Hydraulics - side weirs as flow distribution in a channel 13

[hobeex1]

I am trying to calculate how flow will be split over 4 side weirs coming off a common channel. The channel is a feeder channel to 4 identical process tanks (stopping just after the last weir), and the aim is to design the system such that as equal a flow split as possible is achieved between the 4 tanks fed from it.

I am trying to convince someone that, although there may be some variation in discharges over the 4 weirs, it can be kept to a minimum by selecting the correct channel cross section and weir heights, thereby eliminating the need for a distribution chamber. I'm hoping to keep the weir heights equal if possible.

I'm at the edge of my hydraulic knowledge with this one, so I'd appreciate any advice on whether I'm correct in my assumption, and how I can prove it.

 

[MikeHalloran]

Sometimes you have to convince people, or their replacements, over and over again. Doing something silly may save time in the long run. Silly might include forking the channel and forking the forks, with completely symmetrical arrangements, so that ANY FOOL CAN SEE that the flows should be equal. The supply of fools is unlimited, and if you have to deal with them individually, you won't have time for anything else.

One thing you might consider instead is having the four weirs cut by a laser or a waterjet under CNC control, so that they are all equal. Do not allow anyone to deburr or radius their edges, of course.

You might even put a bridge over the top of the weir proper, with a row of tapered needles of different lengths projecting down toward the design flow level. It's no trick to put all those features in the weir plate, so no adjustment is possible, and they'll all be the same, etc. Then anyone can tell at a glance if the weirs are balanced, just by counting the number of needles breaking the flow surface. If your liquid is free of flotsam and foam, that should provide continuous verification that the flows are balanced.

It doesn't provide a means for actually balancing the flows. One way to do that is to put vertical slots in the weir plates for the mounting bolts, so that you can slide the plates up and down at startup to balance the flows.

You will still need a barrel and a watch to convince the hardheads.




Mike Halloran
Pembroke Pines, FL, USA

 

[bimr]

There is no way to get an equal split if you are coming down a deadend channel and you desire to take flows off the side equally using weirs. If you tried to put adjustable weirs on, you would be able to precisely split the flow at a certain channel velocity, but then you would have to readjust the weirs at each change in channel velocity. It may be workable if you have a constant channel flow.

A better solution would be to come to the end of the channel, go under a velocity dissipating baffle wall and then have the flow go upward over the 4 weirs. This may be what you are referring to as a distribution chamber.

 

[braxtonlewis]

Assuming I have understood your problem correctly where the flow on your "side" weirs would be perpendicular to the flow of the common channel; a greater pressure head would exist at the first or most upstream weir. This would allow more flow through that first weir. However, as you suggested this problem can be alliviated through design of the system by altering both width and height of the weirs. But this solution would only work assuming a steady state flow situation, which is most easily attainable by using a distribution chamber.

Braxton V. Lewis

 

[kchayfie]

If the head loss along the channel is minimal and all the weirs are at the same height, this kind of flow split is accurate enough for most practical purposes.

Keep the channel wide enough to ensure there isn't a high velocity and the head loss between the first and last weirs is of the order of a few mm and the difference in flow over the first and last weirs will be negligible. Use adjustable weir plates as Mike suggested to make up for any variations introduced when constructing the channel.

 

[JedClampett]

I'm just a structural slug, but our company does flow splitting design quite often. What I've observed is that the two weirs at the end of the channel will have the flow biased toward them. Where the channel ends there is a reversal of flow, the liquid tend to "pile up" and more flow goes over these weirs. It seems as if it should be the opposite, but it's not.
If you need a precise split, you'll need to try what bimr suggests above.

 

[Ussuri]

It is possible to use a flow splitter chamber to split flows equally to four process units with a weir arrangement. You need to arrange your design so that the velocity in your chamber is zero and you have sufficient hydraulic gradient in your upstream pipework to develop a head in the chamber and back up the system. You can do this by a low level entry into the chamber and set your weir height above this.

If you have a velocity of any kind in the chamber your backwater profile will vary with distance, which will vary the head over each weir and consequently the discharge.

Couple of potential problems, depending where you are in the process, silting up and solids deposition may be an issue. And you do require a reasonable hydraulic gradient.

A useful document on side weirs is 'Hydraulic Design of Side Weirs' By RWP May, BC Bromwich, and CE Rickard

 

[49merc]

What flow rate are you designing for? A fairly simple distribution box, like those used in filtration plants, will probably work. They are often set up with flow going under one set of weirs and over another. This stabilizes the distribution and eliminates the effects from wave action. If the flow is usually steady, the design will be pretty straightforward.

 

[hobeex1]

Thanks for all the advice - in answer to the last question, the flow rate is variable between 347l/s and 573l/s (sorry if you don't use these units) but is generally constant at about 521l/s.

In order to avoid settlement of solids I need to keep the velocity above 0.4m/s.

I ran a computer simulation of the problem, and discovered what most of you had said - it works great for really low velocities, but to avoid settlement I had to vary the weir heights. This worked for one flow, but the split was out by about 10% at low flows.

I have alleviated this slightly by gradually narrowing the approach channel to increase the velocity and depth. I have also decided that adjusting lots of weirs is pretty tricky, so to control the head in the channel in a more user friendly way, I am putting sluice gates in to provide orifices of varying sizes coming off the channel, before the weirs. The weirs will all be at the same height. I think that a simple level indicator such as a gauge board or ultrasonic head in the channel could be used to verify the depth variance and hence accuracy of flow split along the approach channel.

I found JedClampett's point about the piling up at the far end, and although I can't model this, hopefully fine adjustment of the sluice gates will help.

Needs further work, but judging from the model it seems to have some potential (achieving splits with a 2-3% variance at low flows), however I need to ascertain what tolerances the plant will accept.

I confess that I am leaning further toward the separate splitting chamber solution, but am going to discuss the above solution with the plant manufacturers.

 

[MikeHalloran]

There is a weir shape that is intended for constant proportion metering over variable total flows. I forget what the name is. It looks a little like a v-notch weir, except the notch faces are convex, so at low flows the height varies a lot, and at high flows the height doesn't vary so much. I had an equation for the shape ... on a hard drive that's not accessible now.



Mike Halloran
Pembroke Pines, FL, USA