NPSHR for Parallel Pumps 1

[Aggiedog]

Hello there! Been a lurker for many years while in school and finally took the plunge to register. I will try to contribute as much as I can (though I still have alot to learn).

I understand the very basics of what series and parallel pump configuration does to net head on flow capacity. However, for parallel configurations, I am a little unsure about NPSHR. I would assume that you would add each of the NPSHR for each pump and that would be that correct?

Also, if anyone knows of any good rotating machinary book I could use as a reference, I would appriciate it.

Thanks in advance!

 

[JJPellin]

You are not correct. The NSPH available to a pair of pumps piped in parallel is not affected by the presence of the other pump except for the fact that the flow to the second pump may increase the total flow through part of the suction piping and thus may increase the flow losses. But this is not always the case. The NPSH required for each pump is a property of the pump. It is not changed by the fact that the pump is piped in parallel or series or alone.

An example may help to illustrate my points. I have two pumps piped in parallel. They both tie into a common suction manifold and a common discharge manifold. There are no control valves, orifices or other imposed pressure drops in the piping that connects the two pumps. Rather than running one pump at its rated flow of 1000 gpm, I run both pumps at 500 gpm each. The total flow is unchanged. The NPSH available to each pump increases very slightly. The flow through the common piping is still the same and so it has the same pressure drop and flow losses. But the flow through the individual suction line to each pump is running at a lower flow, lower velocity and so it has a lower pressure drop. For a typical API centrifugal pump, the NPSH required by each pump will drop as a result of the lower flow through that pump. So, I have a higher NPSHa and a lower NPSHr. In theory, I end up with a better NPSH margin and the pumps are less likely to cavitate.

If this is the case, why don’t I prefer to run all of my pumps in parallel? First, running at 50% of rated flow, each pump will have a lower efficiency. Running further from the best efficiency point, each pump will have higher radial loads on the impeller. Each pump will be running further back on the curve. Depending on the shape of the curve, the pumps may not be sharing the flow equally. Especially if the curve is flat in this region, any slight differences in the condition of the pumps will drive one to carry a higher portion of the load and one to carry a lower portion. The pump at lower flow could drop below the minimum flow required to avoid suction recirculation cavitation or other destructive conditions.

It is much more common in the real world for pumps running in parallel to be less reliable than pumps running individually. But it really depends on where each individual pump is running relative to its best efficiency point and how much NPSH margin it has above required.

As a reference to the principals involved, I would suggest any of the following:
“Pump User’s Handbook” by Bloch and Budris
“Centrifugal Pumps – Selection, Operation and Maintenance” by Karassik and Carter
“Cameron Hydraulic Data” edited by Heald


Johnny Pellin

 

[BigInch]

Johnny, I'm wondering about this part you said, "First, running at 50% of rated flow, each pump will have a lower efficiency." Of course its true if each pump is sized for 100% of total station flowrate. However when I need two pumps, with no standby capacity, I would size the pumps such that 100% capacity for each pump is 50% of the total station flow. Doing that would of course allow both pumps to run at BEP = 50% of total station capacity. If it was required that one pump runs and the other should provide 100% standby capacity, they would not normally ever be run at the same time, so would never run at 50% of BEP.
I'm confused why you said that. What's up?

**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[wimple]

BigInch

I guess that it is refering to the fact that the efficiency tend to increase with flow increase. Is n it Johnny?

Wimple

 

[Artisi]

I also looked at the 50% comment and was a bit puzzled however I decided Johnny P was probably thinking of 100% pumps so that there is always full capacity if one oump goes down. My first preference for critical installations is normally 3 x 50% pumps - but of course final selection is based on a number of points and assumptions.

 

[JJPellin]

In my particular example, a single pump was able to produce the entire station flow. Even if it was running very close to the end of the curve, dropping to 50% would usually result in a drop in efficiency. You are, of course correct that there are examples when double pump is more efficient. This is probably more common on pipeline applications. In our refinery, I can only think of three pairs of pumps out of about 2000 where this is the case. But, that is the problem when trying to give a generic answer to a generic question about generic pumps.

Johnny Pellin

 

[JJPellin]

Big Inch, I missed the second part of your comments. In our plant, we often run out of capacity with a single pump. So, for the sake of the example, our production people want 101% of the total flow that a single pump can produce. So, instead of living with only 100%, they start the second pump and run them at 50.5% each. In a pipeline application, this might be foolish. But, in a refinery, that extra 1%, could be worth $ millions. It is a constant struggle to balance the production needs with the reliability of the pumps.

Johnny Pellin

 

[BigInch]

Well what you're saying is certainly true, IF you can make millions more with a 1% increase in flow. However you have to realize your perspective is skewed to the point where you are not really worried about pump efficiency at all, so (and I mean this in the most friendly way) "don't act like you are". Where pump efficiency does matter, because it is a critical part of profit, you would never run 2 at 50.5% BEP. In a refinery profit isn't the direct result of,

Profit = Q * (TransportCost/Q - K * Q/eff)

but at a pipeline that is profit and that equation quarantees you'll never ever never even think about doing it. So, I quess what I'm saying is efficiency must be significant to you in order to be able to mention it in the first place. Know what I mean?

**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[JJPellin]

I fondly hope that some day we will consider efficiency and energy costs in our decisions. But right now, we really don't. The economics of production overwhelm the energy consumed by the pumps.

We have a crude unit that runs a booster pump at 100,000 barrels per day. If they had an opportunity to sell more product from this unit, increasing the unit capacity to 101,000 BPD by running both booster pumps might pay out like a slot machine. The extra 1,000 BPD could generate a net profit of $3 per barrel. And that is truly net profit after accounting for raw materials costs and all processing costs. Over the course of a year, that 1,000 BPD would generate $1.1 million in additional profit. The energy cost for running two pumps at lower efficiency might be wasting 200 HP. At our current energy costs, that 200 HP would cost us about $70,000 per year. Who among us would not spend $70,000 to make $1.1 million? Even if I add in a penalty for reducing the pump reliability, I can't compete with the production incentives. The only thing that could trump this decision would be a significant impact on safety or environmental compliance


Johnny Pellin

 

[Aggiedog]

Ah, thank you for the clarification. I have been buried deep into this issue and can not see the forest among the trees.

 

[BigInch]

The book you're looking for as far as centrifugal pumps is concerned is Stephanoff "Centrifugal and Axial Flow Pumps"

You do know about McNally?

http://www.mcnallyinstitute.com

You could do some lurking around there too.

**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[Artisi]

A further good reference is Pump Handbook - Karassik - krutsch - Fraser - Messina. ISBN 0 07 033302 5

 

[divya90]

hai

For transfer of crude oil from storage tanks to offshore facility about 5 km and the route is more or less flat.

My question is whether we need booster pumps before get in to shipping pumps to transfer oil?

or

we can instal directly the shipping / transfer pumps nearer to tank farm for transfer of oil?

what coudl problem will anticipate if we avoid booster pump.

the flow capcity required is about 45,000 bbl / hr at the rate of 325 psi.

Can any one suggest and comment on that.

Regards,

vijay

 

[BigInch]

divya, You should start a new thread please.
Basically no way to tell without knowing the entire configuraion affecting the hydraulics (headers, valves, viscosity and pipe diameter). You might only need transfer pumps. Supply the details in your new thread.


**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/