Designing a Water Hammer Flow Loop - Where do I start?

[vgurevich]

I am working on a project where I need to design a flow loop capable of generating a water hammer event.

It needs to be fairly simple, and fairly compact (approx 2x3 feet).

I was thinking of something simple - a rectangular flow loop with a pump, and a valve on a straight section. When the valve is actuated, it stops the flow and generates a water hammer event.

My background is aerospace engineering, and most of this work is entirely new to me. Any advice or helpful reading links would be greatly appreciated.

 

[btrueblood]

I'd think a bit bigger...but you don't say what line size you want to consider. The problem with short lengths is that the event zips by so quickly, you may not learn much. You will want to generate fairly high line velocities, which requires head, so choice of pump will be critical (you might consider a variable speed drive). You will want to build this with a tub underneath, and a means of draining same, so you can collect and dispose of the fluid after it all comes out when a pipe/fitting joint ruptures. Finally, you will likely want a clear plastic splatter shield.

 

[vgurevich]

Space constraints are a bit of an issue for going bigger. This is meant to be a pretty small/simple experimental set up.

The goal is to trigger a water hammer event on a straight run pipe, and measure the response of the piping to the pressure surge associated with the water hammer. It would be a parametric study - changing parameters such as pipe diameter, valve closing time, fluid properties (fresh vs salt water), etc.

Ideally there would be no ruptures or failures, and the system will be over designed to withstand the elevated pressures.

My main problem is I don't have a feel for how all the components will go together. Does the pump need to be in-line with the flow?

 

[MJCronin]

If you are just interested in generating a pressure wave on an instrumented setion of piping, I would consider the methods used by the "pumpkin chuckers" on the Mythbusters TV show.

The air-cannon configurations seem to use a compressed air configuration with a quick opening butterfly valve.

Substitute a slug of water for the pumpkin.......and you got it.

However, I do agree with the above post regarding length of piping.

I believe that you will need a much longer run of piping to get results.

Specifically, what effects are you interested in evaluating ?

 

[vgurevich]

Without getting into too much detail, I'm interested in studying the interaction between the propagating pressure wave and the surrounding pipe.

Unfortunately the motivation for this project is similar to a flow loop with a valve, and so the small scale version must be fairly similar.

 

[hydtools]

The event will travel at about 4700ft/sec, 0.0002sec/ft. You will need some fast response measuring equipment to capture the initial event before it is confused with reflections in the short piping lengths.

Ted

 

[vgurevich]

Wouldn't a 20kHz accelerometer be able to accurately pick that up?

If not, how long of a pipe section would I need? Let's assume a 6 ft x 1 ft rectangular flow loop, with a valve in the middle of the 6ft 'test section'. 3 ft would be ~.0006 sec in between reflections, or ~1.7 kHz. An accelerometer with a 20kHz frequency response should be able to resolve the different peaks.

 

[btrueblood]

Another part of the problem is whether you can actually get your valve to close fast enough to see a shock pulse generated within that short a distance, and for the pulse you are seeing via the accelerometer to not be confounded by motions induced by the valve armature itself. Would suggest adding pressure transducers as well.

 

[vgurevich]

Yeah, I will add pressure transducers where I can. Pressure transducers will be hard to come by in that frequency range.

The pump I will be using is a Taco Inc Model FE 1206 3500 RPM, 5.45" impeller. It is capable of ~100 gpm flow w/ approx 75 ft head.

 

[MortenA]

You can use the Wikipedia text:

http://en.wikipedia.org/wiki/Water_hammer

Theres an equation to calculate the min valve closing time in order to avoid a hammer.

Best regards

Morten

 

[gustorf]

We have built a demonstration rig to show water hammer effects comparing swing check valves and nozzle check valves. see attached video.
Maybe this helps?

 

[stanier]

Check out Delft laboratories website. There are some technical papers on their test rig there.

Any pressure transducers must be fast acting and data logger nust be capable of at least 100 points per sec. Check out

http://www.ventomat.com

and see their explanation of of their research into air valves and waterhammer. Without these equipment they missed the peaks in surge events.

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

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

 

[davefitz]

As I recall, some other variables that affect the test are the pipe materials youngs modulus ( steel vs pvc would have major differences), and air entrained in the loop would act as a compressible damper and snubber.

To the extent piping flexibilty yields minute changes in piping volume for minor pipe deflections, then it might follow the piping system flexibility would affect results as well.

 

[MiketheEngineer]

Hydtools -

4700 ft/sec is like 3200 mph or about Mach 4.5 - faster than an SR-71.

If anything was flying around my house at that speed - I would leave!! If there was anything left to leave!!!