PD PUMPS TO CENTRIFUGAL PUMPS - SERIES OPERATION 3

[Energeer]

The scenario is this --- PD pumps (Gaso triplex plunger) pump crude oil via a 6" pipeline into a breakout tank (crude gathering system). From this point, 2 centrifugal pumps in series (mult-stage pipeline pumps) are used to draw from the tank (using a booster pump) and into another pipeline (8") for delivery downstream.

The tank now needs to be placed out-of-service for inspection/repairs. The operator wishes to bypass the storage tank and feed the centrigugal pumps directly from the PD pumps.

I am concerned due to the puslations of the PD pumps (even though the last pump is 15 miles upstream) and the the elaborate controls likely to be required, e.g., PRV on suction side of centrifugal pumps, startup and shutdown philosophy. The centrifugal may "want" to pump more, but will be deprived by the PD pumps's steady flow - even with a programmed control valve on the pumps' discharge.

I've always tried to avoid this due to what I feel would be an unstable hydraulics scenario, however, I've never tried it. Before I suggest against it and instead suggest the more expensive installation of a replacement breakout tank, has anyone tried this or have any thoughts to share on this.

Thank you in advance.

 

[stanier]

Could you design a bypass from the discharge of the centrifugal pump to the suction of the pump. This would allow any excess from the centrifugal pump to be bled back thus equalizing the net output of each pump. With only a % recirculating there shouldnt be a heat build up problem.

Efficiency is the only concern but you may sacrifice this for the short term the tank is out of action.



Geoffrey D Stone FIMechE C.Eng;FIEAust CP Eng

http://www.waterhammer.bigblog.com.au

 

[Energeer]

Yes, perhaps an orifice plate bypass would help..... Thank you.

 

[pipehead]

good plan guys, the 15 mile between the station tells me there will be no pulsation that interfers, we have the same arrangement less than 2 miles apart (a little more compressible hydrocarbon mix).

 

[BigInch]

Doesn't look like a real problem. Nothing wrong with normal pulsations. Are they normal?

Either use the recirculation configurations mentioned above if you can, or another option would be a control valve in the centrifugal's discharge closing% to hold minimum suction pressure. Not very elaborate.

http://virtualpipeline.spaces.msn.com

 

[JJPellin]

BigInch has the method that I would suggest. Pinch on the centrigual discharge to hold constant suction pressure and the centrigual pumps will hold to the same flow rate as the PD pumps.

A spill-back line would work, also. But I would suggest a controlled spill-back that is set to hold constant suction pressure to the centrifugals. An orificed spill-back could be a trial and error process that might take several iterations. And if there is any change in product conditions, it might not work well over time.

Johnny Pellin

 

[Energeer]

Thank you all. You have given me some very good ideas. I plan on installing a controlled spill-back line. I also expect flow through the spill-back line at all times. Otherwise, the PD pumps will be "pumping through" the centrifugal pumps. Please let me know if this does not make sense.

 

[BigInch]

Hold on. Repeat. Pulsations hitting the centrifugal or even "Pumping through" the centrifugal will not be a problem. Don't really know why you're worried about it, as long as pump suction and discharge pressures are not exceeded.

If a pulse ever makes it that far (doubtful, they are reduced by line friction and pipe expansion along the way), you will simply, momentarily and probably minutely lift the suction and discharge pressure a bit. The flow through the pump will also increase, perhaps reducing the differential pressure as well (with no net change in discharge pressure (?).

This is something that I do all the time and I highly doubt you need to be worring about this at all.

http://virtualpipeline.spaces.msn.com

 

[BigInch]

If you want to give me the data, I'd be willing to run a transient simulation as proof.

http://virtualpipeline.spaces.msn.com

 

[Energeer]

I agree that the PD pump inherent pulsations should be dampened out sufficiently over the length of the pipeline. My concern is that the flow rate of the PD pumps will exceed the flow rate capacity of the centrifufal pump.

 

[rzrbk]

I would have to guess that your centrifugal pump capacity greatly exceeds your PD pumps' capacity, otherwise, it could not keep the tank pulled down during current operation. But these flows need to be know before you design your system.

Without knowing more, your most likely scenario is a bypass control valve back to the centrifugal suction to keep suction pressure from dropping too low, as discussed above. Depending on the PD flow suppying this pump, you will have to watch for overheating due to recirculating much of the flow.

As this system will not be very forgiving of flow swings without the surge tank, you might want to consider tripping the centrifugal pump on low suction pressure. (loss of pd pump)

 

[BigInch]

If you don't know the capacity of your centrifugal, I also suppose you don't have the curve.

As rzrbk suspects, with VERY good reason, I also doubt the recips capacity are greater than the centrifugals. Once again, It doesn't look like you will have a problem. However to be sure, it would be an excellent idea to know a little about your system before you even think of modifying anything. I thought you might have divulged some details by now. So then, what exactly do you know about it?. Recip's flowrate, recip discharge pressure, centrifugal curve, centrifugal's shutoff pressure, discharge pipe diameter off the centrifugal, or do you only know you have 15 miles of 6"? If you don't know any of this stuff, you don't really need to be thinking about modifying the system ....and, with that in mind, so far your asking all the questions in the wrong place. We don't have those answers. Maybe better to come back later when you've got the details.

http://virtualpipeline.spaces.msn.com

 

[Energeer]

I do have all the info. necessary to design the controls, etc., but I can work through the details - my question was strictly general in nature. You and others on this thread have provided me with some good insight and comfort regarding this arrangement as workable. I will proceed accordingly - thanks again.

PS - Regarding pump flow rate capacities, the PD pumps on average are lower than that of the centriful pump. However, there are peak periods when a high number of PD pumps are running simultaneously and the total flow rate would exceed that of the centrifugal pump's capacity.

 

[BigInch]

Peak periods? Questionable. The system can move only as fast as the slowest mover, so there would never be any need to operate 2 parallel trains of recips or any configuration yielding a net higher flow going to one centrifugal pump with a lesser capacity. If you got a peak flow into the tank from the recips, it wouldn't do you any good. You'd only have to shut the recips down and wait for the centrifugal to catch up before you could restart the recips and net average flow over a complete cycle would be exactly the same.

As you will only be in the modified configuration for a short time, I suggest you intentionally limit the capacity to that of the 1 centrifugal. Too bad you simply can't just turn down the speed of the recips, or are they constant speed electric driven?

In any case don't accept peak ops when in the temp configuration.

http://virtualpipeline.spaces.msn.com

 

[dcasto]

what should happen is; when multiple PD pumps run, the line should be at nearly MAOP. At the high pressure, the suction to the centifugals should increase and the head they need should decrease. This will back the centifugals up their curve and get more flow. HOWEVER, don't know all the hydaulics...... just a scenerio to look at.

 

[BigInch]

dcastro, I don't follow when you say "if at near MAOP, the head they need should decrease". I would think the downstream pipeline curve head value would also increase, thereby resulting (even if centrifugal suction pressure increases) in a net increase in total pump head in order to maintain a higher flowrate. Maybe you mean the differential head decreases, no? In any case, more flow requires more net input head, recip power needed increases, centrifugal differential head decreases and power increases thereby eventually limiting the flowrate to some maximum.

When reaching MAOP (of the downstream pipeline), centrifugal discharge pressure would probably be held constant by a control valve, the net head in the system (might remain the same, but) would not therefore increase and the downstream pipeline would not increase flow, so no movement on any curve would occur.

http://virtualpipeline.spaces.msn.com