Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations IDS on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

How to make recirculation line from discharge to suction line of centrifugal pump? 1

Status
Not open for further replies.

Zabarieq

Mechanical
Oct 3, 2016
8
Dear Seniors,

I am new to pump installation work and this time I face situation that it seems I have to put recirculation line from discharge to suction line of the pump.

I will use centrifugal volute pump, with flow of 265 LPM at 7.7 Bar (BEP), 2.5" suction and 2" discharge port. The media is water at ambient temperature. This pump will serves 10 point with identical flow (26.5 lpm each). The operation option may go to 10 point at the same time at max, and 1 or 2 point at the same time at minimum. This situation will make the operating flow to be 10 or 20% of the BEP. Based on the internet source I read, it is said that if the operating flow is below 50% of the BEP, then we need to put recirculation line from discharge line to suction line, but I don't know how to do it.

Some source said that I will need suction tank to receive excessive flow, but the other source said that I can directly put the excessive flow back to the pipeline before the suction of the pump. I prefer to choose the second source since I don't have space to put suction tank.

So please kindly advise what should I do?
Can I make recirculation line from discharge to suction line?
What component should I put on the recirculation line?

Thanks,
Zab



 
Replies continue below

Recommended for you

First get the min required flow value from the pump vendor. 50% of BEP is not a bad first number to use, but isn't fixed and firm.

So long as you have some forward flow (i.e. at least one of your points is working) then you don't have to worry too much about the water heating up.

Yes you can flow back to the inlet via some sort of pressure restriction.

The best is to use an automatic recirculation valve or ARV. This will open when flow is low and may need an orifice to create the back pressure, but will close when there is enough flow.

A very simple re-circulation will use a constant flow orifice, but then your pump needs to be oversized for full flow + re-circulation flow rate, which IMO just wastes money.

Lots and lots of information on this site as this comes up very often.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.
 
Generally a circulation line would just be a line running from the discharge to the suction. I am not sure what you mean by 10, 1 or 2 point. I would add these suggestions to the general solution:

Have a pressure gauge at the pump outlet before the recirculation branch. Add a throttling valve in the recirc line that allows you to throttle the recirc flow until the pressure gauge shows pressure at the best efficiency point. If the system varies during operation, you will want to look into automating the valve from the discharge pressure gauge as LittleInch suggests.

If the recirc line is installed at any point above the pump suction, have a bleed valve to atmosphere or leak tank to let you purge air at commissioning. (Make sure you comply with regulatory requirements before releasing anything into the atmosphere. Perform a job safety analysis before any such work).

The tank comes into play if it is a closed system. It allows expansion and contraction of the fluid and maintains a head on the system and you can run the recirc purge line back to the expansion tank. The tank, if used, it should be installed above the system an sized based on the volume of fluid in the closed system.
 
@LittleInch

Thanks. Your explanation is very helpful.
I just knew about ARV, initially I thought that Safety Relief Valve will do the job even I have doubt that the spring will not last for long time. I will try to find ARV sellers in my country.

Regarding to put the excess flow to the suction line, do I need to put check valve there? I am afraid that the water from existing tap-in point will flow to this recirculation line and disturbing something. Please find my PFD sketch on the attachement,

PS: Regarding to find out minimum flow from my vendor, unfortunately my vendor is very slow responsing my query about this. But I will keep trying.

Thanks.


@SandCounter

Thank you. The attached file I mention above may explain about that 10 points. There are 10 points of spraying nozzle operated by solenoid valves (SV). The pump will be interconnected with these SV's, the logic is the pump will start if at least one of SV is energized. Based on this operating condition, then YES, as LittleInch suggestion, seems I need an equipment that automatically release the excessive flow whenever the pressure raised to certain kind level.

And Yes, you are right, the re-circ line will be above the suction line, so I will follow your suggestion by putting bleed valve to release the trapped air on the system. And thank you for reminding me about JSA.

Cheers,
Zab





 
 http://files.engineering.com/getfile.aspx?folder=5aa21aee-17b5-4cf5-b630-1f4fc7a578d6&file=engtips1.pdf
Zabarieq

Normally for centrifugal pumps, the pressure relief cannot be used as a minimum flow as the pump curve is too flat at low flow and the valves are not accurate enough, but it might be feasible depending on your pump curve. These types of valves though do tend to wear and then don't seat fully when needed after a while.

Yes a check valve is a good idea. Normally you have a non return on the pump discharge and not the pump inlet.

If your control system is available for programming, then an alternative might be to set up a re-circulation line with an actuated valve and orifice sized for your min flow ( say 30% of rated flow).

If less than 3 of your SVs are energised, then your control system opens the min flow bypass.

If 3 or more SVs are open, then the min flow bypass closes



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

Thanks for your suggestion. First I will try to find ARV first, if I have difficulties to get it then I will go to flow control programming thing.

About orifice sizing, do you have any link where I can learn how to do it?

Regards,
Zab
 
Your best bet is to search on this site under "restriction orifice"

Or look at something like engineering toolbox. You're almost certainly in choked flow.

You need to use that term and not "orifice plate".

However the restriction orifice route is a bit brutal and an ARV or even more sophisticated control valve route is more gentle, but costs a bit more.

With clean water and a relatively small pressure drop you're probably OK, but bear in mind that RO's when used for long periods can suffer erosion and gradually become bigger than they were originally....

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

Got it. I will try to learn about it.
Thanks!
 
For small pumps, we use a globe valve in the recirc line. For smart customers, we leave the handle on... For less smart customers, we dead head the pump and set the bypass to satisfy the pump's minimum flow requirement, then remove the handle. The minimum flow for small pumps (3" suction and smaller is what I'd call a small centrifugal) is a fairly small fraction of the flow at BEP, but they tend to be much happier with an external flow than they are being dead-headed.

If operation at zero through flow for more than very brief periods is probable, we run the recirc line back to the tank rather than immediately back to the suction. However the OP's situation sounds like running right back to the suction might be OK because there will always be a few downstream users drawing flow.

3" suction is the minimum size I'd start thinking about using an automated recirculation valve if just thinking about the health of the pump.

In other cases, a backpressure regulator is used to dump flow back to the tank whenever the pump is being used as a pressure source rather than to satisfy a particular flow.
 
Moltenmetal makes a good point - I'd forgotten the size of this unit and was thinking of much bigger ones.

You're only looking at about 4 or 5kW pump here so the added cost and complexity of anything more than the simple units may not be justified. So maybe not an ARV, but a simple SV and a bit of programming of a simple PLC is probably worth it if the pump is running a lot.

Depends how often it is running and how much the options cost.

Also if the re-circ is left open all the time you need a bigger pump so factor that in as well.

Design the options, cost them including OPEX costs and make a decision.

Just come back and tell us what it was please.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.
 
If you dump the recirc right back to the suction, you need to be aware of the temp rise through the pump. Take the pump efficiency at the min flow, assume all the power lost goes into heat, and you can get a conservative estimate of how much the pump will heat up the fluid. For larger pumps, this is why it isn't a good idea to put it right back to the suction, fluid will heat up too much. Smaller pumps it should be fine, but do a quick back of the napkin estimate of the temp rise.
 
@moltenmetal

Thank you. As per your explanation, it seems back-pressure regulator is the best solution since you are right, this pump is being use as pressure source rather than the flow. The suction of this pump will receive direct tap in point from existing 2.5" pipe, this pump will have role as booster pump for dust suppression system on the conveyor. The maximum head of this pump is 90 meter, but we will only use it around 77 meter with total operating flow would be only 0.3 m3/min (BEP around 0.57 m3/min, I get it from its performance curve and I attach it here). At a minimum point of application, the operating flow will be only 0.06 m3/min which is about 11% of the BEP. Please kindly let me know your opinion. And also if I am going to use recirculation line, what is the minimum pipe diameter that I have to put?

@LittleInch

The pump will be driven by 15kW motor at 3000 RPM. I found the technical data sheet just now, and please find it on the attachment. What do you think if I use back-pressure regulator as per Moltenmetal suggestion? But still I will have a discussion with my electrical engineer whether the local control station can make arrangement or not to activate SV at recirc line in case I can not find back pressure regulator in time.

Regards,
Zab
 
@1gibson

Thank you. Unfortunately I don't know how to calculate the heat that will come up.
If performance curve said that the efficiency drop to 25% then can we assume that 75% of 15kW will goes to heat?
I think there is a lot of homework for me to get along with this project... >.<

Regards,
Zab
 
Zabarieq,

You could/should get a smaller pump - you're operating too far to the left of BEP. Looking at that curve my guess is you are looking at about 0.15 to 0.2 m3/min minimum flow. For a max flow of 0.3m3/min, that's a lot of energy you're just throwing away.

Like I said work out how many hours per year you're going to flow and roughly how many hours you think you're going to use a min flow. Then you can work out how many kW you might throw away and put that into $$$.

using a pump that is, quite frankly, far too big is very inefficient and will cost you $$$ in electricity compared to buying the right size unit.

I kind of get what MM says, but this isn't a constant pressure thing - again you can do it that way, but I just hate throwing energy away....

I wouldn't worry too much about the heat thing - you only need to worry about that if you've got no forward flow. Your control system says you have forward flow at all times, even if quite a low flow, but you might end up with a bit of an increase in temperature. You cannot pump more power into the fluid than the shaft power, so in your case it would be about 6kW. The 15kW is the max power input into the motor. If the pump doesn't need it it won't take it. So input heat of 4-5kW into a forward flow of 0.6m3/m = how much temperature increase??




Remember - More details = better answers
Also: If you get a response it's polite to respond to it.
 
LittleInch: I hate throwing energy away too, but when flow is highly variable and head is fixed by the users, your practical choices are few.

You don't set the backpressure regulator to spill back to the tank at a flow equal to the BEP- you set it only to satisfy the minimum flow requirement plus a wee bit to ensure the pump is operating at a stable operating point and isn't surging. And yes, when only a few users are on line, you're operating the pump way to the left of the curve. With all users on line, you should be running near BEP and the backpressure regulator should be fully closed.

In small pumps, you're safe to operate well to the left of BEP essentially forever, as long as you a) meet minimum flow, b) don't let the thing surge and c) get rid of the heat properly. When pumps get larger, forces get larger in a big hurry and from a mechanical perspective you're much more worried about operating anywhere outside that narrow sweet-spot around the BEP for extended periods.

Calculating the heat input is dead easy: go to the pump curve and find the motor power at dead head. All that motor energy will be converted to heat in your pumped liquid. That heat has to go somewhere.
 
MM,

Could you sketch out where this backpressure controller is on the pump system and where you would set it on the pump curve.

I'm not critising here, just wondering how it would work in practice. Normally I only set back pressure controllers at the far end of systems to maintain pressure in the pipeline to avoid slack flow, and for some reason can't understand what you mean.

I'm sure a nice little sketch would put this to bed.

thanks

LI

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.
 
No problem, but no sketch needed I think. You shut off the downstream users and then manually set the back pressure regulator, adjusting it until you get the minimum flow required by the pump coming out the discharge of the BPR plus a little for comfort. On a curve, that set pressure will be just below the deadhead pressure, i.e. when the downstream users are demanding no flow, the pump is operating in the left hand portion of the curve near the dead-head pressure. As soon as the downstream users start demanding, the BPR shuts off and the users ride the pump curve. That assumes of course that you don't have big fluctuations in suction pressure- if so, you'd either need the ARV (which contains a differential pressure regulator arrangement of sorts), or you might be better served with the throttling valve which we normally provide, or the RO that some people use instead. I don't like ROs so much as they tend to change in size of their own accord (via plugging or corrosion), and can't be adjusted easily when they do.

If the pump is unhappy and surges or makes noise operating that far to the left of BEP, then you set the BPR's set pressure a little lower, i.e. spilling a little more than the minimum flow required by the pump back to suction. That wastes more energy of course so you only do it if required.

 
MM,

Thanks and now can see how you are thinking and I attach a sketch - I'm a sketch person. so you start at head H1 and then adjust to H2.

My issue with this type of control and with the pump listed in the OPs post above, is that the curve is quite flat at the low flow end and hence getting an accurate enough setting is quite difficult, especially if the inlet pressure can vary even a small amount - I note you say big fluctuations- but I think it would only take a few percent (<5) of the developed head to make a difference). This means you would either get more flow than you need or less flow. Hence it will work well in the right location with steady pressures (such as pumping from a tank which has a fairly constant level), but in other places might not work as well, or at all.

If you're actually mainly working at the BEP or duty point is much lower down the curve and inlet head is fairly constant, then it could work well, but when you're using a pump which is essentially too big for the duty - as presented here, you really need to use something that measures or restricts flow which is not dependent on pressure.

So as ever, it depends on the requirements and details of each particular system.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.
 
 http://files.engineering.com/getfile.aspx?folder=5faac2d9-b72d-4b1b-b1a9-0eb85d68209e&file=20161007102043.pdf
@LittleInch

As moltenmetal said, I have limited option when user state pump specification. This is my first time to use centrifugal volute pump for pressure boosting instead of using vertical/horizontal multistage one. As per my calculation, with same discharge head it only need 7.5 kW of multistage pump with flow just around its BEP.
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor