Minimum flow rate of centrifugal pump

Some centrifugal pumps or pumping conditions may specify minimum flow rates.

This would be required for pumping hot fluids or pumps with cooled seals {stuffing boxes.)

The cooling of the stuffing boxes will depend on pressure differences which may be nonexistent at closed valve or low flows.

All manufacturers have detail on internet. You can try KSB pumps, Worthington Simpson or "centrifugal" and I think you should look at horizontal split casing or multi stage pumps

Note :

A manufacturer will not specify a minimum flow rate without a good reason therefore if it is specified it would be best to apply it.

In general running at closed valve or low flow will heat the water and the seal and it should not be maintained for longer than necessary.

Some pumps offer an option between stuffing boxes and mechanical seals.

Hi Technocrat

The term is "minimum continuous stable flow", and while there are some thermal reasons for this, the main reason is rotor dynamics. With low flow, and at that speed, then something is blocking the normal calculated duty point flow. Such as mis-sized for the head (pressure) duty, or a semi-closed valve, or some other back pressure causing the flow to be low (at that speed). We say 'to the left of the curve'. Then at that low flow point, inside the impeller there is tremendous recirculation around the impeller shrouds, wear rings, inside and around the vane exit shroud, volutes, diffusers (etc) and vortexes build and collapse, gasses come out and go back in solution . . . the fluid heats up making the vapor pressure lower making this worse, bla bla bla . . . and the result is the shaft shakes like heck and the rotor vibrates and pump case vibrates. It shakes so bad it destroys itself, and if flow it totally blocked, as we say 'at shutoff', then life expectancy of a high energy pump is 1-2 minutes, not hours. Minimum flow is defined as the point of low flow where the pump exceeds the vibration limits set forth in API 610 (ISO 13709). And that point is not allowed to be inside the recommended operating window also stipulated by API of 70-120% of the BEP (best efficiency point). The desired operating point is stipulated as 80% to 110% of BEP, but, the min flow vibe thing is not to invade the 70-120% window. Now, when you go out to the right of the curve towards what we call 'runout', then there is no head pressure so the pump just free flows. This is where maximum power is demanded and the poor little rotor become very unstable with too much power being pulled off of the designed shaft and bearings. So, to the left = rotor and case high vibes, heat, and those damages . . . . . and to the right = rotor instability and damaged bearings and broken shafts. The pump guy guesses when min flow is on the proposal curve based on experience with that pump. The only way to know is to test it and check vibes and draw a new line. It will be different than the proposal curve, normally lower than expected (the proposal curve is conservative to protect warranty).

Hope that explains this.

George

correction " . . making the vapor pressure higher..." not lower. Jeeeez I'm an idiot ! Forgive me.

My mother-in-law was right. Her daughter should have married the dentist and not the idiot pump dude.

Hi PetroPower. Is this similar?

I have seen applications where the pump was oversized or too many pumps were running in parallel for the desired duty.

In both type of cases the pump curve were as such that the closed valve pressure was less than the maximum pressure. (we actually had the pumps on a test bench to confirm the curve).

What happened is that as soon as the pressure rose above the closed valve pressure the pump may start to hunt or when the peak is reached the flow is reduced dramatically. With the system flow lower the pressure dropped and the correct duty was running again. Water hammer effect.

Solution suggested -

Single - replace with smaller pump.

Station - Reduce number of pumps running when flow approach the minimum stable flow point.

Hello Hendrik

Not sure I fully catch your story . . . but here goes !

In parallel, all pumps must have EXACTLY the same shape curve, especially the head rise to shutoff characteristics. If one curve is different due to different pump, speed, impeller diameter, worn out . . whatever, it will not pump (open the check valve and enter the manifold). And each curve must be steep, like 20% rise to shutoff from the BEP so you can get them on line fast (open the check valve). With running too many pumps, or too large, the system curve starts to move to the left because you are stuffing more fluid into a pipe than it can handle. So, you are right to use smaller or less pumps to match the system curve to the pump performance curve. The hunting was likely due to check valves opening and closing as the manifold pressure 'complained' and changed the system curve shape, and also some hunting can come from heating up the fluid due to moving that pump to the left and getting gasses in the fluid which makes for head capacity drop and flow drop . . . a vicious death spiral.

Not sure this is spot on, or even accurate, but it sounds pretty profound so I'm going to print it and show it to my mother-in-law and my wife ! Perhaps there will be romance in my future . . . let's see a dentist top this story !!!

George the idiot

Hi George

The sketches are the best I can do at present.

Left arm in plaster, right hand bandaged, face looks like a panda. (no! I am not a boxer)

I may have been a bit groggy while reading your post and could not understand it. My reply was not clear either.

I have to agree with your addition of the word "stable". . I could not find any reference recommending a minimum flow anywhere in the info at my disposal.

Forty years ago I was assisting a farmers periodical advising farmers about engineering problems.

All the problem cases with unstable pumping conditions investigated were because of pump curves as in sketch with the duty point to the left of the stable point.

I have never thought of obstructions as the cause of instability, it was rather seen as a cause of reduced flow because of the increased pressure head. An obstruction combined with a poor pump curve will however lead to instability. Will give your {corrected} explanation some serious thought.

Top : Exaggerated pump curve of bad pump. It really was unstable at low flows and high pressures.

Bottom : The pumps in a multi pump station do not have to be the same. I haven't met a pump capable of detecting the size and number of the other pumps in the system yet. A pump can only balance volume against pressure. A mixed station actually affords better control at low flow conditions. But it is agreed that the pumps need to be the same or at least matched for pressure when all the pumps have to run together.

For clarity, my first post was aimed against the rural myth that centrifugal pumps cannot work against closed valves.

Maybe we should just call them centrifugals, the pumping (of water) seems to have some sort of romantic connotation.

Hendrik the battered idiot

I cannot get the image to load properly.

will change it and retry later.

Centrif can work for 20 years against a closed valve if the energy density is very very very low. I ALWAYS evaluate theory in extremes to make a point. A 100 mm impeller turning 10 RPM with a 20 mm B gap can spin all day and never heat the water, never have recirculation caused vibes etc. But it would never pump anything against a high head requirement either. It would barely pump out of an open discharge. But a 10 MW boiler feedwater pump turning 7000 RPM 14 stage diffuser pump with a B gap 99% of max impeller cannot run more than 1-2 seconds against a closed valve. 2 seconds would cause serious damage. Uhuummm; you may be wondering how I know that !

Maybe the question "What is your application?" had to be asked first.

The statement "all centrifugal pumps" however include my normal water pumps and i responded to that.

The highest temperature I did was a cold 60deg C.

I had some costly bad experience. I once had a 6" pump cavitating within minutes on a 1 meter suction head. The 1st and 2nd replacement pump did the same. Only then did i check the suction hose to find that the inside liner was loose and obstructed the flow.

Long ago my boss taught me the KISS rule. Keep It Simple Stu--d.

Cavitation in a 6" would be hard to detect as it isn't big enough to hear (probably), and you may not have had a process benchmark (cavitation causes the flow to go down) and precise instrumentation (instruments would cost more than the pump). So the only way to find it would be at the failure.

George,

Thanks a lot for the explanation.

Technocrat