Water Hammer Pump Sudden Shut off

[TechnoZA]

I would like to know what the effects would be if a pump had to suddenly be stopped due to, for example, a power failure. Is there significant risk of water hammer? What measures are recommended for preventing water hammer?

In my specific situation, I will be running 3 pumps, sucking and discharging into common manifolds, with each pump delivering around 870m3/hr. Suctions 18" and discharges 16".

 

[BigInch]

A pump stop can be just as bad or worse as a pump start. Case in point, with three pumps, if one stops, the other two pump flows will try to take the path of least resistance. If that path happens to go through the stopped pump, it will attempt to backspin the stopped pump, so you not only get the transient from the loss of product velocity when the first pump trips, you get the reverse surge from the other two pumps product when that turns around and runs back into the tripped pump.

Reverse flow can be avoided by a check valve, however it may still be a good idea to take a look at what can happen, if the check valve malfunctions or breaks.

There are a number of ways to prevent or mitigate transient surges, a fast acting control valve, a soft starter, a surge "accumulator", a pressure relief valve (although sometimes they might complicate the problem), a recycle valve back to suction that can be slowly closed. Best way is dependent on how you want to operate and the cost of the device.

See this recent thread for another case I just finished discussing.
thread378-222072

"If everything seems under control, you're just not moving fast enough."
- Mario Andretti- When asked about transient hydraulics

http://virtualpipeline.spaces.msn.com

 

[TechnoZA]

Thanks for that. We will definitely have check valves on all pump outlets. We will also be operating with VSD's in starting and stopping. The problem comes in when we have a sudden power failure, which happens all too often in South Africa unfortunately. In the case of sudden loss of power, and no VSD's to shut down the pumps, how quickly will the water come to a stop, and will the effects be damaging?

 

[BigInch]

I can't tell you that without knowing all the details of the pipe, pressure, flowrate and configuration of the piping system. Also depends on the moment of intertia of the pump and driver's rotating parts and the mass of the spinning fluid around the impeller as to how fast the pump slows down. What happens in the pipeline depends on the initial velocity, its momentum and the compessibility of water and the "expandability" of the pipe. Reflected transient wave arrival timings depend on the length of the attached pipe. Most of the time, you can generally assume that the transients will have dissipated within 4 or 5 round trips down the pipeline and back (at the velocity of sound in the pipeline's fluid), but that's a very very general statement and it is NOT applicable to all piping configurations.

The theoretical maximum transient pressure can be found by assuming that velocity head (v^2/2/g) is totally converted to pressure head, which you must both add and subtract to the operating pressure at any point within the system to find the total pressure. Add transient pressure to the operating pressure when the wave is outbound and subtract on the return. Actually some energy is dissipated via friction and expansion of pipe and product as the wave travels, so it should be less than the max, but reflections from any points where the velocity changes, at reducers, tees, etc. can affect the original waves at inopportune times, causing two or more transient waves to add together, so the "theoretical maximum" is not including reflected additions when multiple waves are possible.

A computer analysis is really the best way to solve anything other than a very simple system.

And you're absolutely correct in that VFDs might be nice for startups, but do not help at all in power failure situations. Transients have also been known to shatter the disks in check valves ...among other things.

"If everything seems under control, you're just not moving fast enough."
- Mario Andretti- When asked about transient hydraulics

http://virtualpipeline.spaces.msn.com

 

[waross]

Hi BigInch. I thought that the pressure transients when a valve was closed fast enough to cause water hammer were positive on the upstream side and negative on the downstream side. The positive transients may easily reach pressures that will burst piping and pump cases. The transient on the downstream side is negative may approach zero absolute pressure, but will not exceed the the magnitude of the original head in absolute units. (PSIA)
If I am wrong, please be gentle.
By the way, once a motor has tripped, there is no connection to permit a "back surge". However, the other two pumps will abruptly shift to a different point on the pump curves. I would not be surprised to see a load surge developed by the two running pumps. This may be mistaken for a surge from the tripped pump.
Thanks.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[BigInch]

waross,

Yes it is correct that the transient is positive on the upstream side and negative on the downstream side of a closed valve. The downstream wave loses amplitude with friction and compression of product ahead of it. When it is reflected, it continues to lose energy on the way back and the amplitude of that one wave is always less when it returns. If the pressure goes below atmospheric pressure, or external water pressure on an offshore pipeline, a pipe of relatively thin wall thickness to diameter ratio might collapse. At least all of that is true for a flat pipeline and a valve that does not have a feedback and control function. If I said or implied something differently, sorry for the confusion.

Now, once you start adding pumps, control valves, relief valves and realistic elevations to the Joukowski equation, it loses its conservativeness. You must consider the action time of the controls in conjunction with the transient pressures wave leaving, arriving or passing by and how much energy you have stored in the pressure expanded pipeline, in the compressed product itself and the potential energy of all product at higher elevations downstream (and sometimes upstream too). If you are pumping to a high elevation, which is what a lot of pumps do, there is a lot of stored energy up there that is more than ready to go back to where it came from. When the pump stops, it can be free to do so. Mr. Murphy says it tends to do it at inconvenient times.

I'm not following what you mean by "connection". Are you saying a discharge shut-in valve, or a check valve closes to "disconnect" the pipeline from the pump?

An example of a case I was referring to would be something like, only one pump running and discharging into a downstream pipeline that has pressured the product and pipeline to its operating point pressure. The pump's motor abruptly stops running and the pump stops turning in a 1/2 second or so. The pipeline is still pressured to operating pressure and that hi pressure product will begin to flow back through the pump's discharge into the suction of the pump, tending to reverse spin the pump, effectively turning it into a turbine. If two additional pumps were running and discharging into the same manifold, they would only make things worse, since the manifold and pipeline would then have a continuing source of product and pressure, instead of a decreasing amount of product and pressure as was the case with only one pump running and tripping.

I've had cases where a slight overpressure is relieved through a relief valve, which stops product flowing into the pipeline downstream as pump discharge flow is diverted to the relief valve. The pressure near the relief valve begins to decrease and the relief valve closes. Product already downstream of the pipeline then comes to a stop and velocity head is converted to pressure head, increasing the pressure and reopening the relief valve again, just as the pump's discharge control valve is opening back up again. It created a continuous cycle between the pump discharge pressure control valve, the relief valve and the pipeline downstream that was difficult to stop without shutting down the pump and completely restarting the system again from 0.

"If everything seems under control, you're just not moving fast enough."
- Mario Andretti- When asked about transient hydraulics

http://virtualpipeline.spaces.msn.com

 

[BigInch]

If you think about it from design perspective, typically as you say, reflected pressures are going to be less than the initial operational + transient pressure and the pipe remains intact, but from an operational standpoint there is still a chance it might not work.

"If everything seems under control, you're just not moving fast enough."
- Mario Andretti- When asked about transient hydraulics

http://virtualpipeline.spaces.msn.com

 

[waross]

Thanks for taking the time and effort to respond to me, BigInch.
I misunderstood one of the responses. I thought that someone was expecting a power surge of some type and was pointing out that the trip that had caused the problem would have disconnected the electrical feed to the pump, making an electrical surge from the tripped pump impossible but still allowing load surges from the running pumps.
Sorry for any confusion I may have caused. My bad.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[katmar]

I have found this thread very interesting because it is an area that I know very little about, but about which I know just enough to know that I need to know more! So, rather than making a comment, I will extend the original question. Being South African also helps me identify with the reason for the OP having this concern.

If there is no control valve that slams shut on power failure, and the pump simply stops pumping, where does the shock wave start? I accept that the flowing liquid stores significant energy as velocity head and that this will be dissipated into the pipe supports as it slows down, but I find it hard to imagine this being catastrophic. Shouldn't pipe support design take this case into account?

If the pump is pumping against a zone of stored pressure (compressed gas or high static head) I guess the liquid would reverse quickly after the pump stopped and the check valve on the pump discharge would slam shut and I can see this being a problem. But what happens in the situation where we have a cooling water circuit and the water is simply circulating back to the tower? Is it dangerous to stop this type of pump suddenly (as opposed to closing a valve suddenly)?

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[BigInch]

On a pump trip the shock wave starts at the first point where the velocity begins to change and start the conversion from velocity head to static head and the wave amplitude develops proportionally to the rate of deceleration of the fluid. That point is (somewhere) inside the pump. So the pressure wave of positive amplitude starts at the pump and moves upstream, while downstream a negative amplitude transient is first released. The reflections of each wave are opposite in sign.

I have not looked at a continuous loop, but it seems obvious that the positive wave continues upstream and the negative wave continues downstream and, if the two can add together at any point, trouble errupts. I'll add that to my list for future study.

To tell you the truth, I haven't thought very much about what happens to the supports, but I suspect it could involve rather large forces.

"If everything seems under control, you're just not moving fast enough."
- Mario Andretti- When asked about transient hydraulics

http://virtualpipeline.spaces.msn.com

 

[Compositepro]

What kind of pumps are you guys talking about? Shutting of the power to a centrifugal pump will not suddenly stop the flow through it. Isn't that the type of pump that would be used in most of the application mentioned so far? Even large gear pumps would have some ability to be driven by flow and reciprocating pumps have spring-load valves that would allow forward flow even if the pump is off.

A piping system that can't tolerate a power failure to a pump won't last very long.

 

[stanier]

“In physical science the first essential step in the direction of learning any subject is to find principles of numerical reckoning and practicable methods for measuring some quality connected with it. I often say that when you can measure what you are speaking about, and express it in numbers, you know something about it; but when you cannot measure it, when you cannot express it in numbers, your knowledge is of a meagre and unsatisfactory kind; it may be the beginning of knowledge, but you have scarcely in your thoughts advanced to the state of Science, whatever the matter may be."

Lord Kelvin [PLA, 1883-05-03]

Stop messing around and go model the system. If you cannot hire someone who can. Then you will have some "numbers" not just "feelings".

Go to

http://www.pipingdesign.com

and look at the papers there on surge and its prevention.

regards

Geoff Stone

Surge Analyst.

 

[TechnoZA]

Compositepro: The pumps I'm using are centrifugal. I had very much doubted that there would be any significant reduction in speed on sudden shut down, but others that I was talking to weren't so sure, and hence the question.

Stanier: I agree that numbers are essential. However, there is some theory behind the numbers, and even should you get a whole lot of numbers, you need the understanding in order to determine whether or not those numbers are realistic. Furthermore, we don't all have the access to the engineering tools needed to model this, and we also don't all have the backing of management to support the idea of taking such a "possible" problem (because this is all it is in their eyes) to some specialist who is just going to cost the company a whole lot of money. My working environment is not an engineering company, actually it's a farm. So for me, at least this is a starting point to get an idea of how likely a problem this is in my scenario.Thanks for the link.

And thanks all for the input from others...

 

[Shozza]

TechnoZA in the case of power loss to VSD's the motor will acutally become a generator and depending on which VSD is being used some are capable of recovering the kinetic energy of the motor and load and use this energy to keep the DC bus powered long enough to allow the motor to ramp down to a stop rather than just allowing the motor to coast to a stop. It basically allows you to perform a soft stop even though the mains power is disconnected. Check your VSD manual and see if it has this feature.

 

[TechnoZA]

Thanks Shozza. Very interesting, I'll look into it.

 

[BigInch]

compositepro,
Centrifugal pumps both start and stop spinning rapidly enough to cause surge.

Shozza,
If the motor becomes a generator, I can only see the pump and fluid slowing down faster, possibly creating a larger net surge, since the motor would be sapping that energy from the spinning momentum of the pump. Am I incorrect? If I am, why?

TechnoZA,
What's the rated power, head and flow of these pumps, suction and discharge pressure, and diameter and length of the manifold and the downstream pipe?




"If everything seems under control, you're just not moving fast enough."
- Mario Andretti- When asked about transient hydraulics

http://virtualpipeline.spaces.msn.com

 

[TechnoZA]

BigInch,

To start with, submersibles will push water out of a river to our delivery pump station, and delivery pumps from here to one of our dams.

Submersibles:
3 operate in parallel. (Figures below per pump)
dP = 200kPa
Q = 868 m3/hr
P = 57.7 kW

Delivery Pump Specs:
3 operate in parallel. (Figures below per pump)
dP = 617kPa
Q = 868 m3/hr
P = 180kW

Pumps delivery through 2 x 400mm and 1 x 500mm pipelines in parallel. Manifold dia 600mm

Total system static head 60m
dH of subs to deliveries 12m
Length from subs to deliveries about 80m
Length from deliveries to dam about 4500m

 

[BigInch]

No promises, but due to the high level of interest in the forum, I'll try to have a closer look at these... if I get some free time, so let me add some more questions,

RPM submersible pumps?
RPM delivery pumps?
How high are the delivery pumps over the submersible pumps?
How high is the reservoir over the delivery pumps?
Do you have check valves between the submersible pumps and the delivery pumps?
And between the delivery pumps and the manifold?

"If everything seems under control, you're just not moving fast enough."
- Mario Andretti- When asked about transient hydraulics

http://virtualpipeline.spaces.msn.com

 

[TechnoZA]

Submersibles and Delivery Pumps at 1450rpm
Submersibles to Delivery Pumps has dH of 12m
Delivery pumps to downstream reservoir dH of 48m
Certainly check valves btw. delivery pumps and manifold. Not 100% sure if there are check valves between subs and deliveries. We're busy upgrading, so could always add.

Thank you!

 

[waross]

Does this sound reasonable;
There will be an equilibrium flow and dynamic head with three pumps running.
There will be an lower equilibrium flow and dynamic head with two pumps running.
When a pump is stopped the system will coast down from one state to the other.
Think shifting your car out of gear and coasting rather than hitting a wall.
The rate at which the system "coasts down" will depend on the inertia of the fluid in the piping, the dynamic and static heads, and on the differential head across the pump.
Think coasting on the level versus coasting up hill.
I would expect the flow to continue through the unpowered pump due to inertia, and the flow rate to decay with no serious pressure transients.
HOWEVER, the flow through the stopped pump will drop to zero and then reverse. At this point the check valve should close.
But, if the speed at which the flow decays and reverses is short in relation to the closing speed of the check valve, so that there is significant back flow through the check valve before it can physically close, the check valve will slam shut generating pressure transients that may damage the check valve and/or other components.
In the event of a power failure, all six pumps will stop. I would expect the flow rate to drop to zero and then reverse. Any pressure transients would depend on the closing time of the various check valves relative to the time for the flow to stop and reverse.
Hello BigInch. If this post has any validity, great.
If this just causes confusion then please RF it.
Thanks.

Bill
--------------------
"Why not the best?"
Jimmy Carter