Pump Minimum Flow Recycle Return Line

[Sawsan311]

Dear All,

I have always understood that as a good Engineering practice, pump minimum flow recycle line (irrespective to the recycle scheme involving continuous line with RO or automated flow control valve) shall return back to connect to the suction source drum or to the suction header (if a source drum is not available). Connecting directly to the individual pump suction shall be eliminated and not followed to avoid turbulence in the pump suction with associated non-uniform distribution/separation of the flow ahead of the impeller eye in addition to overheating.

However, I have seen several schemes adapting the direct return of the minimum flow recycle line to each pump's suction line (without any cooler) while relying on a suction high high temperature trip as well as maintaining a minimum straight piping length upstream the pump inlet (although a reducer is sometimes unavoidable) .

Have you come across any lessons learnt or operational issues associated with returning the pump recycle line directly to the respective pump's suction piping instead of the common suction header and do you agree to its acceptance as an optimized design approach considering less piping length towards the suction header.

I provided a picture for visualization from the below site:

http://kb.eng-software.com/eskb/pip...ices/sizing-a-minimum-flow-recirculation-line

Thanks

Regards,

 

[LittleInch]

A case where "Good Engineering Practice" meets the challenge of cost and practicality.

The cases above seem to be where you have zero or very low forward flow. In those cases you are probably better off having a timer and then cutting off the pump once you get no forward flow after say 30 seconds.

Min flow is usually about 25-30% or so of the max flow so velocities are lower than full flow.

If you're at least 5D back from the pump you should be ok.

flow back to source or the header is better, but each situation is different and these sorts of things are usually picked up on the HAZOP.

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

 

[EdStainless]

If we couldn't get to a tank, drum, or header then the second choice was 50d upstream. If we couldn't do that we wouldn't allow continuous operation in bypass. These were just internal guidelines.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[georgeverghese]

Returning recycle line back to pump suction line is only done when there is no other practical solution. In such cases, yes, a TSHH on the pump will be required. I have seen this only on fluids with low vapor pressure, and where there are other means to minimise the possibility of the pump operating < MCSF.
If your pump may be operating often at less than MCSF, then make the effort to tie in the recycle line at the source vessel.

 

[MJCronin]

The discussion of centrifugal pump protection under blocked outlet conditions, should be broadened... IMHO

First, not all pumps require this active bypass protection. Positive displacement pumps must be protected with a spring loaded relief device or by a rupture disc. Only centrifugal pumps above a certain energy level require a bypass

https://blog.craneengineering.net/what-is-a-centrifugal-pump-minimum-flow-bypass-line

https://www.eng-tips.com/viewthread.cfm?qid=388813

Where there is a massive amount of energy input into the liquid (Powerplant Boiler Feed Pump), extreme care must be placed into the design of a bypass system. I have specified and purchased boiler feed pumps and always required the pump vendor to supply a properly sized and tested ARC valve to be included in the installed system

http://www.wermac.org/valves/valves_arv.html

https://www.emerson.com/documents/a...omatic-recirculation-yarway-en-en-5198076.pdf

https://fetterolfvalves.com/pdfs/13.pdf

Second, there are various ways to protect these pumps besides installation of an active bypass. Installed safety devices such as Pump motor shutoff on high discharge pressure (via pressure switch) or Pump motor shutoff on Low-flow (via flowmeter) may be less expensive solutions to consider. Power Monitors are inexpensive devices that can signal deadhead conditions. Centrifugal Pump damage by blocked flow should be part of every system HAZOP evaluation

https://www.eng-tips.com/viewthread.cfm?qid=395580

Third, there should be some thought given to materials selection and pump physical configuration under conditions of overpressure. It is generally agree upon that cast iron pump casings and CI impellers are a bad idea for pumps that can be deadheaded...particularly water pumps. There have been cases where a water pump has been deadheaded, the contained water turns to steam and the pump casing explodes ! .... Under conditions of overpressure, some centrifugal pumps will relieve themselves by destroying the seals

Keep us in the loop and tell us about your final solution

Best Regards

MJCronin
Sr. Process Engineer

 

[Sawsan311]

Thank you for all for your great inputs,

LittleInch Thanks, for the water pumps I am dealing with, the MCsF is as high as 60% of the pump rated capacity which is considered relatively higher than the usual case due to the large pumps selected. Also, while reading one of the pumps books, I have come across that for subcooled fluids specially the pumps dealing with water, the cushioning effect of the bubbles which may form during recycle is absent (specially at temperatures of 49 C and at higher suction energy levels). Hence, do you foresee concern in returning the recycle line directly to the pump suction in this case?

Additionally, EdStainless , you are recommending 50 D straight length, however, will this be downstream the pump suction nozzle reducer?? since genuinely straight length shall not include it.

I would also like to ask what is the typical duration though which the pump can be allowed to run in recycle mode. Should recycle operation be limited with time or it it would not matter as long as we are employing a good design of minimum flow recycle control?

Additionally, for very high pressure drop across in order of 150-200 bars the MCSF would a suitable cage/trim design suffice in mitigating the vibration and thrust experienced during operation or it would be better to consider as part of design a combination of FCV and RO?

Regards,