Pump thrust load reverse

[Rreach]

We have a boiler feed water pump that consistently fails at the thrust bearing (2-stage, between-bearing, radial split pump, see attached picture). Out of the four pumps operating in parallel, the failure always occurs at the B pump location. Upon examining the failed bearing (back-to-back and outboard inner race overloaded) the thrust load is reversed from the expected direction which is toward NDE.

Years ago, the OEM increased the 2nd stage suction wear ring to reduce the thrust load and factory test showed that the thrust load is towards the NDE. Yet the B pump continues to fail at this particular location. We have confirmed that there is no piping stress, alignment is good and the NPSHa is adequate.

Does anyone have any idea why the axial load could be reverse for this pump? What else can we check?

Thanks in advance.

 

[Doug Hunter]

My first impression is to look for something else installed backwards or oversize. Is the pump impeller directional in nature, and can it be installed backwards accidentally? I would guess not from the picture.

One problem solving approach that could be applied in this situation is swap studies. The first cut would be to swap pump A, C or D with pump B to see if the problem is related to the pump or the pump install location. Then continue to work downward to the root cause from what is learned there. How long does it take the bearing failure to occur, or how long does it take to start seeing signs?


Best regards,
Doug Hunter
Altarium Technical Consulting

https://www.altarium.us/

 

[Rreach]

We recently installed a brand-new B pump from the factory. After running it for 4 hours, the oil in the NDE thrust bearing turned dark, and we discovered metal chips in the oil. We haven’t removed the bearing for inspection yet. I plan to conduct another test run and want to gather all necessary information before disassembling any parts. I’m considering whether internal cavitation, possibly at the 1st stage impeller eye, could be affecting the axial thrust balance. However, I’m unsure how to obtain data to confirm this, as the suction pressure gauge shows we have sufficient NPSH.

Does anyone have any suggestions?

 

[georgeverghese]

Narrative on page 10-23 of Perry Chem Engg Handbook re corrections to be made for extrapolating cold water NPSHr for hot water service application may be of interest.

 

[Rreach]

Correct me if I'm wrong, but I understand that NPSH required for hot water is actually lower when tested with cold water. I do not have the Perry Chem Engg Handbook, thus could not check that.

 

[LittleInch]

What are your details?

Remember NPSH is not the cavitation limit.

Cavitation starts at anywhere from 1 to 4 or 5m higher than NPSH.

What is the location and set up of the inlet system? Drawing or sketch might give us some ideas.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[TenPenny]

If it's always the pump in B position, and you replaced it with a new unit and the same thing occurred, it's an indication that it may not be the pump, it's a piping or operating issue. Verify operation of all valves, check valves, etc.

 

[georgeverghese]

Yes, there is a reduction in NPSHr for hot water service.

Is the entire suction piping sloping upward continuously towards the de aerator? Show us the PID with process controls, process datasheet also. Have you got calcs on NPSHa ?

Does this pump have an NPSH inducer impeller ? A dedicated booster pump upstream is the preferred option instead.

 

[Doug Hunter]

Rreach,

Regarding your latest question: one approach, if you have the time and money, is to artificially restrict the suction for station A, C or D, to see if you can re-create the failure mode of station B. This is focused specifically on your desire to confirm cavitation.

I think you have enough data to say that something is wrong with station B. It might be worthwhile to tear it down and rebuild or replace it, or swap it with the hardware from stations A, C or D, to see if the problem moves with it. Do you have a diagram of the discharges coming out of each pump? Are they all plumbed together?


Best regards,
Doug Hunter
Altarium Technical Consulting

https://www.altarium.us/

 

[Rreach]

I attached a sketch for the system and basically, we have pressure gauges on both suction and discharge of the pump and individual recycle line with flow reading for each pump.

As per design
Suction head 41.2 m, pumping temperature 130 degC. SG 0.935. Vapor Pressure = 2.7 barA (29.4 m), NPSHa = 11.8 m NPSHr = 3.7 m.

But the site reading shows
Suction pressure 3.3 barA (35.48 m), pumping temperature 114 degC. SG 0.948. VP = 24.97 psi = 1.72 barA (18.49m), NPSHa = 16.99 m NPSHr = 3.7 m

I should still have sufficient NPSH in this case unless the pumping temperature is actually higher, then the margin will be reduced.
Suction pressure 3.3 barA (35.48 m), temperature 126.7 degC. SG 0.939. VP = 35.43 psi = 2.44 barA (26.49 m), NPSHa = 8.99 m NPSHr = 3.7 m.

 

[LittleInch]

Whats the flow rate and the pump curve for each pump. NPSHR varies with flow.

Also how effective is your de-aertaor?

Pump B is the end of the inlet system - how do you regulate and equalise flow between each pump in operation?
how many operate at the same time?

Are they all fixed speed?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Doug Hunter]

Also, would like to see the actual pressure and flow gauge readings for all 4 stations. Are the recycle line valves manual, and if so how are they set for the 4 stations?

Best regards,
Doug Hunter
Altarium Technical Consulting

https://www.altarium.us/

 

[TenPenny]

The fact that B is the furthest from the deaerator is very suspicious. Remember that the only NPSH you will have is the elevation from the pump centerline to the level in the deaerator, it is very suspicious that B is being starved, as it is furthest from the DA.

 

[georgeverghese]

How did you figure temp may be 126.7degC ? Analogue temp gauges are notorious for wrong readings, do you have a temp transmitter readout for BFW at de aerator exit ? Flowing velocity down the main suction header should be max 1m/sec approx when all pumps are running full throttle.

Do friction drop calcs match up with PG (when was this last calibrated?) readout at B pump suction at max flow ? We also see the PG is upstream of suction strainer - have you checked this strainer? What is suction specific speed? - should be > 9000 preferably.

 

[Rreach]

Thanks to all for your suggestions.

We will do some check (strainer cleaning, instrument checks etc) before taking another test run. Will update when I get more data.