minimizing water hammer 1

[marc400317]

Hello All,
I am designing a pump station ( 3 pumps, 2d/1s) pumping brackish water with a total flow of 70m3/hr through a DN160 Ductile iron pipe buried 1.5 m from FGL. The total length of the pipe is 2700m discharging into a Box Culvert. To reduce the water hammer effect I have installed a pressure tank at the pump station upstream and a swing Check valve along the line on every 200m. Is 200m a reasonable distance? is there any reference showing the right distance between Check valves?
Thank You
Marc

 

[BigInch]

your velocity is only 1 m/s, so I would not expect too much of a water hammer problem. Certainly the pressure tank would be enough, if that is required at all.

I am not aware that check valves would help the problem, should it exist.

If you do develop a problem, it will most likely be from pump start/stops, in which case a soft starter would help on startup, but not do much good on a trip.

From "BigInch's Extremely simple theory of everything."

 

[KiwiMace]

I don't think you need all the swing checks either. I have had such systems modelled in the past by 3rd party consultants. Solutions has generally involved the pump start/stop or weighting the swing check to get a particular slow closing rate.

 

[marc400317]

Thank you all, well that's what I thought in the first place however our consultant would like to be convinced somehow. anyone aware of a good water hammer simulation software or reference?
Regards,
Marc

 

[chicopee]

About checking some of the past designs used on hydroelectric plants. Water hammer had to be controlled in the event water flow had to be shutdown at these plants.

 

[BigInch]

Assuming the water does not reach vapor pressure in a negative pressure wave event ...

Those checks would keep a pressure wave being reflected back from the end of the pipeline from returning back to the pump, but reflected waves are much less than the original pressure, so any damage has essentially already been done. A closed check would also be a cause of reflection of any subsequent waves emminating from the pump station, which such valves being closer to the pump station would reflect proportionally higher pressures and reflect them sooner.

A very small check valve as a bypass around a fast closing valve that was causing a waterhammer problem could be effective at allowing some of the wave to pass that valve thereby capping the pressure rise ahead of the valve and any wave reflected back to the pump station, but still, the highest pressure at the pump has already occured. Reflected waves are of lesser pressure than the original pressure event causing the wave.

From "BigInch's Extremely simple theory of everything."

 

[marc400317]

Well Said BigInch and thank you for your answer. I am somehow conviced though as you said if reflected pressure waves are lesser than the original pressure then why do we bother going for a transiant hydraulic analysis measuring for the water hammer pressures ? or is it only when a valve downstream is closed suddenly?

 

[BigInch]

The largest pressure occurs at the location of the event causing it, ie, a valve closing downstream will cause maximum pressure at the upstream face of the valve. When that pressure rises to the maximum, reverse flow begins to carry the pressure wave back to the origin of the pipeline, however frictional resistance to that reversed flow causes the pressure along the pipeline now to fall from that maximum at the face of the valve toward the pipeline's origin. In effect the face of the valve becomes the new pipeline "inlet" and you now have to deduct friction from the reverse flow velocity for each and every foot that the reverse flow travels as it goes back to the pump. When that wave reaches the pump and presumedly a now closing discharge check valve there, the pressure has been decreased from the maximum considerably. When the wave hits the closed check, now at a slower velocity, the pressure spike from that reflection is not as high as the first one originally created at the downstream valve, and it reverses again; flow is once again in the original direction, but now at slower velocity. That occurs 3 to 5 times, each reflected spike less than the previous, before things eventually calm down. The time between spikes is the time that the pressure wave takes to move between original closed valve and pump at the sonic velocity of water, close to 3000 ft/sec. So if the valve was 3000 feet from the pump, it would be 1 second before you saw the first reflection reach the discharge check and 2 more seconds before you see the next spile reach the pump, now the 3rd reflected wave. At about 4 or 5 reflections, there probably isn't much left of the original max pressure wave.

BTW a partially closed valve, or a reducer in the pipeline will reflect a spike proportional to the amount of x-sectional flow area reduction it gives, ie a valve 40% open will reflect 60% of the wave back to the pump while allowing 40% to pass through. A 6 x 3 reducer will reflect 75% of the spike allowing 25% of the spike to pass.

Checks in the line would probably stop most of the reflections from traveling from the closed valve back to the pump, but wouldn't do much of anything to mitigate the original pressure spike at the face of the closing valve. IMO.

From "BigInch's Extremely simple theory of everything."

 

[stanier]

Prof ARD Thorley in Pressure Transients in Pipeline Systems likens swing check valves to the most basic of agricultural machinery. Adding weights only complicates the ability to actually resolve any waterhammer issues. fast acting non return valves ( Noreva , Mokveld etc) with a low mass of moving parts is what is needed to best avoid surge. That said without modelling th system as a whole you are just guessing as to what the dynamic response will be now and into the future.

By pressure tank do you mean a vessel with a pressure sustained by compressor or gas charged bladder. If so it is more important that the check valve at the pump station is fast acting of the non slam type.

Putting check valves every 200m appears to be inconsistent with any solution I have known. You may solve the challenge with one intermediate check valve but its location depends upon the pipeline profile and modelling.

Without the pipeline profile it is difficult to estimate the pressure spikes.

Suggest you invest in some engineering and have the system modelled by a competent waterhammer analyst. Using Joukowsky to estimate the maximum presure transient is courageous.

Without know your maximum transient pressure you dont know the design rpessure of the pipeline.

As for software you can choose between AFTs Impulse, Hytran, Kypipe, Flowmaster, Watham and numerous others. Be warned that purchasing the software does not make you a transient analyst.

"Sharing knowledge is the way to immortality"
His Holiness the Dalai Lama.

http://waterhammer.hopout.com.au/