10

I am making a testing facility for pumps. This facility has to be able to handle 500 m3/h with at little turbulence as possible. The basin itself is 5 meters long, 2,5 meters wide and 2 meters tall. The water level in the basin is kept at 1,9 meters.

The basin is designed in two sections: the first is where the discharge flow comes into the basin (pipe ends about 1 meters in from the bottom). In that area the water flows freely.

After that it has to flow over the retaining wall (1.7m high), which should create a somewhat uniform flow.

overview

Are there better solutions to do this?

One other idea I have is a retaining wall somewhat like this:

wall idea

This makes for three plates with cutout stripes.

The idea behind this is to have a more uniform flow across the entire basin instead of a overflow, making it create a small bit of turbulence there.

Arjo
  • 103
  • 5
  • Do you have more specific requirements? Can you suggest an adequate Reynolds number or criteria for turbulence? – willpower2727 Mar 30 '16 at 12:33
  • 3
    Is the retaining wall just there to smooth flow? What's the point of trying to eliminate turbulence in your tub - you're almost certain to have turbulence at the suction of the pump at those flows. – Chuck Mar 30 '16 at 12:56
  • @willpower2727 I have been looking into Reynolds, but i find it hard to come up with a good solution to make that calculation. What would be a good goal? should i be using the area above the wall as surface area? – Arjo Mar 30 '16 at 13:02
  • 1
    @Arjo I rarely do fluid calculations, but you could look into finding a hydraulic diameter for the tub:https://en.wikipedia.org/wiki/Hydraulic_diameter – willpower2727 Mar 30 '16 at 13:08
  • @Chuck we want to eliminate the turbulence as much as possible to get a clear measurement infront of the pump. – Arjo Mar 30 '16 at 13:25
  • 3
    I'm hardly an expert, but I would consider replacing the upper portion of the solid partition (above the level of the discharge pipe openings) with a couple of layers of some sort of metal mesh or screening. I assume the tank is kept fairly full (1.5 - 1.8 m) during a test. – Dave Tweed Mar 30 '16 at 14:15
  • @Arjo - What is "In front of the pump"? What is the diameter of the pump suction pipe? Where is the suction pipe relative to the edges of the tank? Where is the test equipment relative to the suction pipe? What measurements are you trying to take? You're pumping at a high enough flow to swap the water in the tub every 3 minutes - why not just measure the effluent? I don't think your tub is nearly large enough to get to no turbulence, and I'd wager your inlet pipe will never be laminar, so I don't know what you're doing that "low" turbulence (still turbulent flow) gets you anything. – Chuck Mar 30 '16 at 17:56
  • @Chuck The connection for the suction hose is 300mm in diameter. this is the most far right valve on the basin. after that there is 1-2 meter of flexible hose, also 300 in diameter. Then there is a measurement tube as described in the iso 9906:2012. Its lenght is 5 times the diameter. (with a pressure sensor at 2 x diameter.) And after that comes the pump. Essentially we want to take a full test of how the pump runs, which include pressure, flow but also the electric behaviour. – Arjo Mar 31 '16 at 11:12

2 Answers2

2

In wind tunnels usually meshes and/or honeycombs are used to reduce turbulence.

However, to really achieve low turbulence it's probably not enough to add a single component to your facility. I rather suspect you need to evaluate the overall design. Some kind of diffusors at the pipe end(s) might be helpful. If you look at the flow behind bridge pillars, you can see that turbulence can be preserved over long time scales also in water.

This is non-trivial.

Robin
  • 679
  • 4
  • 8
2

Assuming you are wanting to minimize turbulence in the region of the pump, you would probably need to use flow straighteners somewhere upstream of the pump. You would have to study the optimal diameter and length of the straighteners. MIT simply used drinking straws to do this in a wind tunnel. Here are a couple of the more well known papers on the subject:

Wes
  • 503
  • 3
  • 7