Go here: Centrifugal Pumps

You need to read a complete explaination of pumps to understand them.

Doing it here in less than a thousand words won't help you very much.

Well, if there is maximum flow with no head developed, what is the use of installing the pump <rhetorical question - NNTR>?

Here is a simplified set of pump curves, and you can clearly see the run out point, dead head, etc. I hope you find this almost self explanitory.

There will never be a point at which there is No System Resistance which is what Head overcomes. Even if you disconnect the discharge piping right at the pump flange there will still be the resistance caused by the pump casing and the discharge flange area. Besides that is the physical limitation caused by the flow area of the impeller and the resistance of the walls and the vanes. Another limiting factor would be the suction eye area, the entrance angle of the vanes, the NPSHR vs. that Available at a flow well above the Design Flow at the Best Efficiency Point (BEP) and finally unless the driver is extremely over sized the overloads would either kick it off or the turbine's or engine's HP/KW rating would be exceeded and would slow down. Increased turbulence at extreme flow would most likely also cause cavitation which would also prevent such an increase in flow so that the head curve would in practical terms never reach zero at the "flow rate axis" as you put it

Actually some interesting questions here, and not as much clear info available as you would think, simply because pump specifiers do all they can to avoid selecting a pump that will operate much beyond BEP (Best Efficiency Point), let alone right up to EOC (End of Curve, or run out point) so it is a subject that rarely comes up, and people don't dwell on it. The main reasons for this are that the absorbed power curve and, in particular, the NPSHr curve start to rise rapidly, so you can have problems with motors overloading, cavitation, vibration,etc..

Why the head vs flow rate won't cut flow rate axis ie. Max flow with zero head developed? Is this because of pressure drop in outlet port of pump? - You have really answered your own question here. You will never see a zero head situation as there will always be friction losses within the pump itself (not just at the outlet port, but right through the pump).

Which decides this point? - (ie EOC or Runout point) - Good question. Literature and web-info is very poor on this, and I have never seen a good definition of this. I have my thoughts, but I wonder what others think. Exactly how do you define or determine EOC point?

You're right that pump specifiers try to select a pump that will operate near BEP, but not all centrifs have power curve rising continously with flow. Plenty of them have a hump on the flow/power curve. NPSHr is likely to increase at higher flow, but in principle, the suction pressure can be made to exceed it.

Also friction losses within the pump don't come into it. The manufacturer's curve shows generated head across the pump, ie allowing for internal losses. It's the losses that mean the efficiency is less than 100%.

Also in principle, the pump could run at zero head. This could happen if the static head were negative. Of course you wouldn't normally bother installing a pump in those circs, but if both suction and delivery levels are variable it might have to be taken into account, I've known cases like that. eg could be drawing from a tank say 10m deep, and somebody opens a valve to atmosphere on the discharge side. Might handle it by ensuring velocities are OK, the motor has enough power (if power curve rises to that point) or stopping the pump on vessel levels, or relying on motor overload trip, but it's worth keeping in mind.

Since we're talking about a basically incompressible fluid being pumped there is no question that the NPSHR will increase since the eye of the impeller is not going to enlarge to accommodate an increased flow. While the NPSHA can be increased by reducing temperature and vapor pressure and/or pressure in the suction vessel I would assume that the OP's premise is that we hold the original conditions on the suction side and only predicate the changes going on internally and on the discharge side. Internal friction losses are taken into account for the normal range of flows and not necessarily for the extreme condition of absolutely uncontrolled discharge with a limitless driver input.

Zero head can probably be accomplished by having a variable speed driver and running the pump at such a low speed that the TDH cannot overcome the internal friction of the eye, entry vane losses and internal impeller losses. This should register as zero output in pressure but of course the pump is developing a head, just not enough to get through the pump. Of course, at such a low flow the internal losses would be nil but then the ability to move the fluid through wouldn't be so great either.

You are right about power and NPSH. The point I wanted to get to (eventually) is that what we are told in our infancy is not necessarily the case. There is absolutely no reason why power considerations should prevent you from running a pump beyond BEP, even to EOC. If you fit a large enough motor, it will cope. As you probably know, some major petrochemical companies have always insisted on motors sized for EOC rather than duty point, specifically to ensure motor overload will never occur. Likewise, there is absolutely no reason why NPSH considerations should prevent you from doing likewise. As long as NPSHa exceeds NPSHr by a comfortable margin, the pump will cope. So, if these are not an issue, then what determines EOC point?

I have not worked with centrifs for many years, and have no one handy to ask, but very brief (I'm sometimes too impatient to be a real engineer!) internet research has thrown up some nonsensical definitions for EOC, such as - "Pump Runout is the maximum flow that can be developed by a centrifugal pump without damaging the pump". I have not yet seen what I would consider to be a real definition of EOC.

There is one other point worth delving into. I have heard (or read) it said that noise and vibration can increase towards EOC. I have to admit that I know absolutely nothing about this. Is it due simply to the excessively high flowrate through the pump? Is it something that could cause damage to the pump (or seal)? I have no experience with this, but I cannot conceive of it being half as bad as conditions at the other end of the curve (closed valve), where you will almost certainly see all sorts of problems, including, overheating, cavitation, seal issues and radial loads (assuming single volute) at their highest. I still cannot see that this would determine EOC point.

"Also friction losses within the pump don't come into it" - I have to disagree with you here. In fact, this could be the crux of the biscuit. There are certainly friction losses within the pump, and as with any hydraulic restriction, the losses increase dramatically as flowrate increases (not to mention viscosity, but let's not mention that …… oops!). If you were to pump a fluid through a pump at that pumps' maximum flowrate you would certainly see losses probably far exceeding those from an equivalently sized (vs. suction/discharge port size) 90deg bend. Even if you removed the impeller, you would still see very significant friction losses. Look at the path the fluid has to follow, at a very high flowrate. It couldn't be any other way.

I think Spinco has hit on exactly what I was thinking - "of course the pump is developing a head, just not enough to get through the pump" . So I would be inclined to define the EOC point (or Runout point) as the point at which TDH (Total Differential Head) equals total internal head loss.

Thanks a lot Mr.Holzfeller and all other members for your informative points and time. Sorry, one more basic doubt arised is, if head developed is less than losses in pump, then how discharge 'l come? But at EOC flow is max.? Could you explain in detail? Thank you.

I guess this goes back to your original question - "What'l happen if the system curve is away from run out point ie. no operating point?" (ie. system curve is to the right of EOC). The official response would be as follows:

A centrifugal pump will always perform in accordance with the suction/discharge conditions with which it is presented. It will perform at the point on its curve that intersects with the system curve. What you describe is a situation where the system curve is to the right of (beyond) the EOC point. It doesn't matter what occurs beyond EOC, what the pump "sees" is a totally open valve condition, ie. zero system resistance. It cannot go beyond the EOC point, so it will always run at the EOC point.

However, I have no experience or evidence of this at all, and therefore must remain open-minded and ignorant of the true facts. Interesting question - under these conditions (at or beyond EOC), what would happen if the discharge head were actually negative (eg pumping into a vessel under high vacuum)? What would be the flow?

Some good points there and I think we're pretty much in agreement. It's a while since I worked with petrochemical companies' specs, but I'd expect them to ask for motor rating based on maximum power drawn (usually plus a %age) in case max is not at EOC.

I'm not sure either about precise definition of EOC, but some manufacturers show a max flow on the Q/H curve which should not be exceeded (tho it's not EOC as the curve goes beyond that point).

On the subject of power vs flow, most mixed- and axial-flow pumps have a power curve increasing with falling flow, often with a sharp rise as zero flow approached, so rating the motor to cover that would be a big jump over duty power, so better to ensure the pump can't run dead-headed. In that case can't start against a closed valve, unlike most centrifs which prefer it as lower power drawn.

Back to friction loss in the pump - assuming typical efficiency 65% and taking TDH say 20m, losses are equivalent to 20*(1-0.65) = 7m. If typical pipe velocity at the pump = 2m/s, velocity head = 0.2m, so losses = 35 velocity heads. That's far more than that due to water flowing through the pump with stationary impeller, so I conclude most of the loss is due to turbulence etc caused by the impeller.

For your definition of EOC, I don't think it's the TDH that = total internal head loss, rather TDH + internal head loss (which we might call gross TDH) = internal head loss, but isn't that just another way of saying at EOC, TDH = 0? That's assuming EOC is where TDH = 0 on the curve, which seems to me a reasonable definition.

Sorry Codemaster, I've been having problems with Internet Explorer, and my response seems to have been totally lost! I'll get back to you over the weekend (unless it recovers).

pumps are designed for best performance at a certain head pressure and flow rates. the curve rates are available from the manufacturer.

a pump with no or little head pressure is normally called a pusher pump. larger vanes = more volume.