Triplex Cavitation with NPSHr met? 2

[tkasp]

Good afternoon all, I am having trouble explaining a cavitation issue where we apparently have adequate NPSHa for an existing triplex plunger pump moving produced water.

Pump:

https://www.pentair.com/content/dam...a-60m-60h-triplex-plunger-pump-sell-sheet.pdf

Aplex 60M 1.500" plunger @ 314 RPM (Constant speed on/off)
Manufacturer provided NPSHr = ~ 15 ftWC at these conditions.

Problem is: the pump is audibly cavitating according to operations and the manufacturer's tables indicate ~22 gpm flow at the above conditions. We are seeing only 11.74 gpm (2.67 m3/hr), which may be explained by reduced capacity due to the cavitation. This pump is tank fed with a low level shutdown set @ 0.5m above the pump suction elevation so I am checking NPSHa to confirm we have the required suction head.

NPSHa = Hatm - Hvp + Hst - Hf + Hvel

Atm pressure:

Hatm = 89 kPaa ~= 29.8 ftWC​

Vap Pressure:

Hvp 2.34 kPa ~= 0.78 ftWC (@ 20C, high to be conservative)​

Velocity Head:

Hvel (2.67 m3/hr in 3" pipe) ~= 0.004 ftWC​

Frictional Head Loss:

Hf = 0.35 ftWC (calculated frictional losses from tank nozzle to pump suction at 2.67 m3/hr in 3"& 4" piping)​

Static Head:

Hst = 1.64 ftWC (minimum, usually operate at higher level)​

NPSHa = (29.8 - 0.67 + 1.64 - 0.35 + 0.004) ftWC
NPSHa = 30.4 ftWC

In this regard NPSHa >> NPSHr, what other explanations can explain the reported cavitation and loss of capacity? What am I missing?! My first inclination was to size a single stage centrifugal charge pump to increase NPSHa on the triplex suction but it seems I shouldn't have to? Thanks in advance for any advice.

 

[JJPellin]

The tendency of this pump to cavitate is also very much dependent on the length of the suction line. Recall, that the liquid in that line has to accelerate and decelerate as the pistons stroke. I have personally seen this on a triplex pump very similar to yours. You need to include acceleration head in your formula above. Acceleration head includes piping Length in the numerator. How long is your suction line and how many elbows are in that line? According to my old copy of Cameron Hydraulic Data, a standard NPSH analysis "begins to lose validity if the length of the suction line exceeds 50 feet, simultaneous operation of more than two pumps, more than 3 bends in suction line or complex mixtures of fluids." I seem to recall good papers on this subject at the International Pump Users Symposium over the years. I was prepared to install a pulsation damper on our system to decouple the pump from the long suction line. In the end, we converted to a centrifugal pump

Johnny Pellin

 

[bimr]

A rule of thumb is to oversize the piping by a factor of 3 for plunger pumps.

 

[georgeverghese]

See GPSA for calculation of additional acceleration head loss. This loss may be reduced by including a pulsation dampener installed close to the pump inlet manifold.

 

[tkasp]

Thank you guys for the response. Acceleration head definitely seems to be the big culprit here, I appreciate you folks catching that. We do indeed have a decently long run of suction piping (40' straight pipe, 102' of equivalent length including fittings)

Estimating Hacc from GPSA:

Hacc = LVnC/Kg = (40')(0.51 ft/s)(314 RPM)(0.066)/[1.5*32.174]
Hacc = 8.76 ftWC

NPSHa = Hatm - Hvp + Hst - Hf + Hvel - Hacc
NPSHa = (29.8 - 0.67 + 1.64 - 0.35 + 0.004 - 8.76) ftWC
NPSHa = 21.64 ftWC

I suspect actual acceleration head would be worse as there is six 90's on that suction piping from the tank as well. If I were to estimate with equivalent length that would put acceleration head up closer to potentially 20 ftWC, bringing NPSHa below NPSHr.

Back to my original plan then to look at installing a single stage centrifugal on the suction as providing 12 gpm at a TDH of even 35 ft would be easily doable in a relatively cheap small pump. I suppose a suction pulsation dampener could be explored, but I am afraid it wouldn't provide enough relief and force me into a boost pump regardless...

I welcome any additional thoughts or concerns and appreciate your guys time and expertise, thanks for the help!

 

[georgeverghese]

Moving produced water? So there may be dispersed oil in the water, since most produced water systems are bad actors? If so, you should use the vap pressure of the dispersed oil phase.

 

[tkasp]

It is produced water, however the tank is treater fed and samples very clean. It's a good point, I'll be sure to add margin of safety on the vapor pressure head. This tank is skimmed.

 

[LittleInch]

also you seem to be using the wrong flow rate.

You should be looking at what the pump is trying to do, at least in the initial movement of the plunger. On your chart this is around 4.8 m3/hr, so double the flowrate and velocity you're using.

ditto your solution needs a bigger pump. At 315 rpm your pump should be pumping about 21.5 GPM... Also give it as much pressure as you can reasonably afford.

"Produced water" often contains a lot of dissolved gases of different sorts which impact the vapour pressure.

And try and reduce the number of elbows



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

 

[tkasp]

Fair points LittleInch, I was ruminating on it more last night and settled on chasing adequate head performance at 27-40 gpm to allow for that higher flow requirement on the suction stroke of the plunger, with the rationale being that as far as flow is reduced due to plunger limitations, we're just safer on the head.

The current disposal flow rate, even at reduced capacity, is sufficient for production requirements. However, this pump has a terrible control scheme (i.e. no control scheme) where operations has matched the speed roughly to the production and it operates continuously in that balance. I can't afford installing VFDs etc and it seems that it would be harder to find a booster in these conditions near BEP judging from curves (Tbd with vendor). So I am aiming to speed the triplex up to 24-27 gpm, size the booster to provide adequate head at 40 gpm, and then drive the entire pump train in on-off service from the tank level transmitter (high start - low stop with HH/LL SD). Not an ideal control system, but this is a small battery with marginal economics. I am not going to get any money to be fancier.

I am considering the below pump that I got as a budgetary quote.

As you mention to give it as much pressure as I can reasonably afford and I've seen rule of thumb to aim for 25 psi for triplex suction; if I set a desire to put 60 ftWC on the suction flange of the injection pump and accounting for minor frictional losses of a conservative 1 ftWC. a TDH of 60 ft wc + 1 ft wc (Hf) - 26 ftWC (Est new NPSH at minimum tank level) = 35 ft wC should put me in a good place.

I think the pump below is putting me in a good spot with a 6.125in or 6.5in impeller, no? I will go back to the vendor to see if there is a better pump that could operate closer to BEP, but I think this is a good start? This should provide well in excess of the plunger pumps flow requirements with NPSHa >> NPSHr with room to grow and the required addition and modifications would fit within my budget.

I appreciate your time and advice, pumps have not been an area of high exposure for me.

 

[LittleInch]

Be careful here as you seem to be mixing NPSH ft which is essentially absolute pressure, with guage pressures.

So if you want 25 psig at the pump suction then you're going to need at least that at the pump discharge flange. Forget about NPSH at this point.

It looks like at min level in the tank you only have 1.64 ft head so less than 1 psi inlet pressure which more or less matches your calculated frictional losses.

So to get 25psi at the pump inlet flange ( sounds a good plan to me), your pump needs to deliver a differential pressure of 25 psi or about 57 feet.

This piddly little pump is only doing 35. You need a bigger pump.

At this level don't fixate on BEP. The efficiency here is horrendous, but it's only going to be 1 hp or so so don't fret too much. Small pumps are not efficient.

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

 

[tkasp]

Fair point, my bad as I'm fairly sure the other thread I read was 25 psia, so I took at as a straight head conversion from NPSH. It's taking getting used to thinking of TDH, differential head, NPSH, and their respective pressures for sure. I'm overcomplicating it at this point.

Just for my own understanding: the pump I showed above WOULD work here as it is delivering NPSHa > NPSHr for the injection pump, correct? The only discrepancy being that we are adding additional head (35ft to 57ft) in order to meet the rather arbitrary 25 psig on the injection pump suction flange as additional margin of safety. I suppose i'm getting too granular. For most experienced folks then to spec this pump we know we're gonna meet NPSHr on the booster from our previous NPSHa calcs, we know we need whatever 15-25 ftWC to meet NPSHr for the injection pump, and we know that 25 psig is a good safe suction pressure for triplexes this size; so take 1.5-2x your plunger flow rate and take 25 psig (or higher if desired for safety margin) minus any positive static head for your and run with it?

For BEP I was more worried about vibration etc, but for this small pumps really as long as you're in the yellow so to speak call it good and start worrying about it when you get into the 30-50+ hp units?

 

[LittleInch]

It's usually only process engineers that use psia...

It just needs to be consistent as to what you use.

Yes it should work but if you're going to put a pump in put one in that has a bit more pressure. The benefit if positive pressure (i.e. above atmospheric) is that you can install a pulsation damper if required.

Flowrate is up to you to determine. Its best not to oversize your inlet pump too much as you have a margin from your rated flow of 20-30% before you run off the curve and if you run it too low you can run into trouble. But otherwise yes your description is about right.

You always need to look at efficiency and power as some companies will off you 30% and others 60% for the same duty, but don't go too much on BEP until you get >30hp is about right.

Looking back you appear to have positive pressure ignoring acceleartion head ( staic + 1.6ft, friction losses 0.3 (though you were using the wrong flowrate so this might go up). If you added a large pipe (say 12") vertically teed upwards to the max height of liquid level in your tank right by the pump inlet you might be able to reduce your acceleration head a lot and get fluid in there. The more water in the tank the better and you would need to see if air is being gulped in, but if you make the pipe big enough this shouldn't happen. You might need a little more water in the tank to make it work.

Easy to do and see if it makes a difference before buying a pump and wiring it in etc etc.

A slightly less convoluted pipe run from your tank will also help.

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