Pump Turbine Flow

[BRIS]

I am reviewing a water supply booster pump station design;


Pipe length 40 km, booster station is located at mid point, diameter 0.5 m, design flow 0.26 m3/sec, static lift 51 m, dynamic lift 125 m, two duty and one standby pump arrangement. (design duty each pump 130 l/sec at 176 m)

Pumps are specified to have maximum speed of 1400 rpm which means vendor will likely need to supply dual stage centrifugal pumps.

The designer has not provided a by pass around the pump station but has provided surge tanks on the downstream of the pump station to prevent surge pressures dropping below atmospheric.

My initial comment is that a by-pass should be provided around the booster station. My calculations suggest a by-pass would adequately limit surge pressure to within allowable without the need for surge tanks. I would provide a 500mm diameter by-pass with a nozzle check valve. However I have to review the design as it is not as I would do it!

With the arrangement as it is - on a power failure, or even a normal pump shut down, positive surge pressure in 20 km of upstream pipeline will create a head of up to 100m across the booster pump station, without a pump station by pass flow will continue under this head through the pumps. (for at least 35 seconds).

My question is what affect will this head have on the pumps?
Under a power failure will the pumps continue to rotate (turbine) or will they stop with flow passing through the stopped impeller possibly giving cavitation problems?

i.e. what happens to a stopped pump that a has a pressure difference of 100m between suction and delivery?

 

[TenPenny]

i.e. what happens to a stopped pump that a has a pressure difference of 100m between suction and delivery?

--------------
Typically, the pump will rotate backward, which, depending on the pump design, will cause impellers and shafts to unscrew and fail. Or, if you are lucky enough to have the pump auto restart while rotating backward, the shaft snaps.

Typically, vertical turbines will have NRRs (non reversing ratches) on the motors for this reason; another solution is something like the GE Multilin MM2 motor protection relay, which allows to to prevent restarts on reverse rotation.

 

[Artisi]

TenPenny
The impellers cannot unscrew if running backwards from reverse flow, the impellers can / might/ maybe (depending ondesign) unscrew if DRIVEN backwards by the driving unit?

A backstop on vertical turbines is usually fitted only to prevent running into critical speed or to prevent damage should the unit attempt to start with running in reverse as you have already pointed out.

 

[BigInch]

Are you sure a bypass would help? A closing ball or gate valve will essentially stop flow only over the last 10% of travel, so will allow pleanty of time for any surge pressures to enter the station. In order to protect the station, the valves would have at best only a few seconds to close, if the actuators were triggered by an initiating event well away from the station, such quick closing action in itself possibly triggering another event as well. You may want to do much more than preliminary calculations where pump ramp-down, actuator response time and valve characteristics are all considered.

http://virtualpipeline.spaces.msn.com

 

[BRIS]

Big Inch - I understand what you say but with a by pass the upstream pressure can only exceed the downstream pressure by the amount of head loss through the nozzle check valve so the pressure difference across the pumps can only be a few meters.

Without a by pass the pressure difference across the pump station is controlled by the flow and head loss through the pumps.

The upstream and downstream pipelines are both 20Km long and the wave period 2L/a is some 36 seconds. I would expect pumps will come to a complete stop within 10 seconds - I have therefore simply calculated upstream and downstream pressures using the Joukwosky equation and that gives me a maximum 100 m head difference across the PS on a pump trip. With A by pass this head difference will not materialize: the by -pass check valve will open as soon as the rising upstream head has risen above the falling downstream head. I therefore don't consider that I need to provide additional PS protection.

 

[BigInch]

I'm just wondering why the surge would choose to go through a 2" check valve when it has a perfectly good pump station with a nice large diameter entrance to hit. Surges also take the path of least resistance.

http://virtualpipeline.spaces.msn.com

 

[BRIS]

for 2" read 500mm or 20" in those big inch units

 

[Valvecrazy]

Surge, water hammer, or transient pressure waves in a booster system such as this, are caused by the pumps starting and stopping or valves opening or closing. The best way to eliminate the surge is to never completely stop the flow. Continually varying the flow rate to match the demand will eliminate transient pressure waves from pump starts and stops. A stand pipe or surge tank at the end of the line, can help eliminate surges from valves closing and opening. Then a surge anticipator / pressure relief valve can reduce the surges from pumps stopping because of a power failure.

Controls for the pumps that can vary the flow rate, must react instantaneously as with fast acting constant pressure valves. Slow responses to changes in system pressure, as with soft starts or slow ramping VFD controls can actually accentuate the pressure waves.

See this link;

http://www.cyclestopvalves.com/waterhammer_3.html

 

[BigInch]

Right Bris, ok 20".
Are the pump's suction and discharge diameters 20" also, or do they reduce to smaller diameter pump flanges?

http://virtualpipeline.spaces.msn.com

 

[BigInch]

Bris, From what little you said about this system, I made a quick model and made a few quick approximate runs. "Results" show no problem for starting or stopping the pump, provided the pipeline inlet (KP 0) and outlet valves (KP 40) are open. Note: this is with very sketchy info for the pump.

Worst case scenarios appear to be,
Opening pipeline inlet to running pump.
Closing pipeline outlet on a running pump.
Starting pump against closed pipeline outlet.

During which pressures reaching 550-600 psig were observed without any antisurge devices.

I'd suggest that you use slow open/close valves on pipeline inlet and outlet and do not start pump against a closed pipeline discharge valve. If you were to follow those recommendations, it is possible you might not need surge devices, depending on MAOP of the system. If >= 600 psig, probably not required.

http://virtualpipeline.spaces.msn.com

 

[BRIS]

Big inch - sorry for the delay in acknowledging your reply.

1) I was in the desert (Sahara)with no internet 2) I had to recall what a psig was in real units. 600 pssi = 41 bar.

Pipe is DI and class K9 with PN 40 flanges and has an allowable PMA (pressure rating including surge) of 48 bar. So we are within allowable -( there are a couple of valleys along the profile where pressure are near the limit).

That is without a by-pass or other surge alleviation device (surge tanks). I have already advised the designer that increasing flange and fitting rating from PN 25 to PN 40 will give a much more robust solution than providing surge tanks. (Bearing in mind this is North Africa and the last time anybody maintained anything was in 1942 when Montgomery repaired flat on his Jeep).

I am still back to my original question that without a by-pass on a pump trip due to power failure (a regular event) there is a 10 bar (145 psi) driving head across the pump station. This will force water through the 'stalled' pump impeller. I guess that the pump continues to rotate until this head dissipates ??.

I am not comfortable in not having a by -pass around the PS but I don't have an argument other than preference.

 

[BigInch]

No worries.

I was 10 years over to the eastern side. Yes. Simple is the way to go, the more the better.

Not knowing the true profile, I just ran it with a steadily increasing head from the pump station to the end of the pipeline.

Yup 40 something barg.

If there's no discharge check, the downstream head will tend to reverse flow through the pump until a pump station block valve shuts. I wouldn't think you'd want to do that and I would put in a discharge check, if there is not one now.

Other than that, I didn't see anything astounding other than what I've already mentioned. Remember, that is... with a steady slope (no valleys) on the last half of the pipeline. Valleys can change things, esp if any pressures are near vapor pressure at the peaks.

http://virtualpipeline.spaces.msn.com