Centrifugal Pump Dead Head

[BenLaidlaw]

Hi all,

I am placement student completing a project for a valve testing facility, where a water system is to fill up valves at low pressure (3-5 bar). The water comes from a large tank through a pump and enters a flanged valve (closed system).

My question is whether running a centrifugal pump repeatedly up until zero flow will have adverse effects on the pump long term.

Ie is it more suitable to use an accumulator in the system to fill up the flanged valve.

Thanks,
Ben

 

[Artisi]

Yes - but.
How long has the test facility been in operation, 5 minutes or 5 years?
Why are you concerned?
What size pump is it?
How many cycles a day / a year?
How long does the pump run at deadhead per cycle?
How does the pump sound / operate when running deadheaded?

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.)

 

[BenLaidlaw]

Hi Artisi, thanks for replying!

I should mention that I am designing a complete new system and only currently in design stage. Ie test facility not in operation and I am still to obtain components.

The requirements are approx. 10 gpm and relatively low pressure (5 bar).

The water system will be used 2-3 times a day 365 days a year.

As I have not purchased any components I do not currently have data of the pump running deadheaded.

My main concern is whether to design the system with just a centri pump, and fill the flanged valve up to 5 bar. Or - if this has adverse effects on the pump - to create the system with a charged accumulator and allow that to fill the valve.

Cheers,
Ben

 

[LittleInch]

What you probably need to do is cost the difference between an accumulator, which will slowly reduce in pressure as flow out occurs or a simple Automatic re-circulation valve or ARV set at approx. 30% of the full flow of the pump back to the tank. This is essentially a fit and forget whereby someone can turn the pump on and not worry about the pump until he' doesn't need water anymore.

This is but one example -

http://www.schuf.com/products/pump-recirculation-valve

Most pumps of that sort of size could dead head for a few minutes without significant long term harm - the issue is more one of how long and how do you ensure that the pump isn't left running for much longer periods by accident - maybe use a timer to give say 5 minutes of pumping at each press of the button.

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

 

[Artisi]

If it was left to me, I would buy a high quality process pump from one of the major manufacturers, select a pump that is a bit oversized, run it with a reduced diametet impeller and operate it at deadhead 2 or 3 times a day. Probably last for virtually ever.

I phave seen process pumps running on filter presses at deadhead for hours on end with surprising results - the
major issue was erosion / wear from the abrasive product - otherwise the pumps performed mechanically ok.

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.)

 

[HPost]

Every time you shut the flow off the seal will have no flow to remove the heat generated by the friction between the sealing elements.

Repeated flow shut off will cause filming in the seal faces and the seal will start to leak. if the temperature gets too high, the bellows around the seal will melt and the seal will leak.

Yes, in theory it is possible to dead head a centrifugal pump, but it doesn't do them much good.

 

[electricpete]

I guess you are trying to increase internal recirculation for cooling. But it is a fixed / limited volume of fluid involved. There does not seem to be much mechanism for heat removal. If it doesn't work, who will your employer blame?

I have nowhere near the pump expertise that Artisi does but my inclination from a conservative engineering standpoint would be to avoid deadheading unless you find manufacturer's literature that explicitly allows it.


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

 

[BenLaidlaw]

Thanks for all replies.

Inevitably if the system fails it will be my responsibility.

Additionally the water system has to produce a steady pressure of 3 bar for valve seat testing. I would imagine an accumulator would have the advantage of producing a more steady state flow to the valve test piece.

Thanks for all comments, I shall cost a pump and accumulator set up, a pump and ARV set up, and speak to specific manufacturers about deadheading certain pumps.

Ben

 

[Artisi]

You will have flow thru the seal if you allow partial flow thru' the seal to waste.
Boear in mind the spec is 2 to 3 runs per day, it is not continuous operation.
Also run the pump 4 pole speed not 2 pole.


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]

The simple alternative is to use the tank to fill the valve and then pressurise it with a hand operated piston pump. Easy squeezy - simple, cheap and fool-proof

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.)

 

[1gibson]

What is the total system volume, and how did you get 10 gpm? What kind of flow would you get from the tank due to gravity? If pumping the initial fill only saves 5 minutes twice a day, then it can't be worth it. Gravity fill, then pressure up, maybe an air operated diaphragm pump.

 

[bimr]

Why reinvent the wheel?

Buy a valve testing stand. There is much more to the valve testing stand than just the pump.

http://www.efco-dueren.com/upload/prospekte_ps/englisch/e_35b12_prospekt_ps_15_30_50_75rev01.pdf

 

[BenLaidlaw]

bimir - the test unit is already in place, however I am to design a water recirculation system to fill the valves either in the test unit, or larger flanged up valves.

Ie water into valve from tank, down to sump, and a submersible sump pump back to the original tank. The design I am currently considering is the best means to fill up a potentially large valve to a 3 bar pressure.

The tank volume is approx. 500 gal with 5 metres of 2" pipe to the valve and 10 gpm is not essential. I shall look into whether the valve could be gravity filled and pressured up however it would need to be raised due to the tank output being below the valve.

Thanks,
Ben

 

[bimr]

A recirculation loop will be more reliable and consistent than a gravity loop. You don't have to worry about filling it up either. Set the pressure in the loop equal to your test pressure.

Install a simple recirculation loop:

http://kb.eng-software.com/display/ESKB/Sizing+a+Minimum+Flow+Recirculation+Line

 

[BenLaidlaw]

Thank you for all the responses

Bimir/ LittleInch - Although I have no experience in designing recirculation loops, from reading it does not seem to difficult. Ie size an orifice to set the pressure in the loop equal to 3 bar; which will provide a steady 3 bar pressure to the valve test piece.

http://files.engineering.com/getfil...d07-d6ec76dce69d&file=recirculation_line.docx

 

[LittleInch]

An orifice is one way to do this - you end up oversizing the pump a little as the bypass flow continues under all flow regimes - which is where an ARV comes into it's own - but for this sort of thing - it's simple, cheap and does the job.

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

 

[BenLaidlaw]

Although an ARV would be ideal, for the small application and budget it cannot be justified. After doing some thinking about the most fool proof way to provide a constant 3 bar pressure in the valve test piece I have decided to use a manual globe valve in the minimum flow spillback loop as opposed to a restriction orifice.

I am currently sizing a pump for the application, and although I have some experience of sizing pumps from university days, I was hoping you could give me feedback on my proposed method.

Due to the pressure against the valve test piece being the main concern (to provide 3 bar pressure against the seat for 5 minutes), ie only flow for this time is through spillback loop; the pump will be operating towards the left half of the curve.

For pump:
Approx flow rate needed = rated flow to fill up against valve + pump minimum flow for spillback.

Total pump head needed = required head for main line + required head for spillback line.

Where required head for main line = Total Static Head + Head Loss (frictional and miscellaneous) + Head required to produce 3 bar pressure on valve test piece (30.6 metres).
And required head for spillback line = Total Static Head + Head Loss (frictional and miscellaneous; where globe valve flow coefficient is included is miscellaneous losses).

Thanks in advance for comments and advice,
Ben

 

[Artisi]

Pick any small pump that will give you the pressure you need before the allowable minimum flow rate, don't oversize the pump as you will then have problems burning off excessive pressure / flow thru your control valve.
For me, I would probably select something to run a bit left of the minimum flowrate - remember - you are talking about 2 to 3 runs a day for 5 minutes - you are not throttling a 2000 Kw recirc. pump.

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.)

 

[LittleInch]

For a fixed speed pump you have two operating points -

the first is your 3 bar outlet pressure at your minimum flow, so your static head minus the friction losses to the point where your valve take off is,

the second is the flow into the valve to fill it up plus the flow going round your fixed re-circ line (unless you isolate that while filling). this pressure will be less than the first, but you probably want as much as possible so aim for say 90% of the static pressure and see what you get back from the vendors. This will probably be best written as your "rated" duty point.

~The point about an orifice plate is that you can pre-calculate the size you need fairly well, whereas the globe valve needs ot have a flow measurement on it somewhere to set it up and then you need to fix it in place (take the handwheel off) otherwise someone will either turn it off or open it up - people just can't resist fiddling with things....

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

 

[bimr]

It may take some time to find the optimum pump.

These pumps are close:

https://www.pollardwater.com/pdf/Spec_Sheets/AMT_Sprinkler_Booster-elec-Specs.pdf