Centrifugal pumps - 2 questions 4

[mladenf]

Hi everyone from Croatia.

This is my 1st post so I hope that dumb questions are allowed for newbies

1. Centrifugal pumps are turn on with closed discharge valves - why?
I have 2 potential explanations:
- if it is a point of lowest engaged power, especialy having in mind that starting current is 2-3x higher than nominal, and there is less load on wires, instalation etc... If power = pressure x flow -> when the outlet valve is closed, there is a very small flow trough relief valve or a gap ring (ring between case and rotor) -> as a result power is low.
- starting with closed valve will produce high pressure for instant cooling and lubrication of mechanical or seal with packing
2. why simering seal cannot be find on centrifugal pumps (what is a downside comparing to machanical or seal with packing)?

THX

 

[itsmoked]

Centrifugal pumps are mass-flow devices. If no mass flows they are doing nothing and there is no load. This allows the motor to start, essentially, unloaded. Of course there is some mass to accelerate (the impeller) but that's not much.

Leaving a small leak around the valve will prevent seal failure from overheating with regard to a blocked off pump churning a little water. Even a small leak is usually more than adequate to keep a pump cool. This is the reason fire trucks get wet underneath them as fire fighters turn off all their hoses occasionally and the leak keeps the pump from toasting.

And just for your information motor starting currents are usually more like 8x running current not 2 or 3x.

And welcome to Eng Tips mladenf.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[Artisi]

Most pumps are not started against a closed valve, many start against open discharge while others start against a full discharge line. It is entirely dependent on the installation / application and a number of other considerations.

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.)

 

[georgeverghese]

Starting a pump against a closed discharge valve is not a good idea. The min flow recycle loop ( and any other safety devices ) should be active and not isolated during startup, at the least. If you have excessive high current draw, one solution would be to ask the electrical engineer to implement a soft start at the MCC for this pump.

 

[nondimensional]

Starting currents (locked rotor amps) of 2x to 3x times only lasts for a fraction of a second when compared to full load amps (FLA). For some units, this LRA can be 10x to 15x running currents (FLA). This is the case for across-the-line start. If you have a soft start (auto transformer, or solid state type), these can be shaved significantly. A variable frequency drive completely takes the LRA out. This is NOT the reason for starting pumps against a closed valve.

Starting low specific speed pumps against closed valve is almost always recommended. This is typically done to reduce the sudden shocks to the system - think transients. Power at zero head is not a big deal usually as it is much lower than running horsepower at rated conditions. If mechanical seals are specified, then the shaft deflection at the stuffing box should be carefully calculated not to exceed 2 mils (typical design criteria for mechanical seals).

For high specified speed pumps - typically 7000 and higher - the power at shutoff head (zero flow) starts to rise higher than rated power. At even higher speeds, pressure rise and power rise at shutoff is dangerously high - particularly so for axial flow pumps. For this reason, you will not see a closed valve on a very low head, high flow pumps - think storm water pumps here - New Orleans, Southern Florida, Houston, etc.

In water distribution systems, specific speeds are usually between 2000 and 4000 and starting these pumps against a closed valve which takes anywhere from two to eight minutes is VERY common.

 

[Valvecrazy]

I find that smaller centrifugal pumps, less than 100HP, usually have 6 to 9 times FLA for inrush, which can be reduced to 2-3 times FLA when started against an almost closed valve. Of course you wouldn't want to do this with axial flow pumps. A VFD maybe able to start slow enough to prevent all inrush amps with some centrifugal pumps. But with other centrifugal pumps like submersibles, a VFD still causes inrush currents because of the need to get the sub up to 50% speed in 1 second.

 

[electricpete]

usually have 6 to 9 times FLA for inrush, which can be reduced to 2-3 times FLA when started against an almost closed valve.

The ac portion of starting current (neglecting decaying dc offset) is roughly locked rotor current upon start (typically 6x FLA as you mention), regardless of valve position. Starting against a restricted/closed valve won't change that, but it will reduce the time that the current spends high before dropping lower (and your instrumentation might think that represents a lower value due to some averaging built in)

Of course you wouldn't want to do this with axial flow pumps

A worthwhile comment imo. For op, see faq237-1543





=====================================
(2B)+(2B)' ?

 

[Valvecrazy]

Reduced time is certainly important. But I also see a reduction in peak starting amps. Either way a straight centrifugal pump will start much easier against a closed or almost closed valve compared to wide open start.

That link is confusing. "Opening up on a system throttle", and "adjusting flow in the direction towards BEP". Sounds like double negatives to me. For a straight centrifugal pump I would have said "closing down on the throttle valve will decrease BHP and current draw", which is not the case with axial and mixed flow impellers.

 

[electricpete]

But I also see a reduction in peak starting amps.

The highest instantaneous current will occur within the first half cycle (8.3 millisconds in 60-hz land). Consider how little difference in speed there is going to be in the first 8.3 milliseconds, remembering that acceleration torque must accommodate both inertia and pump torque.... BUT the pump fluid torque is proportional to speed squared (neglecting the friction component which does not depend on lineup). If the pump is accelerating from 0 to 1% full speed during that first 8.3 milliseconds, then the fluid torque is changing from 0 to one ten thousandth of full speed torque. Any difference in speed resulting from a difference in that minuscule component of fluid torque during acceleration over that very short time period is neglibible. Speed is effectively the same in both cases over the first half cycle. There is no reduction is highest instantaneous current. Your instrumentation is not measuring the highest instantaneous current, but performing some kind of averaging. I will say the terminology for describing this type of measurement is tricky. The true peak instantaneous can be up to 2*sqrt(2) time the (rms) locked rotor current. The true peak is obvious if you have an waveform recording. Many instruments will try to give an RMS which is not an obvious conversion for a non-sinusoidal waveform.

Accordingly, the instantaneous relay settings can be determined for a given motor without considering the fluid lineup. The overload relay settings do consider the fluid lineup.

That link is confusing. "Opening up on a system throttle", and "adjusting flow in the direction towards BEP". Sounds like double negatives to me. For a straight centrifugal pump I would have said "closing down on the throttle valve will decrease BHP and current draw", which is not the case with axial and mixed flow impellers.

The link is my FAQ, so I'm always interested in comments that might improve it. But I think you're missing some context in your quotes.

"Opening up on a system throttle valve". The word throttle is used as an adjective indicating the normal function of the valve (we could almost drop the word throttle, but at least it clarifies we're not talking about a recirc valve). Opening is a tense of the relevant verb. It does not seem ambiguous to me.

In a completely different sentence, I said "for mixed flow centrifugal pump, the curve is non-monotonic. As a very rough thumbrule peak BHP is near BEP so current increases when adjusting flow in a direction toward BEP.". What I was saying is that it's not as simple to estimate which direction to change flow to increase BHP on a mixed flow pump as it was on a radial or axial flow pump, but (as a very rough thumbrule) the highest BHP tends to occur near BEP.

That particular FAQ has also been controversial for the statement that axial flow pumps and mixed flow pumps are "commonly referred to as "centrifugal pumps"". I am not saying the principle is centrifugal, only that many people (including some textbook and standards authors) lump radial, mixed and axial flow together under the term "centrifugal". Likewise I observe plant personnel often do not differentiate these types of pumps. But even though some people may refer to them all by the same name, there are differences that become important when you are trying to talk about expected bhp vs flow characteristic. The very fact that the common terminology does not match what an engineer might conclude considering the pump principles is a good reason to mention the terminology imo.

If anything, maybe I should enhance the standard caveats - consult manufacturer's curve for your specific pump rather than relying on these generalities. I also agree my FAQ is now confusing since the 3rd party link that I provided is now broken. Graphs would be nice, along with a label "typical"



=====================================
(2B)+(2B)' ?

 

[Artisi]

electricpete: Your FAQ237-1543: How does hp change with flow for a "centrifugal pump"

"The link is my FAQ, so I'm always interested in comments that might improve it."

A small point for clarification:

"Multi-stage pumps are typically mixed flow design".

This could do with some clarification, high flow multistage pumps can be mixed or partially mixed flow design, whereas low flow high head multistage pumps would be radial flow.

How about, multistage pumps can be either mixed flow or radial flow with applicable power demands.

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.)

 

[LionelHutz]

I would certainly agree with the camp saying that a DOL or ATL started electric motor will draw the same current profile as it accelerates regardless of load. The only thing that changes with load is the acceleration time. The motor manufacturer can even give you the current vs speed curve which shows the current at any speed when rated voltage is applied to the terminals.

 

[LittleInch]

The other reason for starting against a closed valve is to limit the maximum flow which can occur when pumping into a system where the pressure is significantly below the pressure the pump is designed for in steady state.

It is quite easy for systems to have a pump operating at or beyond end of curve conditions for a short while as the flow rate starts very high against not much back pressure which then climbs as the system starts up. End of curve operations leads to high vibrations, high current and pumps tripping before they have really got going.

More of an issue in long pipelines or where pipes have been drained, but starting against a closed valve allows for a slow opening of the discharge valve and essentially some throttling of the flow once it opens.

There is also a bit of SOP here - what individuals and companies get used to doing perpetuates as no one wants to change something that works for them....

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Valvecrazy]

I agree. Done correctly a valve can create a mechanical soft start and soft stop. This can eliminate nearly all transients and water hammer in the pipeline, as well as reduce the inrush current. It maybe more duration than amplitude, but still a much softer start.

 

[electricpete]

"Multi-stage pumps are typically mixed flow design".

This could do with some clarification, high flow multistage pumps can be mixed or partially mixed flow design, whereas low flow high head multistage pumps would be radial flow.

How about, multistage pumps can be either mixed flow or radial flow with applicable power demands.

Updated, thanks!

=====================================
(2B)+(2B)' ?

 

[itsmoked]

Good points I hadn't realized Little!

Keith Cress
kcress -

http://www.flaminsystems.com

 

[bimr]

"Assuming that the impeller is not frozen in place, the seal faces are not stuck together, the pump is full of liquid, properly vented, and the pump is not wired to run backwards; there really is no good way to start a centrifugal pump."

Mc Nally Institute Link

 

[Alex1900]

hello every body ,
i have a question about centrifugal pump with electric motor.
i reduced the pump's impeller to set it to my desired flow rate and head.calculation shows pump's absorbed power in about half of installed electric motors power.i cant re please electric motor.
now question is :in this condition the pump will get correct head and flow rate?(by assuming calculation is correct)
thanks in advance.

 

[LittleInch]

Please start a new post and don't hijack a previous one.

Also some numbers and data would help.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[LittleInch]

Alex,

I'm waiting for your new post, but please spell check it first.

But in essence the answer is yes, but needs some data to confirm.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.