The Clarifier Return Needed No Pump. The Upflow Number I First Wrote Was Still Wrong.

Every clarifier in a small sewage treatment plant has the same awkward problem at the bottom of it. Solids settle into the sump, and from there they have to go two places. Most of them have to go back to the aeration tank, because that sludge is the biomass doing the treatment and a plant that does not return it has no solids retention worth the name. A small share of them has to leave the plant altogether, because sludge is produced faster than it decays. Both of those movements are usually solved the same way: put a pump in the sump and switch it on.

On a recent design I took the pump out and let the whole thing run on gravity. The return and the sludge draw now share one pipe off the clarifier cone, and the plant has one pump in it from end to end. I am writing it up because the arrangement worked and because the first number I calculated for it was wrong, on a basis that looked perfectly reasonable until somebody made me say out loud what was actually rising through the tank.

This is not the chamber post. Our standard underground clarifier detail is a rectangular chamber with a concrete hopper and a submersible sludge pump sitting in that hopper, and that detail is still what we pour on most jobs. The carryover post is about reading a clarifier that is throwing solids over the weir. This one is narrower than either: it is about the line that leaves the bottom, what decides whether that line needs a machine on it, and the arithmetic of what actually climbs to the weir once you have a return on it.

Whatever leaves the bottom either gets lifted or gets to fall

The reason the sump usually gets a pump is elevation. In a buried plant the clarifier is normally the last chamber in the train and its floor is the lowest point in the works, so the aeration tank you are returning sludge to is above you. Sludge that has to go up needs something to push it.

On this job the levels happened not to require that. The clarifier cone sat high enough relative to the chamber the return was going to that a line off the hopper would run downhill the whole way. That is a site accident, not a design philosophy, and I want to be clear about which it is. The design decision was only to notice it and use it instead of drawing the pump in out of habit.

Once the line is gravity, the rate is set by the pipe and the head available rather than by a duty you select. That is the opposite of comfortable. You cannot dial a gravity return up and down from a panel. What you can do is size the pipe for the rate you want and then check what that rate does to everything downstream of it, which is where this stopped being a plumbing decision.

One line, two jobs

The return runs at about five times the forward flow. That is high by the usual habit of two to four, and it is deliberate: the same line is carrying the returned sludge and the nitrate return, so the recycle has to be enough to keep the anoxic zone fed with nitrate as well as keep the clarifier hopper from building a blanket. The carbon arithmetic for denitrification is a separate check and it does not change here; what changes is that one pipe is doing a job two pipes usually share.

At that rate the volume sitting in the cone turns over roughly every forty minutes. That number is the one I care about most, because a clarifier hopper is a place where settled sludge goes septic if it sits. Sludge that goes septic in the sump releases phosphate back into the water, which is exactly the mechanism behind a phosphate result that passes the clarifier and fails at the chlorine tank. A forty-minute turnover is not a target I set out to hit. It is a consequence of the return rate that I then checked for, and had it come back at four hours I would have drawn it differently.

The same line is also the sludge draw. There is no separate waste-sludge pump, no separate waste line, and no separate valve chamber. Wasting is done by taking from that line, which means the operator is drawing off a stream that is already moving rather than starting a machine that spends most of its life switched off. One less pump involved is the whole commercial argument, and it is a smaller argument than it looks: the saving is not the purchase price, it is that every pump on a small plant is an after-sales decision before it is a technical one, and a plant with one pump in it has one thing to re-specify in three years instead of three.

The crest loop, because a buried line will quietly empty a clarifier

A gravity line out of the bottom of a full tank, running downhill to a lower chamber, is a siphon waiting for somebody to start it. Once it primes, it does not stop at the rate you sized it for. It keeps pulling until it draws air, and the thing it is pulling from is the clarifier.

So the line does not run continuously downhill. It rises to a crest above top water level and discharges into open air into the receiving chamber, and that break is what makes the arrangement safe. Air gets in at the crest, the line cannot hold a continuous column of water from one end to the other, and the flow is governed by the head on the hopper side rather than by the drop to the far end. It is a plumber's detail rather than a process one, and it is the single item in the arrangement I would never leave off a drawing. A buried siphon that empties a settling tank at four in the morning is not a fault anybody is going to diagnose from the lab sheet.

The crest also has to be where somebody can see it. An open discharge is a flow indication you get for nothing — you can stand at the chamber and watch the return running, which on a plant with no instrumentation is worth more than a gauge nobody calibrates. That is the same reasoning as reading a settling cone instead of waiting for a solids figure: a thing you can observe on site beats a number that arrives late.

Then I wrote the upflow rate on the wrong basis

The transfer into the clarifier on this plant is not continuous. It arrives in bursts, because the chamber ahead of it discharges on level control. So the question is not the average upflow rate, it is the upflow rate during a burst, and that is the number that decides whether solids get carried over the weir while the burst is running.

I calculated it the obvious way: pump rate divided by the clarifier surface area. That gave about two metres per hour, which is high but not absurd for a short burst, and I wrote it into the design basis and moved on.

It is the wrong basis, and the reason is sitting in the paragraphs above. The recycle leaves down the hopper. Water entering the clarifier does not all have to climb to the weir — only the net forward flow does, which is what comes in minus what goes back out of the bottom. On this geometry the honest burst figure was about 1.26 metres per hour, not 2.0. The tank was never as close to its limit as my own number said it was.

I would rather publish that than not, because the error is the kind that survives review. Two metres per hour is conservative, it reads as careful, and nobody checking the document would have flagged it. A wrong number that errs towards caution is still a wrong number, and it was about to size a chamber bigger than the water required. It is the same lesson as the day a gravity recycle I had sketched turned out to need far more driving head than the site had, once anybody computed it instead of looking at it. A glance is not a method, and an arithmetic habit is not the same thing as arithmetic.

What the honest number changed, and the ball valve on the transfer line

Knowing the real figure did not make the problem disappear. It relocated it. If the burst upflow sits near 1.2 metres per hour at the designed transfer rate, then it sits near 2.3 if that transfer pump is allowed to run at its full curve, which is roughly what an off-the-shelf pump will do on a short line with very little head against it. The pump you can buy is almost always bigger than the duty you want, which is a problem we have met before at the opposite end of the scale, with a dosing pump twenty-two times the required dose.

The remedy is a ball valve on the transfer line, set at commissioning and then left alone, throttling the pump back to the rate the clarifier was designed around. That is an unglamorous answer and I have a mild suspicion of it, because a valve set by hand is a setting somebody can move. It earns its place here for the reason an adjustable weir plate inside a buried tank does not: it is one of the few things on a plant that is still reachable once the cover slab is on. An adjustment nobody can get to is not an adjustment. An adjustment at eye level is fine.

It also means the commissioning record has to state the valve position and the rate it was set to. If that is not written down, the first person to service the plant restores it to fully open in good faith and the clarifier starts carrying solids over on every burst, for a reason nobody will connect to a valve. Turnover is a design input, not an operations excuse.

Drawing sludge by the bucket is a solids decision, not a volume one

The wasting routine on this plant is small enough to sound unserious. It is a draw off the return line, a handful of litres at a time, on a frequency set by what the settling cone is reading rather than by a calendar. People expect a sludge draw to be a tanker event. Most of the time, on a plant this size, it is not.

The reason to waste is not that the tank is full. It is to hold the solids concentration where the clarifier can still cope with it, because the limit that gets hit first is the suspended solids leaving over the weir, not the capacity of the tank holding them. Waste a little, often, from a line that is already flowing, and the concentration stays inside the band the settling arithmetic assumed. Waste nothing for six weeks and then a great deal at once and the plant loses treatment for days afterwards.

Which is also why the destination still has to be written down, however small the volumes are. No plant is zero-sludge, and everything you take out of the plant has to go somewhere — a lidded container at a reachable point, hardstanding within hose reach, and a named route off site. A bucket-scale draw with no destination is just a bucket somebody leaves beside the tank.

Where a one-pump plant is the wrong answer

I would not generalise this arrangement, and I want to say plainly where it fails.

If the clarifier floor is below the chamber you are returning to, there is no gravity return, and that is the ordinary case rather than the exception. If the plant has any reasonable flow to it, five times forward flow down a single pipe stops being a modest pipe and the crest detail stops being trivial. If the return and the waste have genuinely different destinations — a thickener, a drying bed, a tanker point at a distance — then one line serving both is false economy and you will end up valving between them anyway. And if the site needs the return rate to be adjustable because the influent swings hard, a gravity line is the wrong instrument: you cannot tune it from a panel, and on a plant where the biology is marginal that is a real loss. Solids retention is the thing that decides whether nitrification is possible at all, and anything that makes solids retention harder to control deserves a pump and a control loop, not a clever pipe.

The honest summary is that a single pump is not the objective. The objective is that every item in the plant has a reason you can state, and on this job the pump in the clarifier sump did not have one. On the next job it probably will, and the standard detail with the submersible in the hopper goes back on the drawing without an argument. The test is not how few machines a plant has; it is whether every line on the drawing is there for a result somebody asked for.

Where this stops being transferable

All of the above is from one plant, on one set of levels, at one scale, designed for a commercial building here in Cebu where we pour our own tanks and know what the concrete costs within a narrow band. The part that travels is not the layout. It is the two checks: find out what is actually rising through your settling tank rather than what your pump is delivering into it, and find out how often the cone turns over rather than assuming the sludge is leaving because a pipe exists.

Both of those are arithmetic anybody can do on their own plant in an afternoon with the drawings and a tape measure. Neither needs a laboratory, which is the reason I trust them. If somebody re-does the first one on a plant where the return is small relative to the forward flow, they will find the correction barely matters and my 2.0-versus-1.26 distinction is noise. On a plant returning five times the forward flow it is most of the answer. That is how these things go: the same check is decisive in one place and irrelevant a hundred metres away, which is why we look at the site and the numbers rather than carrying a conclusion from the last job into the next one. Reading somebody else's finished drawing is mostly this exercise, done on their geometry instead of ours.

If you want to see the shape of a plant for your own building before any of this detail matters, the STP price estimator and the wastewater design generator will give you tank volumes and a process order from the flow. What the limits actually are is on our DENR discharge standards page, and which parameters your industry is held to is on the significant parameters page. For a plant already built and running, maintenance and operation and our O&M support are the better starting points, and ammonia, nitrate and phosphate each have a page of their own, with the longer nutrient field guide behind them. If you are starting from nothing, getting an STP sets out how we begin, our wastewater treatment services and our work in Cebu cover what we do, what to look for in a wastewater treatment provider in Cebu is worth ten minutes first, and the plants themselves are on completed projects.

If you have a clarifier on a drawing and nobody can tell you what rises through it during a burst, send it over and we will work it out with you. Book a free 15-minute consultation.