Pump Performance Curve Clarification 6

[MajesticFerret]

So I'm looking at a Pump Performance curve.

RPM is stated as constant. Pump casing is a constant. You've got multiple impeller curves that descend in head as you get to higher flow rates.

My question is, how is flow rate subject to change if the impeller speed is constant and your only looking at the curve of the same impeller? If the curve represents a single impeller and it's spinning at a constant RPM, wouldn't the amount of fluid it's displacing be constant, ie, what it can yield as a flow rate?

I must be missing something here...

 

[JJPellin]

The different curves represent different impeller diameters. Do an internet search for "affinity laws".

Johnny Pellin

 

[chicopee]

I think that I understand your inquiry. Looking at only one curve for a particular impeller diameter at a particular RPM, the flow rate will change when the system demand curve changes. The system demand graph is your responsibility as you have to determine friction losses based on several flow rates for a particular piping arrangement which you can plot on the pump curve graph. Intersection of those two graphs will be the expected pump flow rate at a particular head for that piping arrangement.

 

[HPost]

Every impeller will have similar curves and its mirror of the conservation of energy laws. As the flow increases, the head will decrease.

Some are more flat than others, but they all have same curve for a given speed.

The flow will only be constant for a constant head at a constant speed

 

[BigInch]

Pump curves are based on the performance predictions ignoring any effect of all connecting elements, pipe, valves, etc., as if the pump were sitting on a test bench. In a real system, pipe, valves and other elements all have thier own curves. The performance when installed within a real system that contains pipe and valves will be the net result of combining all element's curves wtih the pump curve. The "system curve" (the sum of the effects of all elements other than pump), by itself, does not determine the flow rate of the system any more than could the pump curve, by itself alone. The flowrate and head within pump and pipes and valves together is determined by the intersection of the pump curve alone AND the system curve, not by any one curve alone.

you must get smarter than the software you're using.

 

[LittleInch]

I think you're asking a more basic question judging by the comment "wouldn't the amount of fluid it's displacing be constant". The answer to that is no it wouldn't, because centrifugal pumps work by imparting energy into the fluid as velocity. This is different to positive dispcement pumps where for a fixed rpl, they attempt to deliver a fixed volume of fluid. If a centrifugal pump has low or no flow, the power input just goes down, but the head / pressure rises a little as this is related to impellor diameter, RPM and impellor design.

Centrifugal pumps are essentially (~ 15%) a constant pressure unit within their normal flow range. As the flow increases the losses in the pump increase and hence the head will tend to fall, and power requirement increases until the head of the pump output matches the head required for a certain flow. You need to match these quite closely to get a good design.

As BI and others above say, to work out your flow, you need to know the flow / head curve for your downstream equipment.

you can think of this type of unit like an electrical circuit - your pump in this instance is your supply voltage - change the transformer settings (impellor diameter) and you have a different voltage. How much current you draw is dependant on your circuit. E.g. a supply rated for 100 light bulbs, One light bulb is very low current, but 200 lightbulbs will drop the voltage and possibly blow up your transformer / trip the circuit on overload.

You can take this analogy a little bit far, but it sometimes helps to understand the basic principles.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[MajesticFerret]

Alright, I think I get it. So centrifugals impart higher head onto lower flow rates and less head on higher flow rates. I believe I've seen instances where the pumps were set up to boost pressure without imparting a major change to the flow rate, so that makes sense.

So another question is, you'd determine the minimum flow rate you can run your pump at given a set RPM/impellor diameter by making sure it's head won't exceed the MAOP and you'd determine your maximum flow rate based on how low you'd allow your head to get, right?

So for booster pumps where the flow going into the suction side is far less than the flow going out of the discharge side, are flow control valves basically the only thing that ensures you get the flow rates/heads you want, given a set impeller size/RPM?

 

[1gibson]

Min flow is specified by the pump manufacturer, typically 30-40% of best efficiency point, but it depends on the pump specific speed.

Same flow in and out, conservation of mass.

The control valve is how you get the flow you want, the head is what it is. Note that there is a pressure drop across the control valve, so the pump creates a higher head than you would see measured after the valve. This is important when trying to locate where you're at on the pump curve.

 

[DubMac]

To better grasp the flow/head curve, it may help visualizing a pump running with only a fully open butterfly valve bolted onto the discharge flange. When the pump is turned on, there is no resistance to flow and the pump will be operating on the far right end of the curve. As the valve is slowly closed, friction resistance (head) will increase and cause the operating point to move back to the left on the curve until fully closed (shutoff head).

From a pump manufacturer's standpoint, below some flow rate, the fluid flowing through the pump is not able to move heat of mechanical friction out the pump faster than it builds up. That flow is specified as Minimum Flow. That figure may be adjusted higher in certain pumps due to other considerations such as discharge recirculation, radial shaft loads, or even process conditions, but never lower.

The pump will physically run below minimum flow, down to shutoff head, its just a matter of how long. The right side of the curve is different however. With the valve fully open, the flow will increase to the point at which it loses prime due to lack of NPSH. That is as far as I will offer on that can of worms.

The shape and slope of the flow/head curve is determined by impeller geometry.

 

[Artisi]

As usual, half a story from the OP. A copy of the pump "curve" and the duty requirements would make a great deal of difference to the comments being presented.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[LittleInch]

Majesticferret,

From your second comment I think these pages are not going to answer all your many questions. I would try looking up some basic diagrams and docuemtns such as

http://www.pumps.org/content_detail.aspx?id=5022

or a pump suppliers design guide.

You normally try and design a pumped system looking at the duty the pump needs to do matched to it's downstream system as a complete package.

Normally yes, you try and make sure that at low / no flow your pump won't exceed the MAOP of the system, but it's not the main criteria. Maximum flow should be within the pump and motors normal range, but rated flow should be close to the pumps best efficiency point (BEP). If you go beyond this you either start to get high vibrations or you exceed the motor power and it trips on high temperature or excess amps.

your last para defies the law of constant mass - mass in to a pump = mass out of a pump. Sometimes yes, yu need to insert a control valve downstream of a pump in order to match pump flow/head with that of a varying or lower system curve. To do this on a "normal" basis is just throwing away money as you've created the head, now you're lowering it and turing that energy into heat. However for startup and for a varying demand / head requirement downstream such a valve is useful to stop the pump pumping too much flow and going "off the end of the curve", i.e. going too far to the right hand side of the pump curve.

Hope this helps, but I think you need more than this.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[ProcessHVAC]

Wow.. you guys in the pump forum are really nice. A similar kind of extremely basic question would have been asked in other mech eng forums such as hvac and the OP would have been shredded to pieces likely being accused of being lazy at doing university homework

 

[Artisi]

Yes, we are nice- more than can be said for many of the OP's who can't even raise a thank you or just dump the post mid - stream never to be heard of again -- until they need assistance once again.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[LittleInch]

Artisi - couldn't agree more. sometimes I think we find the responses more interesting than the OP(!)

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[Artisi]

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[MajesticFerret]

Haha, thanks you guys. Just because the OP's forget to reply doesn't mean we aren't grateful for the information.

 

[Artisi]

Sure, maybe we should forget to supply answers and then the OP's wouldn't need to say thanks.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[TenPenny]

So for booster pumps where the flow going into the suction side is far less than the flow going out of the discharge side, are flow control valves basically the only thing that ensures you get the flow rates/heads you want, given a set impeller size/RPM?

I've never seen one of these pumps where the flow going in is far less than the flow going out. Does anyone have a picture or datasheet? I can think of several places they'd come in handy...

 

[Artisi]

A very useful pump whenever you are expecting a major leak via the mechanical seal, otherwise of not much use for day - to -day use.


It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)