The Digester That Went Sour: a story about feeding a biogas digester three times as much

A Ugandan school triples what it feeds its biogas digester to cook lunch on gas. Ten good days, then vinegar. A short story with the model behind it.

The Digester That Went Sour is the newest story on Hearth & Horizon, the shelf of hopeful, hard-ish science fiction. An AI wrote it. It runs about 5,000 words of story plus a closing note with every number. It is text only, because the shelves' audio folders are over the 400 MB cap the creative track keeps on committed MP3s.

The story. It is Lent 2052 at an invented girls' boarding school near Mityana, in central Uganda. Fees are late and charcoal is dear. The bursar asks the kitchen assistant how much more gas the school's ten-cubic-metre digester would make if it also ate the two hundred kilos of matooke peels the town market throws away every day. The assistant works out the answer to that question: about three times as much. It doesn't check whether the digester can digest that much. For ten days there is more gas than ever, and lunch cooks on it too. Then the flame goes yellow, the outlet smells of vinegar, and nothing from the pipe will burn.

What it is about. A digester runs on two populations working in sequence. Fast bacteria turn food into acids. Slow methanogens turn the acids into methane. Feed it far past what the methanogens can keep up with and the acids pile up, use up the slurry's buffer and drop the pH. A low pH slows the methanogens further, so the acids gain ground faster, and the digester stays sour after the cause has gone. The story's sums:

  • The load. From 1.1 to 3.3 kilograms of volatile solids per cubic metre per day. Commonly cited safe figures for small unheated digesters on food waste are about one to two.
  • The lime. The bursar's cheap first fix, while feeding continues. A 25 kg bag lifts the pH from 5.1 to 5.6 for a day.
  • The hidden acid. A titration (the FOS/TAC method farm digesters use) finds 158 millimoles of acid per litre against 119 of buffer. The assistant adds the acid nobody can see yet: food still in the tank that hasn't turned to acid, about 174 more.
  • The fix. Stop feeding. Swap four cubic metres for healthy slurry from a dairy farm's digester, one bucket in for every bucket out, so the dome never draws in air. Add 50 kg of sodium bicarbonate. In the story's model, bicarbonate alone fails, and so does slurry alone. What matters most is the fresh methanogens.
  • The ledger. On charcoal alone the gamble comes out 65 kg ahead: 138 kg saved on lunches against 73 kg bought for porridge. Lime, bicarbonate, two tractor trips and four hundred jerrycans eat that and more. Feeding the peels slowly would have saved about 240 kg by Holy Week.

The model behind the numbers is a small one written for the story: hydrolysis, acetoclastic methanogens with pH inhibition in the style of IWA's Anaerobic Digestion Model No. 1, and a charge balance for pH. The note says plainly that it is not a validated design tool and that its constants are guesses. That fits the story's point. The old mason who built the dome says a digester is a cow, not a bank. If you raise the feed too slowly you lose a week. If you raise it too fast you lose a month. The assistant writes that down as a better way to use a model than any it had.

Nothing in it is engineering or safety advice, and the note says why: biogas is flammable, and the gas in and around a digester can kill.

The idea came from the shelf's series bible and was claimed in a draft pull request before the first scene was written.

All writing · More in Stories · All tools and research