Jump to content



Featured Articles

Check out the latest featured articles.

File Library

Check out the latest downloads available in the File Library.

New Article

Product Viscosity vs. Shear

Featured File

Vertical Tank Selection

New Blog Entry

Low Flow in Pipes- posted in Ankur's blog

6

Pump Startup


23 replies to this topic
Share this topic:
| More

#1 waelchouchani

waelchouchani

    Junior Member

  • Members
  • 15 posts

Posted 08 July 2024 - 10:08 AM

Hi,

 

During the startup of the new RO plant, we encountered a problem with the feed pump. Seawater is supplied to the plant through a pressurized network at 250 psi. At the plant inlet, a pressure control valve (PCV) is used to reduce the pressure to a set point of 90 psig.

The startup procedure involves flushing the RO membranes with seawater. This is done by passing the water on the feed pump without starting it through the multimedia and cartridge filters. After 5 minutes of flushing, the feed pump is started with the flow control valve (FCV) set to a 10% opening. However, shortly after startup, a pressure drop occurs in the suction line, causing the pump to trip due to low suction pressure.

The problem seems to be related to the high water demand during startup, which overwhelms the PCV, preventing it from maintaining the required pressure. We need to find a solution that does not involve hardware modifications.     



#2 Pilesar

Pilesar

    Gold Member

  • Members
  • 1,432 posts

Posted 08 July 2024 - 04:12 PM

Can you start the pump with the discharge valve closed and then open it gradually?



#3 fallah

fallah

    Gold Member

  • ChE Plus Subscriber
  • 4,987 posts

Posted 08 July 2024 - 10:09 PM

Hi,

 

During the startup of the new RO plant, we encountered a problem with the feed pump. Seawater is supplied to the plant through a pressurized network at 250 psi. At the plant inlet, a pressure control valve (PCV) is used to reduce the pressure to a set point of 90 psig.

The startup procedure involves flushing the RO membranes with seawater. This is done by passing the water on the feed pump without starting it through the multimedia and cartridge filters. After 5 minutes of flushing, the feed pump is started with the flow control valve (FCV) set to a 10% opening. However, shortly after startup, a pressure drop occurs in the suction line, causing the pump to trip due to low suction pressure.

The problem seems to be related to the high water demand during startup, which overwhelms the PCV, preventing it from maintaining the required pressure. We need to find a solution that does not involve hardware modifications.     

 

Hi,

 

To get a proper and accurate response, please upload a simple sketch of the system you described...



#4 shvet1

shvet1

    Gold Member

  • Members
  • 394 posts

Posted 08 July 2024 - 11:52 PM

in lack of info I bet on hydraulic surge (aka water hummer) in suction line

https://www.ksb.com/...hammer-data.pdf

https://go.ksb.com/r...vDi6Qdf1pXF-60G


Edited by shvet1, 09 July 2024 - 12:28 AM.


#5 waelchouchani

waelchouchani

    Junior Member

  • Members
  • 15 posts

Posted 09 July 2024 - 02:51 AM

Hi

Please find attached schematic drawing.

Attached Files



#6 waelchouchani

waelchouchani

    Junior Member

  • Members
  • 15 posts

Posted 09 July 2024 - 02:52 AM

 

Hi,

 

During the startup of the new RO plant, we encountered a problem with the feed pump. Seawater is supplied to the plant through a pressurized network at 250 psi. At the plant inlet, a pressure control valve (PCV) is used to reduce the pressure to a set point of 90 psig.

The startup procedure involves flushing the RO membranes with seawater. This is done by passing the water on the feed pump without starting it through the multimedia and cartridge filters. After 5 minutes of flushing, the feed pump is started with the flow control valve (FCV) set to a 10% opening. However, shortly after startup, a pressure drop occurs in the suction line, causing the pump to trip due to low suction pressure.

The problem seems to be related to the high water demand during startup, which overwhelms the PCV, preventing it from maintaining the required pressure. We need to find a solution that does not involve hardware modifications.     

 

Hi,

 

To get a proper and accurate response, please upload a simple sketch of the system you described...

 

 

Hi,

I attached the schematic drawing.



#7 waelchouchani

waelchouchani

    Junior Member

  • Members
  • 15 posts

Posted 09 July 2024 - 02:55 AM

Can you start the pump with the discharge valve closed and then open it gradually?

Hi,

I tried this, but the problem is once we close the FCV the pressure upstream the pump increases which causes the PCV closure. Please refer to the schematic drawing.   



#8 waelchouchani

waelchouchani

    Junior Member

  • Members
  • 15 posts

Posted 09 July 2024 - 03:00 AM

in lack of info I bet on hydraulic surge (aka water hummer) in suction line

https://www.ksb.com/...hammer-data.pdf

https://go.ksb.com/r...vDi6Qdf1pXF-60G

Hi,

The problem is that during the startup of the feed pump there is a high demand for water within a short time, which causes a reduction in pressure because the PCV opening is slow and cannot follow the abrupt startup of the pump.



#9 breizh

breizh

    Gold Member

  • Admin
  • 6,436 posts

Posted 09 July 2024 - 04:12 AM

Hi,

Did you check with the engineering company who supplied the process and equipment? You should have a start-up procedure and control philosophy.

To me the set-up pressure (50-100 psig) is too wide, you controller may be too slow to react. 

 

Note: You may need an additional PI transmitter at the pump suction, cascading on your pressure set point to anticipate the pressure drop. 

 

Good luck

Breizh 



#10 waelchouchani

waelchouchani

    Junior Member

  • Members
  • 15 posts

Posted 09 July 2024 - 05:19 AM

Hi,

Did you check with the engineering company who supplied the process and equipment? You should have a start-up procedure and control philosophy.

To me the set-up pressure (50-100 psig) is too wide, you controller may be too slow to react. 

 

Note: You may need an additional PI transmitter at the pump suction, cascading on your pressure set point to anticipate the pressure drop. 

 

Good luck

Breizh 

Hi,

As per the startup procedure, after flushing the FCV should be adjusted to 10% and after that start the pump. but this created a depression at the suction side and the pump trips by low suction pressure by the pressure transmitter installed at the suction header (the trip set point is 22 psig).

The inlet set point is adjustable as per design it is 100 psi. 



#11 fallah

fallah

    Gold Member

  • ChE Plus Subscriber
  • 4,987 posts

Posted 09 July 2024 - 08:37 AM

 

 

Hi,

 

To get a proper and accurate response, please upload a simple sketch of the system you described...

 

 

Hi,

I attached the schematic drawing.

 

 

Hi,

 

Thanks for sending the sketch...

 

As i reviewed, the problem is mostly due to wrong sizing of the PCV such that at high flow passing, it probably goes to choked flow at a pressure lower than the set pressure (90 psig as you mentioned) hence cannot handle the demand flow rate along with pump trip due to low suction pressure.

 

To follow finding exact root cause, please let's have the performance curve (outlet pressure vs flow passing) of the PCV provided by relevant vendor also the value of highest demand flow rate... 



#12 snickster

snickster

    Gold Member

  • Members
  • 65 posts

Posted 09 July 2024 - 05:05 PM

This is my take on the possible issue.

 

PCV's are not typically installed on suction lines of pumps but I can see why in your installation it is required.

 

When the system is being flushed the PCV is in a very open position.  Then when flushing is stoped and the FCV is set at 10% open, the PCV goes to a very closed postion but not fully closed, since while still with pump not started flow is low through RO unit with only 90 psig pressure downstream of PCV.  When pump is started flow increases rapidly causing an instantaneous evacuation of volume in the suction system volume between the pump and the PCV due to the imcompressibility of the liquid, and the such small quantity of liquid removed from the suction required to drop the pressure.  The PCV cannot respond fast enouh to open wider to offset the pressure drop by letting more liquid in to the suction system to balance the liquid being removed,

 

I think the only way around this without adding hardware to the system would be to put a time delay on activating of the suction pressure switch, 5 seconds or so would allow transient pressure to stabilize and should be ok for pumps for very short periods of start up.

 

Hardware modification might include installing a bladder volume tank on the suction to prevent transient pressure drop as it will quicly fill any liquid suddenly removed giving the PCV time to respond.


Edited by snickster, 09 July 2024 - 05:09 PM.


#13 snickster

snickster

    Gold Member

  • Members
  • 65 posts

Posted 09 July 2024 - 05:26 PM

Thinking about it further it is likely that the reason the operating procedure requires starting up the pump with the FCV only 10% open was to try to prevent the low suction pressure which is happening as I previously described, considering the designers of the system foresaw this potential issue.

 

But there is also an issue when you go from 10% open of FCV under manual set point to automatic operation of the FCV.  Seems like there will also be a transient there too.  So time delay may need to be set to provide enough time for FCV to go to fully automatic operation. Otherwise you might want to slowly manually increase the set point of the FCV until the design set point is reached to minimize transients.


Edited by snickster, 09 July 2024 - 05:41 PM.


#14 waelchouchani

waelchouchani

    Junior Member

  • Members
  • 15 posts

Posted 10 July 2024 - 12:42 AM

This is my take on the possible issue.

 

PCV's are not typically installed on suction lines of pumps but I can see why in your installation it is required.

 

When the system is being flushed the PCV is in a very open position.  Then when flushing is stoped and the FCV is set at 10% open, the PCV goes to a very closed postion but not fully closed, since while still with pump not started flow is low through RO unit with only 90 psig pressure downstream of PCV.  When pump is started flow increases rapidly causing an instantaneous evacuation of volume in the suction system volume between the pump and the PCV due to the imcompressibility of the liquid, and the such small quantity of liquid removed from the suction required to drop the pressure.  The PCV cannot respond fast enouh to open wider to offset the pressure drop by letting more liquid in to the suction system to balance the liquid being removed,

 

I think the only way around this without adding hardware to the system would be to put a time delay on activating of the suction pressure switch, 5 seconds or so would allow transient pressure to stabilize and should be ok for pumps for very short periods of start up.

 

Hardware modification might include installing a bladder volume tank on the suction to prevent transient pressure drop as it will quicly fill any liquid suddenly removed giving the PCV time to respond.

Hi,

You described perfectly the issue.

We tried increasing the time delay for the suction pressure trip to 3 seconds, but it did not resolve the problem. The pump vendor confirmed that the maximum time delay is 3 seconds.



#15 waelchouchani

waelchouchani

    Junior Member

  • Members
  • 15 posts

Posted 10 July 2024 - 01:45 AM

In my opinion, one potential solution is to perform the flushing with the MMF drain valve open. This would allow the PCV to remain in an acceptable position while the FCV is closed to 10%. Subsequently, the pump could be started, and the drain valve closed simultaneously. This approach aims to keep the PCV open during pump startup, enabling it to handle the high water demand.

 

Concerns and considerations:

  • Surge pressure: Closing the drain valve rapidly could indeed create a surge of high pressure. 

  • Drain valve failure: While the check valve downstream of the drain valve offers some protection, a failure to close the drain valve during pump startup could still damage the pump. 

What do you think?



#16 waelchouchani

waelchouchani

    Junior Member

  • Members
  • 15 posts

Posted 10 July 2024 - 11:15 AM

 

 

 

Hi,

 

To get a proper and accurate response, please upload a simple sketch of the system you described...

 

 

Hi,

I attached the schematic drawing.

 

 

Hi,

 

Thanks for sending the sketch...

 

As i reviewed, the problem is mostly due to wrong sizing of the PCV such that at high flow passing, it probably goes to choked flow at a pressure lower than the set pressure (90 psig as you mentioned) hence cannot handle the demand flow rate along with pump trip due to low suction pressure.

 

To follow finding exact root cause, please let's have the performance curve (outlet pressure vs flow passing) of the PCV provided by relevant vendor also the value of highest demand flow rate... 

 

Hi

Please find attached the design cases of the PCV. The pump operating point is 1350 gpm @ 1015 psig

Attached Files



#17 fallah

fallah

    Gold Member

  • ChE Plus Subscriber
  • 4,987 posts

Posted 10 July 2024 - 01:21 PM

 

 

 

 

Hi,

 

Thanks for sending the sketch...

 

As i reviewed, the problem is mostly due to wrong sizing of the PCV such that at high flow passing, it probably goes to choked flow at a pressure lower than the set pressure (90 psig as you mentioned) hence cannot handle the demand flow rate along with pump trip due to low suction pressure.

 

To follow finding exact root cause, please let's have the performance curve (outlet pressure vs flow passing) of the PCV provided by relevant vendor also the value of highest demand flow rate... 

 

Hi

Please find attached the design cases of the PCV. The pump operating point is 1350 gpm @ 1015 psig

 

 

Hi,

 

The data sheet isn't included the performance curve of the PCV. Please let's have this curve or, at least, the value of choked flow of the PCV.



#18 breizh

breizh

    Gold Member

  • Admin
  • 6,436 posts

Posted 10 July 2024 - 10:52 PM

Hi,

Can you share the pump curve and FCV data sheet? 

Breizh



#19 waelchouchani

waelchouchani

    Junior Member

  • Members
  • 15 posts

Posted 11 July 2024 - 12:31 AM

 

 

 

 

 

Hi,

 

Thanks for sending the sketch...

 

As i reviewed, the problem is mostly due to wrong sizing of the PCV such that at high flow passing, it probably goes to choked flow at a pressure lower than the set pressure (90 psig as you mentioned) hence cannot handle the demand flow rate along with pump trip due to low suction pressure.

 

To follow finding exact root cause, please let's have the performance curve (outlet pressure vs flow passing) of the PCV provided by relevant vendor also the value of highest demand flow rate... 

 

Hi

Please find attached the design cases of the PCV. The pump operating point is 1350 gpm @ 1015 psig

 

 

Hi,

 

The data sheet isn't included the performance curve of the PCV. Please let's have this curve or, at least, the value of choked flow of the PCV.

 

Hi,

The curve is not available with me but the as per the PCV datasheet it is linear curve with rated CV of 400



#20 waelchouchani

waelchouchani

    Junior Member

  • Members
  • 15 posts

Posted 11 July 2024 - 12:37 AM

Hi,

Can you share the pump curve and FCV data sheet? 

Breizh

Hi,

Please find attached the FCV datasheet and the pump curve.

Attached Files



#21 waelchouchani

waelchouchani

    Junior Member

  • Members
  • 15 posts

Posted 14 July 2024 - 05:54 AM

In my opinion, one potential solution is to perform the flushing with the MMF drain valve open. This would allow the PCV to remain in an acceptable position while the FCV is closed to 10%. Subsequently, the pump could be started, and the drain valve closed simultaneously. This approach aims to keep the PCV open during pump startup, enabling it to handle the high water demand.

 

Concerns and considerations:

  • Surge pressure: Closing the drain valve rapidly could indeed create a surge of high pressure. 

  • Drain valve failure: While the check valve downstream of the drain valve offers some protection, a failure to close the drain valve during pump startup could still damage the pump. 

What do you think?

Hi,

Could you advise please?



#22 snickster

snickster

    Gold Member

  • Members
  • 65 posts

Posted 14 July 2024 - 09:13 PM

I think what you describe with the drain is very tricky.  You have to open the drain valve enough that the PCV goes open equal to what pump normal flow open is or greater so that PCV can stay where it is or only close further when the pump is started up, and would you do this with FCV 10% closed or FCV set to normal flow rate opening?  Also you would need to close the drain just at the instant the pump is starting.

 

I believe a better option would to start the pump with discharge valve closed as someone suggested.  Yes the PCV should go closed if you do this because without flow (with pump discharge valve closed and pump running) the PCV would need to close completely. to maintain 90 psig set pressure.  Then as you gradually open the discharge valve to slowly increase the flow the PCV opening should be able to keep up with the flow increase.  You could monitor the suction pressure to make sure you do not open the block valve too fast.

 

The question is do you start up with FCV 10% open then increase opening in steps while increasing opening of the block valve in steps too or do you set the FCV for design opening at normal flow shown on the data sheet of 50% to 64% open with block valve closed and then just keep opening the block valve slowly until the design set point flow is reached which should be at 50% to 64% open as shown on the data sheet.  I would try the latter first and see if you can keep the pump from tripping.  Start pump with FCV 50 to 64% open then very slowly open the block valve to increase the flow until the flow equals the theoretical flow corresponding to the set point then go into automatic mode and then your FCV should be right where it should be for the flow you have reached.

 

Of course the pump would need to be operated in the dead head condition initially while gradually increasing flow, where the flow then will be below pump minimum flow for the time required to stabilize the system.  I would discuss this with the pump vendor.  This is a very high HP, flow and pressure pump so it may be very sensitive to operating for any time below minimum flow.

 

Otherwise if there is a way to recycle the flow during pump start up then gradually closing the recycle line block valve this may work too.  Is there a pump recycle line?  When you flush the system before starting the pump do you flush through the pump or do you have a pump bypass line that you flush with bypassing the pump.  If you have a bypass line then you could use this bypass line as a recycle line while starting up.  Then you could try setting the FCV at 50% to 64% open while slowly closing the recycle valve. However starting the pump in the recycle mode with FCV 50% open may still cause a spike in flow that the PCV cannot keep up with.   If there is still a spike in low suction pressure then you may also a need to close the pump discharge block valve and slowly open in combination with slowly opening the recycle valve.

 

Another thing is I notice that your NPSH required from your pump curve is 39 ft @ 120% QR.  I belive your 22 psig set point for your pressure switch gives you about 83 ft of NPSH so I think you can actually go down to 5 psig set point for water at 80F and still have about 44 ft. NPSHR with vapor pressure of about 0.5 psia.  This may help a little with lower set pressure of pressure switch.


Edited by snickster, 14 July 2024 - 10:54 PM.


#23 breizh

breizh

    Gold Member

  • Admin
  • 6,436 posts

Posted 14 July 2024 - 11:25 PM

Hi,

I don't want to be rude, but you cannot ask Operation team to "play " with manual valves to start a process. I'm worried about your design. On a maintenance point of view and operation it will have been much better to have smaller pumps in parallel to cope with the start-up (gradually reach the flow set point starting one pump after the other). For reliability having a spare installed, 3 pumps in running and 1 in standby mode, will have been nice.

I understand you are short in Capex. To me you should sit with your manager / design team and operation team to discuss the startup and modification needed to resolve the issue.  

 

Good luck

Breizh 



#24 carlos infante

carlos infante

    Brand New Member

  • Members
  • 9 posts

Posted 18 July 2024 - 07:28 PM

star the pump with the FCV valve in manual ( control loop not working), when the flow stabilized, set in automatic the FVC with a ramp of opening.  DCS programent should do that.

 

It is normal procedure for large water pumping flow

 

Carlos






Similar Topics