High Pressure Pump, Min flow arrangement

[Watro]

Helow Experts,

We are designing a high pressure pump system and the schematic is enclosed with this thread. Apologies, if the subject has been discussed in depth in the past, but I couldn't find a satisfactory answer in old threads

As you see from the sketch, Is this normal industry and SAFE practice to connect the MCSF in the pump suction line? Is there any issue with temp rise (as I see from old threads). How to estimate that? What are the other design challenges and risks that may arise. Piping is Metallic, Super duplex. The ramp up speed could be reduced to 45 sec max, if need be. Fluid is particle free Seawater with 32 g/l chloride. No other harmful particles present. Note-I don't have a tank upstream. There is a Magnetic flow meter upstream of MSCF line connection, missing in sketch.

I dont have any previous exp. in this kind of MCSF return to suction line, so please advise all possible guidelines and issues that may arise

Thanks
Niss

 

[georgeverghese]

This recycle scheme looks a little unusual - how is the control scheme configured to close the remote - auto operated valve on the recycle line when at BEP ; ie what is BEP to the control system ?

Offhand, if you have no better return location for this recycle line, a typical approach to heat build up in the recycle loop would to be install a TAH alarm and TAHH trip within the loop. This will happen only when net forward flow through the pump is low. If you expect extended periods of operation at or below this low flow, then a VFD would be an obvious choice to avoid heat build up and pump trips.

Presume the FCV is to prevent the pump from going to end of curve and overloading the motor on startup - another option to this would be a soft start feature to ramp up the pump to full speed over 45secs- that would ease the load on your power supply also that would otherwise occur from high startup current at the 1800kW motor.

 

[JJPellin]

We have high pressure water wash systems that are piped in a similar manner. I have attached a sketch of one of these systems. In our case, we are not allowed to direct the spill-back to the supply vessel because of a particular process safety concern. Our pumps are pumping stripped sour water at 120 gpm and about 5000 feet of head. Our pumps are running at 8650 rpm. We have not had any issues or problems in our application. But, we almost always have uninterrupted forward flow. There is a remote isolation valve downstream in this system (not shown in my sketch). If that valve were to close, the spill-back would open to protect the pump for a short period of time. We are not allowed to trip the pump off in this event because a false trip could be the initiating event for a back-flow scenario. We count on a quick response by our operators to shut the running pump down or restore forward flow quickly. There is a chance that the pump could be damaged before they are able to resolve the situation. But, since it is water, the risks are relatively low and we have an installed spare pump, so production would not be impacted.

In short, this piping arrangement is not ideal, but it can be used successfully if necessary. I agree that alarms and trips on high temperature would be a wise addition, if your process allows that.


Johnny Pellin

 

[LittleInch]

I agree with George.

Your sketch is quite strange and contains a number of conflicting information.
It is also difficult to determine from your post what the process is and what and why you're trying to do.

E.g.
Data ( A pump curve would be much better)
Your sketch state a duty? point of 3785gpm, but then also a MCSF (min continuous stable flow??) return line at 2290 gpm - this seems very high compared to the rated flow
The pressures are odd as well - the FO (flow orifice?) states a diff pressure of 145 psi but the diff pressure listed is 692.
The purpose of the flow control valve is unclear and states 50 psi?

As drawn the system simply recycles a fairly fixed volume (FO is close to critical flow I would have thought) until the isolating valve is closed. Quite how this fits in with normal operation or startup isn't clear. If it's just start up then I'm not at all sure you need this arrangement as if you can get to full flow in 45 or even 120 seconds, then the pump will withstand short term low flow conditions usually.

If the forward flow remains at some flow at all time then you don't need to worry about the temperature, but if you have a deag head situation then you will.

Temperature rise is normally approximated by looking at both the inefficiency of the pump at the flow and pressure ( this is translated as heat into the fluid) plus the amount of energy "lost" from the fluid in the return process ( flow x head loss) which again you convert to heat energy and then work out from the volume being re-circulated how much energy is required to heat up that volume of water. Some energy is lost as sound and heat radiation and the pipe walls take up some of the heat, but it's a good starting point.

That is a big pump to dead head and have 2290 gpm re-circulating. I imagine you will get a pretty rapid heat up.

I hope your FCV is something better than a butterfly valve which is how it's drawn.

The system JJPellin puts forward is more normal where a control valve modulates on flow through the pump. The valve opens when flow falls below a set value ( the min flow) and closes when it exceeds this value until fully closed when all the forward flow exceeds the min flow figure.

Reciriculation back to the inlet can be a problem but make the volume of water as large as you can, monitor the temperature and if required build in a cooler, but this is only if you are regularly dead heading the system, but want a large pump to keep running for some time ( minutes) with no forward flow.

Give us a bit more info and you'll get more back, but we don't have enough to understand your system and its operation at the moment.


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

 

[Watro]

Thank you for all replies and apologies for late reply

Basically the high pressure pump is feeding a Reverse Osmosis system, which need to be started/stopped slowly. In order to maintain the min flow, the return arrangement to suction line is envisaged and there is no tank upstream. Please find the pump curve.

1. Please advise if MCSF line with control valve will be sufficient?
2. Is heat generation will be concern, if start up and shutdown is less than 2 minutes. Min flow scenario will be only during start up/shutdown, rest all time the pump will be operating >MCSF
3. What are all other consequences connecting the MCSF line to suction line (arrangement same as JPPellin)

Regards
Niss

 

[LittleInch]

What does your RO flowrate start flow over time look like? Can you post a graph?

The MSCF is intended to be for continuous flow, not a short term event like you describe.

I therefore wonder if you really need such a system. Try talking to the pump vendor and find out what the issues are about a ramp up flow rate over that sort of time schedule ( I assume you will get to MCSF in < 1 minute??)

If you've always got some sort of forward flow, or beyond say 10 seconds, then you shouldn't have an issue with temperature - there should be enough flow to dilute any temperature rise over time. you can work temperature by following the energy lost by the liquid converted to temperature rise of the water. A bit basic but gives you an idea of possible temperature rises.

You haven't answered many of the other points, such as why your diagram only has 145 psi across the FO - It will have the full pump differential pressure (approx. 650 psi) and for that sort of flow and pressure it really needs to be an RO, not a plate or multiple plates in series ( say 4 or 5). A single plate will wear out very rapidly and make a very loud noise with that sort of pressure drop and flow.

Any chance of modifying the motor speed or voltage during these events to give you a soft start / stop then connecting to give you a fixed speed to avoid the losses in a full VFD system?

What is the pressure ramp up like into your system? Does it need a certain min pressure to generate flow?

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

 

[Cecorros]

Adding to some of the comments and directly to your questions:

1. You've stated one rated duty - are there other duty points on your application? Normally you'd need this sort of arrangement when there are very low flows and you have to get them with the same pump. If this were the case and you design the recycling line so that at least the MCSF always passes through the main (pump) line, then it will be fine. Minimum flow required to avoid large temperature rises and overheating will be lower than the MCSF.

2. Strongly doubt that overheating will be a problem during start-up and shut-down. Hence I support previous comments highlighting that the pump will withstand short term low flow conditions.

3. Should you really need the recycling line that you've sketched, then one thing can be a reduction on your NPSH available - but from the numbers written you've got plenty and taking that the fluid is at ambient temperature, there can't be issues in this area.

 

[georgeverghese]

Temperature buildup during recycle is usually only a concern when net forward flow is very low, much lower than MCSF.

 

[JJPellin]

We have perhaps 10 or 12 reverse osmosis systems. None of ours are as large as yours. But, all of them start and stop as you described. All of them have the pumps spill-back to the suction piping rather than back to a supply tank. To the best of my knowledge we have never seen a problem with overheating as a result of this configuration. There need to be good procedures that limit the amount of time that the pump is left in full spill-back. If they are not able to establish some forward flow within a few seconds, they need to shut back down. I would not anticipate any problem with your system.

Johnny Pellin

 

[georgeverghese]

Are you sure the pump flow will always be greater than MCSF all the time ? That may be the case for normal operation only. Have you considered any abnormal failure scenarios, which should include plant operator error also? There will be severe piping vibrations on this line if flow drops below MCSF on a 1800kW pump - the diagram says this recycle line will be closed after the pump reaches BEP flow after startup.

For such a high powered pump, it would be preferable to install both a min flow recycle loop ( which is active all the time), and an independant low flow trip. Since this is only water we are dealing with here, the FSLL could possibly be configured off the same FT / FIC that is used for the min flow loop.