Flow Distribution Chambers and Splitters: Getting an Equal Split to Parallel Treatment Units

How sensitive a weir split is to crest level

Most distribution chambers split flow over parallel rectangular weirs of equal length. Over a sharp-crested weir, Q ∝ b h^1.5, so for two weirs sharing a common upstream level the ratio of their flows depends only on their heads. A difference δ in crest level between two weirs at head h gives a flow error of about 1.5 δ/h. In numbers:

• Head 100 mm, crest difference 5 mm: 7.5 percent more flow over the lower weir.
• Head 50 mm, crest difference 5 mm: 15 percent.
• Head 200 mm, crest difference 5 mm: 3.75 percent.

Two rules follow. Design for a generous head over the weirs at minimum flow, 150 mm or more where the hydraulic profile allows, because sensitivity falls with head; and make the crests adjustable (stainless plates on slotted fixings) and level them with a surveyor's level, not a spirit level, against the water surface at low flow. A chamber poured to normal concrete tolerance of ±10 mm and left unadjusted can mis-split by 15 to 30 percent at low flow.

Approach velocity and the momentum problem

The weir equation assumes the approach velocity head is small relative to the head over the crest. An inlet pipe entering a compact chamber carries momentum that piles water up against the far wall and starves the weirs nearest the inlet, and the jet persists for ten or more pipe diameters. Compare the approach velocity head V²/2g with the weir head: if it exceeds about 10 percent of h, the split will depend on flow rate and the chamber will not behave as designed. An inlet at 1.5 m/s has a velocity head of 115 mm, which against a weir head of 100 mm is not a correction, it is the dominant term.

Model studies on compact splitter boxes with high inlet flows found that the standard remedy, a long approach length before the weir, is rarely available, and that the practical fix is to dissipate the inlet energy inside the box. Best of the arrangements tested was a vertical perforated plate across the full width and depth of the chamber between inlet and weir, with about 40 percent open area; higher porosity passed too much momentum and lower porosity added head. A ramp or sloping floor rising from the inlet toward the weir further smoothed the surface. With plate and ramp fitted, the measured split was the same percentage at every flow tested, which is the definition of a working chamber. Inlet baffles in clarifiers and the stilling zones ahead of measuring flumes obey the same principle.

Symmetry and the sequential-offtake trap

A chamber whose weirs are arranged in a line along a channel fed from one end cannot split equally without correction, because the water surface falls along the channel as flow is drawn off, and velocity head is converted to pressure head as the flow slows (the surface actually rises toward the dead end in a frictionless channel). Its last weir sees a different head from the first. The sound geometries are those in which every outlet sees the same path from the inlet: a central inlet rising into a circular or square chamber with weirs on its perimeter, or a symmetrical tee-and-branch manifold. Where a linear channel is unavoidable, the offtakes should be sized as orifices or submerged ports with a head loss several times the surface variation along the channel, so that the surface variation becomes a small fraction of the driving head, or the weirs should be individually adjustable and set by measurement.

Pipe manifolds follow the same logic: a header with a series of branches splits equally only if the branch head loss is large compared with the header friction and velocity-head recovery between branches. A rule that works for distribution manifolds in filters and aeration grids is a branch loss at least five to ten times the header loss between the first and last branch, which is achieved by orifices or nozzles at the branches rather than by a smaller header.

Submerged orifices versus weirs

Weirs are the common choice because they are visible, adjustable and self-regulating in level. Submerged orifices or sluices (Q ∝ a √(2gΔh)) have lower sensitivity to level differences, since flow varies with the square root of head rather than the 1.5 power, and they are better where floating material or foam makes a weir crest unreliable. Their drawbacks are that they cannot be inspected in service, they block, and the head loss they need to be insensitive is permanently lost from the hydraulic profile. Where a works has limited fall, the usual compromise is weirs with a well-designed approach; where fall is available, orifice splits at 300 mm or more of head are nearly immune to construction tolerance.

Whatever the device, the chamber needs the means to take a unit out of service without upsetting the others: an isolating penstock or stop logs on each outlet, and weir lengths or orifice areas that can be re-proportioned when the number of units in service changes. A four-way split that becomes a three-way split by closing one gate will overload the remaining three by 33 percent, which is the design intent, but the chamber must also keep splitting the remaining flow equally with the approach pattern changed.

Verification: measuring the split and using CFD

Measure the split at commissioning and after any change of flow regime. Methods, from simplest to most precise: head over each weir with a point gauge (flow follows from the weir equation, so a 5 percent flow imbalance shows as a 3.3 percent head difference at the same crest level); tracer or dye timing in each outlet channel; portable area-velocity meters in the outlet channels; or, where the downstream units have their own measurement, the sum check against the inlet flume. A chamber that cannot be measured cannot be adjusted.

For new chambers of any size, or where an existing chamber is to be modified with baffles, a computational fluid dynamics model resolves the inlet jet, the surface profile and the effect of baffle porosity before concrete is poured. Modelling effort is small compared with the cost of a tank that never receives its share of the flow, and the same model answers the related inlet-energy questions in the clarifier inlet design guide. Specialist CFD modelling of treatment hydraulics is the right tool when the geometry is unusual or the flow range is wide. The wider hydraulic profile that sets the head available for splitting is in the flow equalisation guide and the weir hydraulics in our weir flow calculator.

How to design a flow distribution chamber

1. Fix the head budget. From the hydraulic profile, allocate at least 150 mm of weir head at minimum flow, or 300 mm for orifice splits.
2. Choose a symmetrical geometry. Central upward inlet with perimeter weirs, or a mirrored manifold; avoid sequential offtakes along a channel.
3. Kill the inlet momentum. Inlet velocity below 1 m/s into the chamber, a perforated baffle of about 40 percent open area across the full section, and a stilling length or ramp before the weirs.
4. Make the crests adjustable. Stainless weir plates on slotted fixings, levelled by survey to within 1 mm after the chamber is filled.
5. Provide isolation and re-proportioning. Penstocks or stop logs on each outlet, and weir lengths or orifices that can be changed when units are taken out of service.
6. Measure the split at commissioning. Point-gauge heads or outlet flows across the full flow range, and re-level if any outlet deviates by more than 5 percent.