Pump Coupling - Pump Deadheaded

[macmet]

I'm not particularily familiar with "large" pumps, so forgive me if this is a ridiculous question.

We have a boiler feedwater pump for a 500bhp boiler. We were having some issues with a valve and flow transmitter on the inlet of the boiler, so we were runnign the system and dead headed the pump a few times.

My question is, does this cause excessive stress on the pump coupling to the motor?

The reason I ask is that we recently noticed that coupling has seen excessive wear and had to be replaced. Could this have been caused by us closing that valve?

Any input is appreciated. I will check back to provide any other information that could be helpful.

 

[DRWeig]

How long did you run your pump with the discharge valve closed? Could it be that you spun it long enough to boil the water in the pump casing and create some cavitation (with associated vibration)?

Good on ya,

Goober Dave

 

[tr1ntx]

Assumption is that it is a centrifugal pump:

When you shut off the discharge flow, the water that is inside the impeller housing doesn't just stop. The impeller is still acting on it, so water is still moving from the impeller ID to the OD, but it can't go on around the volute and out the discharge, so it flows down the sides and re-enters the impeller ID. This changes the forces that the impeller, and thus its shaft, experience. You're putting that HP into a fairly small, closed system. Different impeller designs handle these forces differently; some accomodate it, some do not.

 

[macmet]

Thanks for the replies so far.

I'm not sure if we ran it "long enough". How long would you say is long enough to see cavitation?

It is a centrifugal pump. Direct drive, 30HP.

 

[DRWeig]

macmet,

tr1ntx's reply makes a lot of sense too. Could be a geometry thing.

How long to boil the water? If your 30 HP pump is, say, 70% efficient, that means about 30% of the input power goes to heating the water -- not a problem as long as the water is once-through. Closing the discharge valve confines a small volume of water in there, plus runs the efficiency down very low. Together, you can get boiling.

Input power is much lower at dead-head, but so is efficiency. Even if it's only drawing 5 HP from the motor, you might have half of that (or more) going to water heating, the rest goes to moving the water from inlet to outlet and back again. Half of 5 HP is 1.9 kW, and if you only have about a gallon of water in the pump housing it will rise almost 13°F per minute. Heat loss through the housing (if uninsulated) will help limit the rise and extend the time, but if well insulated, it might take less than 10 minutes to get to boiling temperature.

I made a bunch of assumptions in the paragraph above, maybe one of the more experienced pump people can correct anything I stated that might be out of line. I have a little practical experience, having boiled a pump once a long time ago. My pump didn't hurt its coupling, but it did blow steam out the seals. It was a 200 HP pump, high speed impeller that was small and thus not much water volume in the housing. I didn't time how long it ran deadhead, but it wasn't too awful long, maybe 5 minutes?

Let's see who else pitches in with better knowledge and experience...

Good on ya,

Goober Dave

 

[macmet]

Ya, you're right. Tr1ntx's response could be the cause too. And if the alignment wasn't right, I could only imagine that that would excalate those forces further.

In regards to deadhead time, it probably would have been deadheaded for 5-10 minutes at the absolute max, but it would have happened a few different times during the day.

Thanks again for the responses, I just wanted to know for a warranty claim if it could be our fault, and sounds like it's possible. It's not expensive, but something to avoid we'll try to avoid doing in the future. Looks like we'll suck up the cost of this one.

 

[flexibox]

I would check the following:
1 Pipe stress. Disconnect the suction and discharge pipes and see if they spring out of alignment.
2 Check alignment figures. Have you checked for soft foot.
3 Sounds like you have a rubber coupling, if the problem persists then check concentricities of hubs to ensure everything is running true to shaft centre line.

 

[ccfowler]

This is very small pump compared to my experience with boiler feed pumps (multi-thousand hp each), but the principles are still the same. If very low flow rates must be tolerated, it is essential that a by-pass line be installed to return a minimum flow of the discharge flow to an upstream feedwater heater (or a heat exchanger to cool the flow) to assure an adequate minimum flow through the pump to avoid both overheating and damaging unintended forces. It will be necessary for you to determine what the tolerable minimum flow will be and how (and where) the temperature rise must be accommodated.

Yes, the energy usage of the pump will be increased by this by-pass flow, but you will find it to be a trivial burden compared to the burden of pump system damage. Also, since the problem has been severe enough to damage the shaft coupling, it is likely that there will be damage to the pump internals as well. It is likely that pump performance and efficiency have suffered from this type of operation and the resulting damages.

Valuable advice from a professor many years ago: First, design for graceful failure. Everything we build will eventually fail, so we must strive to avoid injuries or secondary damage when that failure occurs. Only then can practicality and economics be properly considered.

 

[Artisi]

Back to the coupling, unlikely that running dead headed is causing the coupling failure, power input is at its lowest at this point. More likely misalignment, wrong coupling selection.

 

[ccfowler]

My thinking on the coupling failure is that overheating in the pump could cause axial force imbalance within the pump envelope and axial shifting of the pump shaft greater than the coupling can handle. Depending on the motor bearing types, there could be some axial shifting of the motor rotor as it finds its electrical center (differing from its axial position during alignment procedures). All of this being based on wild speculation, of course, since we know precious little of the exact details of this situation. Most likely, the coupling problem could be mitigated by replacement with a coupling that can accommodate a greater range of movement.

Although the pump heating during blocked flow operation is surely less severe here than in the BFP's of my experience base, it is also likely that this pump has been operated in this mode far longer than may really be prudent to avoid internal damage regardless of the coupling damage. I've seen plenty of smaller pumps turned into mechanical water heaters due to inadequate flow (usually by virtue of unfortunately mis-matched pumps operating in parallel). These pumps never operated long in this mode without suffering serious internal damage (what a surprise!).

Valuable advice from a professor many years ago: First, design for graceful failure. Everything we build will eventually fail, so we must strive to avoid injuries or secondary damage when that failure occurs. Only then can practicality and economics be properly considered.

 

[electricpete]

I agree with others the concerns about deadheading a pump usually have nothing to do with the coupling.

If there is axial force on the pump, it's range of motion will usually be restricted by the pump thrust bearing, but that doesn't rule out fatigue stresses to the coupling as I think suggested above.

We do have one set of single stage centrifugal pumps driven by sleeve bearing motor through shim pack coupling where we see large 0.125" axial movement of the motor shaft at a frequency of maybe 1 hz which occurs only when the pump is at low flow (goes away at higher flow). I figure that the broadband / imapcting forces from small movements of the pump rotor (<0.020" pump bearing clearance) act as base excitation for the spring / mass system consisting of the coupling / motor, and excite the resonant frequenc of that system which willbe quite low due to flexible shim pack coupling and large motor mass. That low-flow operation recurs periodically for periods of hours to days, but we have not seen cracking of the shim packs.

You might want to tell us what type coupling and describe the bearing configurations on pump and motor.

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

 

[Bunji51]

If your coupling is looking rough, you may want to take a peak at your thrust bearing as well to see is there is any damage there also.

Vibration analysis can tell you the story of what's going before the next failure occurs. As others have indicated it sounds like an alignment issue.

 

[1gibson]

Deadhead would have the highest radial loads on the pump, and pump will have highest vibration. OH or BB type?

Also an off chance of higher pressures affecting the piping and nozzle loads, causing movement of the pump relative to baseplate which would of course affect alignment. Grouted baseplate, or I beam / C channel?

 

[desertfox99]

I don't understand why you would intentionally dead head any pump. I could understand throttling (closing) the valve so in order to try and see a change in the flow indicator. By means of increasing the head and looking for a lower flow rate. However, I don't see what would be gained by fully closing a discharge valve and dead heading a pump.

 

[BigInch]

Hot water?

I don't see where anyone mentioned deadheading intentionally.

 

[Artisi]

Seeing as how the OP hasn't appeared since the 3rd Dec.2010, think we can now assume this post is also deahheaded.

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.)