problems with any extra heat.

 

[pdldavis]

-JJPellin and Checman,

The "spill back" or return flow from the balance disc returns to the suction nozzle, ~30 gpm. Total pump flow is ~2,000 gpm. The return line and cavities have been check for debris or plugging, no problems have been noticed.

We should know a little more tomorrow. As an aside, we are also going to install a pressure gage on the return line.

 

[checman]

pdldavis,
I am interested in the minimum flow through the analog recirculation control valve. How much and where does it go? Also is the valve slow open slow close? Thrust loads from valves that fly open and close can raise heck with balance disc or drums.

Regards checman

 

[rmw]

Or from BFW level control valves that are out of tune, or from sudden load swings on the units.

But maybe your term "valves" included all of the above.

Maybe I should have said "flow fluctuations" instead of "load fluctuations" above.

rmw

 

[pdldavis]

Hi, the plant is operated base load so load, flow fluctuations would be minimal.

The Recirc Valve returns to the LP drum. I don't know the flow but I will ask.

Valve control is typically slow acting and I am unsure if this is the case at this plant. This is one of the items they will be looking at.

I think KSB shows up out there today so I should hear something tomorrow.

One other thing that we are currently unsure of is when is this damage occuring. Is it more likely to be occuring as a result of a pump start or stop and compounds with the number of starts and stops. Or is it a result of something happening during steady-state operations. Or something else.

It is an interesting problem.

Thanks for all the input.

Dan

 

[JJPellin]

Have you found your problem? I would be very curious to hear the final result.

 

[pdldavis]

Hello, I have a report on the balancing disk damage on the KSB pumps. We finally got a resolution on this last week and here is basically what was going on as I understand it.

The plant was originally built to operate as a base load unit but in reality it was being operated more as a peaking unit.

The owner was aware that valves and check valves on the suction and discharge of the bfp's leaked by but did not disclose the fact that no preventative maintenance was done on the equipment. The lack of maintenance was because the owner chose to ignore that over profits from peaking. They did not equate that with bfp problems.

There was enough leakage to where there was insufficient pressure in the balancing disk area for the disk to do its job during startup so for a short time the disk was not centered and there was metal to metal contact - gradually trashing the balancing disk.

This was proven out to the point that it is not a warranty issue anymore. It is a repair issue at the owner's expense.

Thanks again for your suggestions and input.

pdldavis