The Plant Was Three Weeks Old and It Failed Its First Test

A plant we had handed over failed its first compliance test. Three weeks old. Everything in it was new — new blowers, new diffusers, new panel, new concrete. The lab sheet came back with ammonia well over the limit for the class of water it discharges to, and BOD sitting just under. The owner read that sheet the way anyone would read it: the plant does not work.

I want to write down what we actually found that week, because this is one of the few situations in wastewater where a failed result is not evidence of a fault.

What we found on site

Nothing mechanical. Air was going in evenly across the tank, the clarifier weir was level, the return line was pumping, the panel was reading what it should. I did the thing I do before I read anything else — I smelled the plant. It smelled of very little. Not septic, not sour. Thin.

Then we took a one-litre sample out of the aeration tank and let it stand. We read a cone rather than chasing MLSS, and on a working plant we want to see something in the range of 50 to 200 mL/L settling out of that litre. This one gave us a thin, pale layer well under that. There was almost nothing in the tank to do the work.

That is the whole story. The plant had been energised, not commissioned. Those are two different dates and on most projects only one of them is written into the contract.

The two speeds, again

A sewage treatment plant is a building that grows a culture. The steel and the concrete are finished on the day the contractor leaves. The culture is not.

The organisms that eat BOD grow quickly — they double in a matter of hours, and given food and air they build up a working population in days. The organisms that convert ammonia grow slowly, roughly a day or two to double, and they slow down further when it is cooler or when alkalinity runs short. We have written before about why ammonia is the parameter that behaves differently from everything else on the sheet.

So a brand new plant recovers its BOD first and its ammonia last, in exactly the same order and for exactly the same reason as a plant that has just lost its air for three days. In both cases the pattern on the lab sheet — BOD respectable, ammonia far out — is not a design symptom. It is a clock.

On this job that clock had a date attached to it. The first sampling had been scheduled by the building's own consultant for two weeks after energising, because that was when the occupancy paperwork was due. Two weeks is enough time to grow a BOD population. It is not enough time to grow a nitrifying one. The result was decided before the sample was taken.

What we did

We seeded it. Return sludge from a plant of ours that was running well, moved across and put into the tank, which gets you a starting population that has already been living on similar water. We keep a considered view on bought cultures — there is a narrow window at start-up where they are worth the money, and a much wider set of situations where people buy them because a plant is failing for a reason no bacteria can fix.

Then we fed it, which was the harder half. The building was at something like a fifth of its design occupancy. A plant designed for 150 cubic metres a day receiving 30 is not comfortably under-loaded; it is starving, and a starving tank will not hold a population no matter how carefully it was seeded. So we ran the plant small on purpose — one train, one tank in service, the rest isolated — so that what little load arrived was concentrated enough to keep something alive. A small plant is harder to run than a large one, and a new building temporarily makes a large plant behave like a small one.

We held the flow steady through equalisation rather than letting the morning peak wash the tank out, we stopped wasting sludge entirely, and we left the chemicals alone. The one thing we did watch was alkalinity, because nitrification consumes it, and on a plant that has just started to nitrify the pH can slide at the exact moment the population is finally establishing. That is a bad week to lose it.

The ammonia came right in the fifth week. Nobody had changed anything about the plant since the first test.

The part that is a design decision, not an operations one

What I take from this job is not about bacteria. It is about which dates get written down.

The acceptance test date on a sewage treatment plant contract is a technical decision that is almost always made by someone treating it as an administrative one. If the first compliance sample is scheduled two weeks after start-up, and the receiving water class carries a tight ammonia limit, the plant will fail and the failure will be attributed to the design. We have picked up work that started exactly this way — a plant condemned as wrong when it had simply not been given time to become a plant yet. Telling those two apart is the first thing we do on a failing-plant enquiry.

Where the class is tight, the arithmetic is worth doing before signing rather than after. Ammonia as NH3-N is 4 mg/L for a Class C receiving water and 3 mg/L for Class B under the general effluent standards as amended by DAO 2021-19, and those are the numbers a young population misses by a wide margin. The limits by class are set out on our compliance page, and which parameters actually matter depends on what your establishment discharges. If you want a feel for what a plant sized for your flow and your receiving class looks like before you talk to anyone, the estimator and the design generator will both give you something in a few minutes.

The commissioning period also belongs to somebody, and the contract should say who. On our own jobs we treat it as ours — on one plant we effectively built twice, we did not raise another invoice until it passed. That is a position, not an industry standard, and it is worth asking about explicitly. It is one of the questions I would ask before signing.

And the population you spent five weeks growing is the same population a new operator can lose in a weekend. Turnover shows up in the lab sheet, and the fastest way to lose nitrifiers is not to fail to grow them but to pump them out of the building. Handover is where operation starts, not where the project ends. You can see the kinds of plants we have taken through this on our completed projects, alongside the broader picture of how we design and build and what we run here in Cebu.

Where this stops being transferable

Five weeks was this plant, on this water, in this climate. Cebu is warm and that helps — the same start-up somewhere cooler takes longer, sometimes considerably. A plant carrying FBBR-style fixed-film media behaves differently again, because the media holds a population that a suspended-growth tank would have washed out, so it can start slower and then hold on better. A plant fed by a building that fills quickly has a much easier first two months than one fed by a building that fills over a year.

And the honest limit of this story: a young plant and a badly designed plant produce the same first lab sheet. The only thing that separates them is what you find when you look in the tank, and how the numbers move over the following weeks. If yours has failed a test and you cannot tell which of the two you have, that is a reasonable thing to want a second opinion on — it is most of what we do on our operation and maintenance support side, and it is the same conversation whether you are still choosing a plant or already living with one.

If any of this sounds like your plant, book a free 15-minute consultation and tell us what the sheet said and how old the plant is. Those two facts usually settle it.