Does NPSH required for gear pump for negative pressure at suction side?

[Tundra11]

Hi guys,

Need your help and advice on this:
There is a pump system in my company that uses a gear pump to pump a melted solid from a holding tank above the pump to another storage tank. The holding tank is under negative presure(close to full vacuum), and the pump has been frequently experiencing no flow problems.
I checked the calculations for the NPSH required and found that the NPSH available is less than that of NPSH required.
However, the pump datasheet provides a suction pressure which when the pump is turned on, provides the NPSH required.
So my question is: when calculationg for NPSHA in this case, do we consider the suction pressure provided by the pump?
If not, I suspect that cavitation has been occuring in the pump leading to damage, loss of clearance and subsequently no flow of the pump.
Thanks all for your valuable input!

 

[MikeHalloran]

NPSH required and found that the NPSH available is less than that of NPSH required.

Bingo. The pump is starved, and will die young, and all of its replacements will suffer the same fate.

You need to increase the NPSHA, e.g. by moving the tank higher so as to provide more gravity head, or by making the suction piping much larger so as to reduce head loss in the suction piping, or both.


Mike Halloran
Pembroke Pines, FL, USA

 

[Artisi]

Think Mike is on the case here.
However, one point for your own clarification, "So my question is: when calculation for NPSHA in this case, do we consider the suction pressure provided by the pump?"
NPSHa is never a function of the pump, it is entirely a function of the pressure available from the system to ensure the pressure available exceeds NPSHr of the pump so that cavitation is not a problem.

NPSHa = Hs _Hvpa + Hst - Hfs

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.)

 

[samanmalith]

Net Positive Suction Head Calculation

NPSHr is the net positive suction head required by the pump which is a characteristic of the pump design. Varies from pump to pump and size. This is shown in the pump performance curve. NPSHr has to be < NPSHa for cavitation free pumping.
NPSHa is determined by the system we are operating and is a function of pipiing size,length, number of valves,etc and atmospheric pressure,

NPSHa = Hatm - Hvap + Zs -Hi
where
Hatm = Atmospheric head (m slurry)
Hvap = Vapour head of water ( slurry)
Zs = Static suction head (m), Suction lifts are negative in sign
Hi = Combined head loss in suction pipework

 

[Tundra11]

Thanks Mike, Artisi and Samanmalith for your replies!

Yes, thats what me and my colleague thought so too when we went throught the NPSH calculations.

Because this system was not designed by us originally and the pressure of the holding tank is negative which makes for an interesting case study with all the problems we have encountered during its operation so far.

So, my main confusion was whether the original design team factored the suction force provided by the pump in their NPSH available calculations so as to overcome the NPSH required requirements and whether is this the correct way to calculate it.

So far, it looks like they made a costly mistake for us to rectify.

Mike, just for information, we probably can't increase the height of the holding tank (too costly) or the diameter of the pipe (pressure drop is minimal). Therefore I'm considering changing pump type to a self priming pump to overcome this NPSH problem.

Would appreciate any other solutions for consideration too. Thanks!

 

[samanmalith]

Your pumping media is melted solid. It may be still hot when you are pumping.So the vapor pressure may further reduce the NPSHa. Changing the pump to a self priming type is the way to solve this issue.

Cheers,


Malith De Silva
Weir Minerals Australia Ltd

http://www.weirminerals.com

 

[LittleInch]

Tundra11 - I concur with the posters above. Please stop talking about "the suction force provided by the pump ". The pump provides no such force. The force in absolute units driving the fluid into the pump is the pressure force above the liquid (in your case fairly low) plus any height difference (positive if the fluid level is above the pump) minus vapour pressure of your fluid minus friction losses in the pipe.

Also your statement "However, the pump datasheet provides a suction pressure which when the pump is turned on, provides the NPSH required." What are you talking about? Please post the data sheet or the relevant section as this makes no sense at all, but seems to be at the root of your misunderstanding.

As you state your pump is below your holding tank a self priming pump will do nothing for you. These are for where your pump is above the liquid level and needs to prime or "lift" fluid into the pump. You don't have that problem.

You only have a few options as has been noted by Mike:
Raise your negative pressure above your tank - even an increase of 0.1 bar might be enough
Increase liquid level in your tank - can you run it higher?
Lower the pump or raise the tank. I've had to put pumps in 3m deep pits before now when trying to pump out tanks at very low pressures
Get a pump with a lower NPSHR - what is your NPSHA and R??
Consider two pumps to move the NPSHR to the left (usually lower) operating at lower speed

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[Tundra11]

Hi LittleInch,

I would have to disagree with you there on the pump not providing a suction force. This pump is a gear pump (positive displacement pump), meaning that it will definitely provide a suction force at inlet side if pump is turned on.
And I know there is a suction force because when the pump is turned off, no liquid is able to flow out to the discharge side via pump drain line(due to the negative suction pressure in holding tank), however when pump is on there is flow.
I have attached the pump specifications for you guys to take a look at.
Perhaps my understanding of pumps may not be right given I have limited experience with rotating equipments.
Look forward to further clarifications thanks!

http://files.engineering.com/getfil...be21-d72e1f35ff49&file=Pump_Specification.pdf

 

[Artisi]

Tundra11, You are either a very game man or naïve to argue / disagree with the people in here about pumps adding suction force. The only force on the inlet of a pump is the NSPHa, which is what is left over from atmospheric pressure after all losses are accounted.
If you look at your data sheet you will not that inlet pressure is - 0.5 (I assume this is -0.5m) this means the inlet is under vacuum - that's why there is no flow when the pump is turned off - you will further note that the NPSHr is 2m, meaning the lowest level in the tank should be no lower that 2.5 m above inlet and preferable more to ensure the system is safeguarded.

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]

Tundra11.

The pump does not provide a suction "force", what it does when it is working is create a pressure at the inlet below its normal static condition - in your case this is a pressure lower than atmospheric, hence no flow from your drain . If you think about this properly, the pump does not "suck" your liquid, the actual force propelling the fluid into the pump all comes from the upstream side - fluid will only flow somewhere if there is a differential force propelling it, be that air pressure, a pump or gravity. In your case the normal atmospheric pressure pushing down on the liquid (1 bara) is not there or very low (0.1bara??) If the height of the liquid in the tank is say only 1m (how much is it?) then you only have 2m at best. However you then need to subtract the vapour pressure of your liquid and any friction losses in the pipe work (you don't say how far or what size / velocity or viscosity of the fluid) to get NPSHA. You say you've calculated that - how about sharing that with us??.

Remember that we can only see what you tell us / show us so have difficulty providing anything more positive. A dimensioned sketch showing distances, height, sizes and flows often works well in these situations.

Your specification / data sheet shows an NPSHR of 2m. You will be very lucky to find anything much lower than this so I can only suggest that you look at some of the other alternatives we have mentioned above. The only I missed out, but Mike said, was make your piping as big as you can / reduce length / reduce bends, elbows, fittings to a minimum.

You have to do something or your pump will continue to not work at all or not work properly. You can initially tell everyone the design was rubbish, but you've now been given the task of making it better. You're getting some good pointers here, but you need to understand the basic physics behind it, come up with some options including some which might be costly, then put them in front of the person(s) who will decide which one to implement.

Await further info

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[moltenmetal]

Your pump needs some head depth of melted solid material above the suction: that depth is your NPSHA, assuming the pump is close-coupled to its melt drum and hence there is virtually no suction head loss due to friction. The operating pressure above that melt layer is irrelevant, no different than pumping any saturated (boiling) liquid. In polymer plants, the melted polymer product is often removed from vacuum devolatilizers or rotary film evaporators by means of gear pumps which spin very, very slowly. The liquid polymer is so viscous that flow under its own piezometric head is very slow, and the tips of the gears are literally grabbing polymer melt as it oozes down onto them.

 

[LittleInch]

The pump spec does have a very interesting 70-700 rpm speed variation. Are you trying to work this pump too fast? try the minimum speed possible first then see if you get flow and gently increase until you hit problems. However IMHO there's no substitute for metres head...

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[btrueblood]

What molten said. I recently ran a test on a ridiculous viscosity level for a product we build, to prove that it was possible - a gear pump that normally pumps a low viscosity (1-2 centipoise) liquid was instead used to pump a 10,000 centipoise liquid. Said liquid was cheapest to duplicate as corn syrup. It pumped okay, but the pump speed had to be reduced from 600 rpm down to about 20-30 rpm to maintain inlet flow. Fun test, but sloooo

http://wwww.

 

[LittleInch]

That's the other thing a bit odd on the data sheet - It says viscosity is supposed to be 4 to 20cP - a little thick, but not polymer standard (in the thousands of cP).

Maybe you just have the wrong pump.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[Tundra11]

Hi all,

I am sorry if I had offended anyone by disagreeing, no disrespect meant.

As to the fluid, yes it is a by product from the polymer reaction process and its viscosity changes greatly depending on temperature as well as on its composition.

As for the pump, we are currently running at around 350 RPM, which gets us the required flow from the holding tank.

Need your advice if my calculations below are correct:

Holding tank pressure: 7.2 kPaA

Total height of liquid above pump (normal conditions) : 8.04 m

Pipe inner diameter: 0.0274 m (1 inch schedule 40S)

Flow rate: 0.21m3/hr

Viscosity: 4.92 cp

Density of fluid: 1020 kg/m3

Total length of pipe to pump suction: 12.834m

Vapour pressure of fluid: 78 kPA
So:

Reynolds number = 1020 * (0.21 / ( pi * (0.0274)squared / 4)) = 305 (Laminar flow)

Pipe loss pressure = 32*( pi * (0.0274)squared / 4) * 12.834 * 0.00492 / (0.0274 squared) = 0.000144 Pa

NPSH available = 7.2/(1020*10) + 8.04 - 78000/(1020*10)- 0.000144 = 0.39m

So since NPSH required from the pump datasheet is 2m, I assume that there is insufficent head therby causing cavitation of the pump?

Thanks.

 

[MikeHalloran]

You've pretty much fenced off the solution space.

One thing that remains is a much larger pump runnning at a much lower speed.



Mike Halloran
Pembroke Pines, FL, USA

 

[Artisi]

Tundra11.
Don't think any of the big people here would be offended with your disagreement as most are always more than willing to assist.

As to the problem, certainly seems NPSHa is the problem, the reason why the tank level was put so far above pump inlet. Not having checked your calculations and taking everything is correct, the question is what is different from the original design and the actually application assuming that the system should have worked.

As to changing pumps, will leave that advise to others experienced with pumping this product with gear pumps.

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]

The odd thing for me is that your Vapour Pressure implies that this stuff is constantly vaporising in the 1" tube until it gets to about 0.5m above your pump, i.e. you have a two phase flow in your 1" pipe with bubbles trying to go one way whilst the liquid goes another - really quite strange, but maybe common in the polymer world.

Your calc looks good to me.

A bigger pipe (4" say) might help in reducing this effect, but on those figures you don't have much room to move unless you can change something physical like height.

As MM says, maybe you need to have two or three pumps rotating very slowly instead of one at 350 rpm or a bigger one going slowly. When it doesn't flow, try reducing the speed until it does, even if this is less than you want and just add more pumps (easy for me to say but I can't see your system).

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[btrueblood]

I can also see a nasty transient problem, unless the pump body and inlet lines are heated: specifically, the viscosity probably rises quite rapidly as the fluid temperature drops a few degrees, and the result is that, at startup with "cold" lines and pump, it's even more difficult to get the fluid to flow to the inlet face of the gears. I could believe that after the pump has moved a few pipefulls of liquid, the temperature comes up and you could run all day long. Try a slower startup speed, monitoring the outlet pressure or a flow switch on the outlet line, and only ramp up the pump speed when you can see evidence of fluid flow. Yes, this implies a variable speed motor or drive for the pump, but they have become cheap these days.

 

[moltenmetal]

If the holding tank pressure is 7.2 kPaA and the vapour pressure is 78 kPaA, you will have no liquid in the drum!

Something is wrong in the set-up of this. At LOWEST, the pressure at the TOP of the liquid level will be EQUAL to the vapour pressure of that liquid.

Otherwise, you have 8 m of liquid head to work with. Unless your suction piping is really long with a lot of bends, it shouldn't lose that much, and if the pump only needs 2 m of NPSHR, everything should work out fine.