Run Out and Centrifugal Pumps

Are you asking what the consequences are of running a centrifugal pump far to the right of best efficiency point (BEP)?

When you say "end of curve" what is that as percentage of BEP?

What is the Nss (Suction Specific Speed) of the pump.

As a very general statement you would not select a pump for a given service more than about 20% to the left of BEP or 10% to the right if practicable. However, closer is better and it also depends on the Nss. The higher the Nss the less you can move the operating point either side of BEP. High Nss only about 105% and very low Nss up to about 120% of BEP flow.

Guest #1

No,in this discussion efficiency and power consumption is completely not being considered. I am just asking whether this system shall work or not ?In fact wanted to insert pump vs system curve here but couldn't.

Some use ful pump data are as follow :

Maximum flow rate= 160 m3/hr.

Minimum Flow rate=40 m3/hr

TMH at min allowable flow=134 mwc.

TMH at maximum allowable flow=88 mwc.

shut off head =135mwc.

specific speed at best efficiency point=778.23 US Unit

suction specific speed= 8940.79 us unit,

NPSH(R) = 7 mwc

pump rated speed= 2930 RPM.

This may help for further analysis and discussion.

Good day.

I mentioned neither efficiency nor power consumption except that BEP defined a point on a curve. I repeat the question - are you asking if the pump will work far to the right of BEP? If so yes, but not for long - if motor can cope. Keep trying to insert the curves it will make understanding your problem much easier.

Guest #1

Exactly, this is the my question i.e pump curve seems to intersect with system curve at far right of BEP as well as end of curve (by interpolating pump curve).

Also, this may be a hypothetical point of intersection because curve given by manufacturer end much earlier and that's why this question arises whether this selected system will work or not?

In which case your question has been answered - Yes, probably, but not for long, as long as the motor can cope. If not, close in on the discharge valve as suggested by Ed.

A pump will find a point of balance even if it not on the graph. But this is clearly not the ideal. A quick fix would be to close the control valve slightly to add the missing head and prevent over pumping and excessive power.

The duty point seem to be at the right of the best efficiency. A bigger pump in the range with skimmed impeller or slower RPM may have been a better choice. (but it is hard to say from a distance)

You should investigate the reason for this choice and deal with it accordingly. For example if it was because of bean counters on your side you may have to fork out some money.

OK,what need to be done is clear, I am asking whether this system will work or not. I am taking this opportunity to elaborate further as :

Case-1: A pump is taking suction from EL. 1107.0 and feeling tank at El.1210.0 (including water ht. in tank),this gives static difference of 103 mwc plus frictional & other losses gives 110 mwc required head of a pump.(normal condition)

Case-2:During flood suction pond water El. goes upto El.1148.0 to fill water at the same tank at same elevation,and reducing static difference from 103 to 62 mwc temprorily. In fact nothing changed physically and system has been already designed according to case-1.And due to flood water level in suction pond goes up and creating rare condition case-2.

While analyzing on Fathom same pump in second condition,system curve & pump curve doesn't intersect. Assuming electric motor is over sized to cater 2nd case power requirement, what will happen ? whether the same shall able to deliver water in the tank or not.

The shape of the pump curve will determine what actually happens.

Some pumps may reach a maximum (curve running almost vertically).

Most pumps sloping down wards gradually will deliver more water and generate more friction in the pipe until a balance point is reached. The bigger volume at a slightly reduced head and lower efficency will consume more power which may exceed the drivers specks.

A control or pressure regulating valve on the delivery side will dispose of the additional pressure and your output will stay the same. You will however be wasting energy.

If this has to be done for prolonged periods you may consider having a second unit to cater for the reduced duty point.

<...system curve doesn't intersect with selected pump curve...>

Eh? This doesn't make sense. The pump curve and the system curve always intersect, for that is what dictates the operating point.

Completely agree with PWS.

You may not be working at the max efficiency point but they are bound to intersect, simply because one is a positive slope and the other a negative slope.

You are of course correct, but what I think he means is that when he plots the system curve against the pump curve the system curve is to the right of where his pump curve ends. If this is the case then the pump will pump as long as the motor can cope with the flow. The pump will not last long for reasons given in previous post.

All you guys need to understand what a pump curve is. A typical pump curve is not a complete listing of performance at all possible flow rates (at a particular speed and liquid condition)

It is the level of performance the pump manufacturer is prepared to sell or what the manufacturer actually tested on his own calibrated test stand prior to shipment (given that few actual installations have calibrated flow measuring equipment).

The high flow end of the pump curve typically is at a region of rapidly falling efficiency and cavitation from inadequate NPSH. Typically on a manufacturer's commercial test stand there is not a whole lot of static head available so the power requirements start to level off at 125% to 150% of the maximum efficiency point with the onset of cavitation. Beyond that they often cannot get consistent test readings and don't want to sell a pump in that regime anyway because mechanical issues like excess bearing loads, shaft deflection, cavitation damage and vibration start to look like warranty problems.

So let's say you have a pump rated at 160m^3/hr at 88mwc and you try to run it at open discharge through a short run of pipe, for example. (note the pipe should end at least a few centimeters above the centerline of the pump so there won't be air in the casing when you start up). If it's an electric motor driving it with a decent starting control it'll quickly get noisy and may kick out the overload circuit. If you are driving it from some other source with a lot of power available at a fixed speed around 3000 rpm it'll likely become noisy from cavitation and may surge some especially if it's an internal combustion engine with a speed governor. Since this is a relatively low specific speed pump with a mechanical design to handle low flows at high pressures without early failure for 60 Hz applications as well as 50hz it seems unlikely to me that you'll hurt it for a short running period.

Note also your flow suggests 80mm piping or thereabouts. Certainly you can afford a small gate valve installed in the discharge line along with an inexpensive pressure gauge between the pump and the valve. Fill your pump and piping with water and start it up with the valve closed "running at shutoff". If it sounds OK then gradually open the valve watching the pressure gauge and an ammeter on the motor connections and note when either it starts to sound troubling and/or the current reading gets too high for satisfactory operation (whatever your electrician thinks the limits are). This is pretty much standard operating procedure for any small centrifugal pump test. Don't try this with a positive displacement pump unless it is equipped with a suitable sized relief valve.

One other thing is important here. If you haven't purchased the motor yet and are planning to buy new you may find that the pump supplier really does not want to guarantee more than one condition point for head, capacity and power. If you plan to purchase the pump with an electric motor attached it is reasonable to ask the supplier for a motor sized to fit higher power requirements if your operating conditions require. Since most commercial centrifugal pumps in the world marketplace are designed for both 50hz and 60hz applications it is likely a motor of larger frame size (and power output) will be readily available.

Ed Weldon

All you guys need to understand what a pump curve is.

Ed, With all due respect I think that a number of "us guys" do understand what a pump curve is.

In spite of the "off topic" I stand by my position based on the way you all carried on the discussion before I arrived. Everybody missed the point that "pump curve" has more than one meaning.

Prof or Henrik: I know you guys are well versed in knowledge of pump selection and system design and have seen a lot of pump performance surves. And that Guest had a reasonable question that might come from an engineer without a lot of experience in these issues. But the issue I raised should have been settled right from the gitgo so that once Guest told us about the "flood" condition the specific issue of actual performance could be explored in detail.

Fathom, for all it's capability, appears to me on the basis of the comments that it takes the pump curve, including the end point, quite literally. Out in the real world the engineer sitting at his desk looking at a computer screen may not fully appreciate that a pump does not just stop pumping at the end of the rating or test curve.(with further decrease in system head).

So to me the real question is can Guest come up with a single pump/motor selection that will handle both condition points? Or will he need external controls such as an operator to man a throttling valve, a variable speed motor drive or motorized valve with appropriate control system or a second pump to handle the excess flow and reduced head of the "flood" condition?

I would look in the pump manufacturer's family of curves at the next larger size of pump with the idea that selecting an operating for the normal conditions of service to the left of BEP might give near the same efficiency as the smaller pump previously selected and be better able to handle the flow resulting from the head conditions during the "Flood". It's all going to be an operating cost vs. equipment up front investment trade off. Ideally there would be no emergency plans to execute or complicated control systems and sensors that would have to be periodically tested and might still fail during the flood.

Another area of exploration would be a high slip motor if available from the motor supplier as a special. This approach was used on simple variable speed pump drives when first introduced some 40 years ago before variable frequency drive technology matured. Such a motor slows down from synchronous speed to full load by a much greater percent. It can still be reasonably efficient at partial loads and inefficiency at the emergency condition is not likely to b an excessive financial drain.

Ed Weldon

Guest #1

Very nice & elaborated reply. It is really a nice forum to interect with such wonderful experienced people and a great learning experience.

With reference to my earlier reply post#3 in reply to #1 where details of pump duty conditions has been elaborated.I am taking this opportunity to interpret the pump curve (which is generaly interpreted by less experienced engineer) in some other way :In any envisaged condition,static head shouldn't be less than TMH corresponding to end of curve,in the case -2, where due to flood condition static head changes from 103mwc to 62 mwc and this is well below 88 mwc at the end of the curve,so system may not work. This is run out condition and hence caviation will occure ?

This run out difinition and its consequences are may seems to be theoretical but generally how run out condition is typically explained not to mention here pumps & its componet life, noise etc.

Again some thing contradictory with reference to cavitation, because either conditions are flooded, only level is changing from EL.1107.0 to 1148.0 while pump centre line is at approx 1098.0 so cavitation is not envisaged.

Please correct me if wrong.

Have a nice day.

"Again some thing contradictory with reference to cavitation, because either conditions are flooded, only level is changing from EL.1107.0 to 1148.0 while pump centre line is at approx 1098.0 so cavitation is not envisaged."

Do not limit your thinking to a pure suction condition induced cavitation (NPSHa being too close to NPSHr).

When you deviate from the Best Efficiency Point there will be an ever increasing discrepancy between the fluid approach angle and the impeller inlet vane angle. This will result in recirculation and vortices forming in the impeller passages and tips and leads to cavitation which can be extremely damaging to the pump. (Similar thing happens at the discharge). This goes some way to explaining the shape of the efficiency curve. You do not mention an NPSH curve so I guess you do not have one, but at flows way to the right of BEP the NPSH required rises very rapidly.

Guest # 1

Well, my understanding level regarding cetrifugal pump in general & cavitation in particular broaden.

Good day.

Ed

Gate valve is a poor choice in my opinion, they are mainly used as shutoff valves and can be very difficult to open against pressure and do not control flow well. A globe valve would be more appropriate I would think.