Fourteen of the Sixteen Air Pumps Were Dead. Fixing Them Would Not Have Made the Plant Comply.
We were asked to look at an existing plant that had never passed an ammonia test. The first number I wrote down on that site was a surprise: the property had more aeration capacity installed than the process needed. Sixteen small air pumps, adding up to roughly 240 cubic metres of air an hour against a requirement of about 175. On paper that is 137 per cent of the air demand, sitting there, already bought and already wired.
Fourteen of those sixteen pumps were not running. The relays had failed, one at a time, over a period nobody could date, and the plant had been operating on something between fourteen and seventeen per cent of its installed air. Nobody on site knew that mattered, because nobody had been told that the thing clicking in the panel was the thing keeping the bacteria alive.
That is a satisfying story to tell and it has an obvious ending: replace fourteen relays, the air comes back, the ammonia passes. I want to write about why that ending is wrong, because the arithmetic we did next is the part that actually earns its place here. We have written before about which machine to specify on a small plant and why noise is a design input, and about a plant that lost its air for three days and how it came back. This is neither of those. This is a plant that had the air and could not use it, and would still not have complied on the day every relay was replaced.
The cheap fix was real, and we named it ourselves
I do not want to be clever about the relays. They were a genuine fault and they were the cheapest fault on the site. A relay is one of the least expensive items in a control panel and one of the most consequential, which is a pattern we keep running into. On a different and much smaller plant years ago, an overload relay that was never fitted was the whole reason that plant kept losing its nitrifiers — one pump, one missing component, biomass going over an overflow every few weeks. That was a plant losing its bacteria. This was a plant that could not grow them in the first place, and the difference is worth keeping straight.
So we said in writing that the relays should be replaced regardless of what else was decided, because a plant running at a sixth of its air is not a plant anyone can assess properly. You cannot diagnose a process while a basic utility is missing. Fix the known fault, then look at the result.
Then we looked at what the result would be, and it did not comply.
Four tanks, and nowhere for the solids to stay
The treatment line was four tanks of about 33 cubic metres each — call it 132 cubic metres of process volume — receiving around 100 cubic metres a day. There was no clarifier, and there was no sludge return.
That second sentence is the whole problem, and it is quiet enough to miss on a drawing. Without a clarifier returning settled solids to the front of the process, the bacteria in the tank have exactly the same residence time as the water. Solids retention time equals hydraulic retention time. In this case 132 divided by 100 is about 1.3 days, and that is how long a given bacterium stays in the plant before it leaves with the effluent.
For the organisms that remove organic load, 1.3 days is survivable. They double in hours. For the organisms that oxidise ammonia it is not. At the temperatures we work at in the Philippines, roughly 30 degrees, nitrifiers wash out of a system somewhere around a solids retention time of one to one and a bit days — that is the point at which they are leaving faster than they can reproduce. The plant was sitting a fraction above its own washout line. To design for nitrification with any safety factor at all you want something near three days as an absolute floor, and for a result you would actually put your name to, eight to ten.
So the honest finding was this: put full air into that plant and you will fix the organic load, the smell will improve, and the ammonia will still fail, because there is no mechanism in the plant for keeping the slow population long enough to build it. That is not an aeration problem wearing an aeration problem's clothes. It is a solids retention problem that was hidden behind an aeration problem.
Why dosing bacteria into it cannot work either
The obvious next suggestion, and one that had already been made to the owner, is to buy bacteria. We sell biological products ourselves and we have written plainly about when we tell clients not to buy them. This site is the clearest case of that I have had to explain.
Dose whatever you like into a plant with a 1.3-day solids retention time and it leaves in 1.3 days. You are not seeding a population, you are paying to put bacteria through a pipe. The product is not at fault and the dose is not too small; the plant simply has no way to hold what you put in it. Monthly dosing on a site like that is a permanent bill with a predictable result, and the honest answer to the owner is that no consumable reaches this.
That reasoning is arithmetic, not opinion, and it is the sort of thing worth asking any supplier to show you. If someone proposes a recurring dose, ask what the solids retention time is and how the plant keeps the organisms it is being sold.
The air was also weaker than its nameplate, and that is a margin problem
There was a second finding inside the aeration itself, and it is the kind of detail that only turns up when you check the duty point rather than the rating.
The diffusers sat about 1.7 metres below water level. That submergence is roughly 0.2 bar of back pressure, and the pumps on site were rated at 0.2 bar. In other words the machines were being asked to work at the exact pressure their published output assumes, with nothing in hand. A diffuser does not stay clean, and as it fouls the back pressure rises; a machine with no margin loses output immediately, while a machine chosen with some headroom absorbs it. The vulnerability there is not capacity. It is margin. The plant's installed figure of 137 per cent is a nameplate number, and the delivered number on a fouled membrane is not something I would want to guess at.
This is why we check what a machine does at its actual duty point rather than what the data sheet says in free air, and it is part of the same habit as asking for an actual influent characterisation for the industry in question instead of accepting a round number.
Sixteen machines is an operating cost argument, not a capacity one
The case for replacing sixteen small pumps with one or two properly sized machines is not that the sixteen cannot deliver the air. They can, on their nameplates. The case is electricity and maintenance.
A row of small machines doing the work of one correctly sized one draws substantially more power for the same oxygen delivered — in the order of three times, on the comparison we did for that site. I am deliberately not publishing a peso figure, because the number belongs to a job that is still live and because the shape of the argument is what transfers. The other half is attention: sixteen machines is sixteen sets of filters, sixteen things to fail, sixteen relays, and no redundancy at all, because when one dies that chamber has nothing. One duty machine and one standby is less equipment, less maintenance and more resilience, which is the same after-sales reasoning we apply to every item we specify.
The honest way to put that to an owner is as cost per cubic metre of water actually treated, not as a capacity upgrade. Nobody should be asked to pay for capacity they already have.
What the plant actually needed
Three things, in this order, and only the first is cheap.
Relays and the air back, because you cannot assess anything without it. Then a way to retain solids, which on a site with four existing tanks and no space to add a fifth means either a clarifier with a return — and we now build the same clarifier detail on every job, so that is a known quantity rather than a redesign — or FBBR-style fixed-film media in the aerated chambers, which holds the slow population on the media instead of relying on a return that does not exist. Media is usually the lighter intervention on a retrofit, because it does not need new civil works and it does not depend on a pump somebody has to keep alive.
Then the aeration machinery, argued on running cost.
That ordering matters commercially as well as technically. A retrofit and a new build are two different jobs, and on a retrofit the owner has already paid for a plant once. Telling him the whole thing has to go is easy and usually wrong. Telling him which of the things he already owns can stay is harder and takes arithmetic.
How we would have seen this earlier
Nothing in the paragraph above needed a laboratory. The finding that decided the case was two volumes and a division: process volume over daily flow. Tank dimensions do not move and daily flow can be established from a bill, so that number is available to anybody with a tape measure before a single sample is taken.
Equipment counts are the opposite. The pump count on that site was wrong in the information we were first given, and it took somebody standing in the plant room to settle it. That is the general lesson I now apply to any assessment of a built plant: anchor the finding on the thing that cannot be disputed. Volumes and geometry are arithmetic. Equipment, settings and histories are reports until verified.
On the day-to-day side, a plant with no sludge return will show you the same story in a one-litre cone: very little settles, because very little is being kept. The cone is the cheapest instrument on any site and it would have raised the question months earlier. So would a look at what the final chamber is handing forward. And where a site has no equalisation volume to flatten its peaks, a short retention time gets shorter for a few hours every day, which does not help.
Reading somebody else's plant is not the same as reading somebody else's drawing
We have written about reading a drawing somebody sends us before pricing it. A built plant is harder, because the drawing and the plant have usually diverged, and because the people who could explain the divergence have often moved on — which is its own recurring theme, since turnover in the operator's chair shows up later as a plant problem.
The practical consequence is that we treat the site visit as a gate rather than a courtesy. Fourteen dead relays do not appear on an as-built, and the absence of a sludge return is easy to overlook on a plan where four tanks and a pipe look like a process. Neither shows up on a laboratory certificate, which is the only document most owners have when they first call us.
What transfers from this, and what does not
The method transfers. Divide your process volume by your daily flow before you argue about machinery, and ask what mechanism keeps the slow bacteria in the plant. If the answer is the clarifier and the return pump, check they both work. If there is no clarifier and no return, your solids retention time is your hydraulic retention time and you can read your ammonia prospects straight off that number. And when you find a cheap fault, fix it and then re-ask the question rather than treating the fault as the answer.
The numbers do not transfer. The washout figure I used sits where it does because of temperature, and a cooler site behaves differently. The eight-to-ten-day target is what we design to, not a law. The back-pressure finding came out of one submergence depth and one class of machine; on a shallower tank it would not have arisen at all. And the conclusion that media was the lighter fix was specific to a site with four tanks already in the ground and no room for a fifth — somewhere with space, a clarifier and a return may well be the better long-term answer, and we would say so.
Most of all, this was one plant, assessed once, with a particular history behind it. I would not tell you your ammonia failure is a solids retention problem because this one was. I would tell you to do the division before you spend anything, because it costs nothing and it changes what you buy.
If you are working out what your own site needs from scratch, the STP price estimator and the wastewater design generator will get you to a sensible starting shape, and the DENR discharge standards page sets out what you will be held to by water body class. The ammonia, nitrate and phosphate field guide covers the site readings, with more detail on ammonia removal and nitrate removal specifically. If you have inherited a plant that already exists, operating and maintaining one and our O&M support are the relevant pages, and if you are starting a new one, getting an STP begins with a conversation about the building. You can see the kind of work this comes from on our completed projects page, on the Cebu wastewater treatment page, and on the general wastewater treatment page. Some of this also sits behind how we cost a job from rates we have actually paid, and what we do as an STP contractor in the Philippines.
If your plant has never passed ammonia and nobody has shown you the retention arithmetic, send me the tank dimensions and your water bill rather than the lab result. Book a free 15-minute consultation and we will do the division together.