Building the Hydraulic Profile of a Wastewater Treatment Works: From Flood Level at the Outfall to the Inlet Works

What the profile has to answer

Three questions: the discharge level when the receiving water is high, the head lost through each unit and connection at the governing flow, and where each tank top, weir crest and channel soffit must therefore sit so that nothing overtops, no free weir drowns and no sewer surcharges. Answering them needs a single consistent datum (the Riverside master plan found its drawings were 67 mm off the newer topographic datum, enough to matter on a launder), a list of flow scenarios, and a loss coefficient for every element in the path.

Head in an open tank is, to within a velocity head, the water surface elevation; in a full pipe it is pressure plus elevation, and the water surface reappears wherever the pipe discharges into a chamber. Energy is lost by friction, at entries, exits, bends, contractions and expansions, through gates even when open, through screens and media, and over weirs. A weir behaves in one of two ways: with a free, aerated nappe the upstream level depends only on flow and geometry and the profile breaks there; with a drowned nappe the upstream level depends on the downstream level too and the chain is continuous.

Boundary condition: the outfall and the flood level

Start at the receiving water. For a river outfall the design tailwater is a statistical flood level, not the normal level: Riverside in California used the 50 year flood of the Santa Ana River; the Winnipeg South End works (SEWPCC) used a Red River flood elevation of 229.0 m against a 50 year level of 229.80 m and an outfall invert of 218.54 m, submerged most of the time. In England the Environment Agency expects new assets to be assessed at the 1 percent annual probability (1 in 100 year) fluvial level with the published climate change uplift, and the profile should show the works still draining at that level or state which flows must be pumped or held back when it is exceeded.

From that tailwater, add the outfall pipe losses at peak discharge: friction (Colebrook-White or Manning with a roughness that allows for slime), the exit loss (one velocity head into still water), bends and the entry loss. At SEWPCC a 1,800 mm final outfall and a twinned section were needed to keep the clarifier launders clear; the modelled launder level at 300 ML/d with the river at 229.0 m was 231.52 m, 0.52 m below the weir crest, where the earlier single pipe had surcharged the weirs in past floods. Non-return flap valves add 0.3 to 0.6 velocity heads when fully open.

Headloss allowances through each unit

Working upstream, each unit adds the loss through its outlet structure, its body and its inlet, plus the connecting channel or pipe. Reasonable allowances at peak flow, to be refined with manufacturer data:

Between units, channel losses are friction plus 0.1 to 1.0 velocity heads at each bend, gate frame, port and junction. SEWPCC, a 90 ML/d average works with screens, grit, primaries, BNR reactors, final clarifiers and UV, has 3.82 m between the screen channel invert and the floor of the chamber at the end of the UV structure, a fair figure for a conventional works without granular filters: 2.5 to 4 m. Sand filters need 2.5 to 3 m more or an interstage pumping station. Weir and flume equations for the individual structures are in the companion guide on weir hydraulics, and the screen allowances in the guide on inlet screening.

Design flows and recycle streams

The profile is checked at several flows, and the governing one differs along the works. Upstream of the storm split the full arrival flow applies: Formula A, the inlet pump capacity or the sewer model peak; SEWPCC used 415 ML/d of influent pumping against a 90 ML/d average, a ratio of 4.6, and Riverside used a peak factor of 2.2 upstream of its equalisation basins and 1.5 downstream. From the storm split to the final tanks, flow to full treatment applies, but with the recycles added: return activated sludge at up to 100 percent of average flow passes through the aeration tanks and final tank inlets (Riverside set it at one times average flow), filter backwash and filtrate return (3.5 ML/d at Riverside), sludge liquors, and any storm tank return. Models add and extract these at the right nodes; a hand profile must do the same, or the final tank inlet is sized 30 to 50 percent light.

At the minimum night flow (SEWPCC checked 20 ML/d against 415 ML/d peak) channel velocities fall below the 0.3 to 0.6 m/s that keeps solids moving, and long weirs run at a few millimetres of head where levelling tolerance controls the split.

Setting levels: weirs, freeboard and the chain of control

With the downstream level known at each unit, set the structure that controls the upstream level. For a free weir, the crest must sit above the highest downstream water level by at least 50 mm (sharp-crested) or by enough to keep a flume or broad-crested weir below its modular limit, and the upstream water level is then the crest plus the head at peak flow. For a drowned weir or a submerged port, the upstream level is the downstream level plus the loss. Then add freeboard: the Riverside study used 150 mm as its criterion for identifying bottlenecks; UK practice is commonly 300 mm on open channels and tank walls at the peak storm case and 150 mm on launders, with every chamber rim checked against the modelled level.

Two checks close the exercise. First, repeat it at the normal river level: a free weir set for the flood case may now drop 1.5 m into the launder, harmless for effluent but on a primary tank a cause of air entrainment and floc breakage, so stepped or submerged outlets may be needed. Second, repeat it with one unit of each type out of service, since the remaining units run at higher flow and head.

Models, bottlenecks and retrofits

Both published studies used hydraulic software and calibrated it against measured levels first. SEWPCC ran steady-state cases at 55, 200, 300 and 422 ML/d with and without the flood tailwater, since a surface wave crosses a works far faster than sewer flow changes. Riverside found no bottlenecks at average flow but four at peak wet weather: a 42 inch interconnecting pipe that submerged the aeration basin and primary clarifier weirs (replaced by 54 inch), a distribution channel, and two 24 inch headworks control valves, each fixed with a parallel pipe or an extra valve rather than by re-levelling tanks. That is the usual pattern on an existing works: the tanks were set with enough fall, and the losses that grew with time are in the connections, the gates and the screens.

For a retrofit, the first task is a level survey of weir crests, channel inverts and tank rims against one datum, followed by logged water levels at a known flow; the gap between measured and calculated losses is usually a drowned weir, a part-closed penstock or a fouled screen. Where a new tertiary stage must fit an existing profile, the head between the final tank effluent channel and the outfall at flood level decides whether it can be gravity fed: cloth filters and ballasted clarifiers at 0.3 to 0.6 m often can, sand filters at 2.5 m rarely can. Where the geometry is awkward, a physical or CFD model is cheap compared with the concrete it protects; the Wigan inlet works model confirmed the detritor split within 5 percent before construction. Our process design service builds the profile as part of every works design, and CFD modelling resolves the structures that hand calculation cannot.

How to build a treatment works hydraulic profile

1. Fix the datum and the tailwater. One ordnance datum for all drawings; the 1 in 100 year river level with climate allowance, or the tidal extreme, as the outfall boundary.
2. List the flow cases. Full storm flow to the storm split, flow to full treatment plus recycles beyond it, minimum night flow, and the n minus 1 unit case.
3. Work upstream unit by unit. Outfall pipe losses, then each unit outlet, body, inlet and connection, using manufacturer maxima for screens and filters.
4. Set each control. Free weirs 50 mm or more above the highest tailwater and within the modular limit; drowned connections as downstream level plus loss.
5. Add freeboard and check rims. 300 mm on channels and tank walls, 150 mm on launders, every chamber rim above its modelled level.
6. Repeat at normal river level and low flow. Confirm no excessive drops, self-cleansing velocities and workable weir heads; then calibrate against survey on an existing works.