10 replies to this topic

[siretb]

I have a problem with a pair of centriguge pumps.
I have two pumps in parallel, one iddle (with valve closed) and one active, that pump water at 63°C, at a flow rate of 700-800 m3/hr (3000-3500 gpm). The inlet line (one pipe, that manifolds to the two pumps) is large enough, but have two sharp bends. I dislike the inlet design, but cannot be sure it is the cause of the problem.
The liquid level is about 5 meter above the impeller centerline, under atmospheric pressure . and the the pump is supposed to discharge at 6-7 bars. We have that pressure, more or less.
The pump is running at 1450 rpm and the impeller (plastic) diameter is 500mm.
According to the supplier, the NPSHR is 6 meters.

I can hear a noise, that , to me, is characteristic of cavitation. This noise is really frightening, although we managed to run for monthes. I cannot measure the actual flowrate. The liquid is pumped from the bottom of a scrubber, and we have some water returning to the sump, from a much higher elevation, by a submerged pipe. It could be that some air is returned by this line and sucked by the pumps, though the velocity in this line is around 1 m/s. We have tried to change the design, we have put some 'baffle' to help for air disengagement between this return line and the scrubber bottom, to no avail. The residence time in the sump looks OK.

I have two questions: despite the high liquid level, could we have some vortexes that would extend from the surface, all the way down to the pump succion line, at the bottom?
Do you think it is a good idea to reduce the pump velocity (the rpm) and run the two pumps in parallel (twe two pumps beeing active), instead of one. I can accept that either (or both) of the flowrate and discharge pressure is changed a bit.

Thank you in advance. Bernard.

[palusa]

Bernard,
is the pump working at its design capacity or not?

If it's running at reduced flow, I've posted in another thread an article regarding risk of damages because of internal recirculation at lo-flow... What's your pump Suction Specific Speed? it is above 9000?

If instead the pump is at its operating point, an additional suggestion that solved me a similar situation is to install a vortex breaker on your suction drum.

I would avoid however to run the pumps in parallel ...

Hope this could be helpful to you.

_Lf_

[rosa]

as palusa said,if your pump does not work on designed capacity ,it is beter to reduce the rpm.in this case your pump work slighter and the noise also reduced.also the inserted load decreases.but reducing the rpm should done in according the performance curve of your pump.also pay attention your liquid discharge temparature.

Rosa,
cheers!

[hamadzzz]

Bernard,
is the pump working at its design capacity or not?

If it's running at reduced flow, I've posted in another thread an article regarding risk of damages because of internal recirculation at lo-flow... What's your pump Suction Specific Speed? it is above 9000?

If instead the pump is at its operating point, an additional suggestion that solved me a similar situation is to install a vortex breaker on your suction drum.

I would avoid however to run the pumps in parallel ...

Hope this could be helpful to you.

_Lf_

if you reduce RPM of pump the tepm will increase then the motor will damage

hamad

[uetian98]

Bernard,

You can have few checks before that you pinpoint the cause.
It seems that your NPSHA is less than NPSHR. Perhaps its the line losses that are making the water to vaporise while traveling down to the pipe eye. Try evaluating the suction losses. Moreover try flooding your scruber (if you can afford) and note if you can still hear wierd noises from pump casing.

Try working out with Palusa's Idea ...

[siretb]

Rosa & Palusa. Thank you. I read the previous posts before posting my own. As told, I am not sure of the exact flow. However it cannot be much different from the design point (700-800m3/h), I would know because the rest of the process would not work fine. So the pump is not running at a much reduced flow. There is a plain (a cross) vortex breaker.

[palusa]

Well, Bernard, if the pump is operating as designed for, my further checks are very much obvious ... being your pump pretty large (24 inches suction I guess to stay below 1m/s ?) do you have T or basket strainer there on suction? if so check their mesh

From simple math I made you should have a NPSHavail: 10 atm + 6elev - 3 vap_press= 13 mwc considering almost nothing for suction prdrops on 60 equiv_metres.
You stated Vendor asks for 6 m so the margin looks large...

However double-check your pipe branches, reductions & bends ... you mentioned 'sharp' elbows: how? I took L/D = 20 but could be much higher ... what kind of vavles do you have there?

A last (very silly) question ... did you try to change over the main and the spare pump? did you notice a different behaviour (no, I guess)?

Anyhow: don't change pump rpm...unless you want to impact on Flow or pump head and this is tolerable to your process.

Good luck

_Lf_

[djack77494]

Bernard,
I would examine the following as possible causes of your problems:
1) Insufficient NPSHa. Calculate this based on the flow you expect through your pump.
(Get flow using measured discharge pressure + pump curve.)
Now compare: NPSHa > NPSHr = Should be no problem. Also, consider unexpected pressure losses due to (for example) plugged suction strainer or a partially closed valve. I note that pumps like well developed flow at their suction nozzles, so I agree with your aversion to the two elbows upstream of the pump. However, this alone should not create the problem you describe.
2) Vapor entrainment. Increasing your liquid level should reduce vortexing. Try it and note if the situation imporves.
3) Low flow. Try to increase the flowrate through the pump to note changes.
4) Mechanical problem. Switch to spare pump and not changes.
Put on your Sherlock Holmes' hat.
Doug

[siretb]

Thank you all. I appreciate a lot.
I saw no one replying to the last part of my querry. Is it likely that, with 5 meters of water, I have vortexes extending all the way down?

[palusa]

With a vortex breaker installed I'd say suction vortexes are unlikely to occurr and extend ...
As above suggested did you try to vary the liquid level and check is something happens?
About vortex breaker: did you inspect it recently? it could be clogging suction pipe ...
If you suspect air entrainment, have you checked for air pockets on the suction (cold pipe) ...

To disengage air on suction, a balancing line between suction and suction vessel is in general suggested, but I undestrand this is a big modification ...

_Lf_

[uetian98]

Bernard,

Your last queries were

"I have two questions: despite the high liquid level, could we have some vortexes that would extend from the surface, all the way down to the pump succion line, at the bottom? .... I havent heard this at least. Moreover, Vortex Breakers are highly recommended for such services.

"Do you think it is a good idea to reduce the pump velocity (the rpm) and run the two pumps in parallel (twe two pumps beeing active), instead of one. I can accept that either (or both) of the flowrate and discharge pressure is changed a bit." Running the two pump in parallel will intensify the situation. More flow ~ More pressure Losee ~ More noises - you may damage you equipment permantly. If you say that you look forward to reduce their rpm (then its some major change). reducing the rpm will give u lesser capacity but as that we will not be following the same pump curve so you will get less outlet pressure. But again, we have never taken this step EVER.

Try,
Increase Suction Head (Increase Suction Level).
Locate Unwanted Suction Losses (Choked Suction line, Choked Strainer, Sharp Edges)
Run the spare pump on trial. If it gives the same set of noises then its not the pump its is the suction line the real culprit. if the pump doesn't offer noises, then it could be a chocked strainer of the other pump or may be some broken impeller scenario there with the running puymp)

hope that helps.
Cheers
AHMED