Today
Aerating the activated sludge tanks often takes half or more of a treatment plant's power. Its setpoints were set at commissioning and are rarely checked again, although the load varies widely over the day, the week and the seasons. At night and at weekends the blowers often blow in as much air as at peak time. The data for this sits in the control system, but nobody has the time to compare months of trends with the effluent values.
An example
Aldevo Wassertechnik runs a municipal treatment plant for 40,000 population equivalents under an operating contract. The agent analyses three months of control system data: between 11 pm and 5 am the ammonium at the outlet of the aeration tanks is below 0.3 mg/l, yet three blowers keep the oxygen at 2.0 mg/l at night as well. It proposes lowering the night setpoint to 1.2 mg/l and coupling the daytime setpoint to the ammonium value, and expects about 70,000 kWh less power a year, with the effluent staying well clear of the 10 mg/l ammonium nitrogen limit. The plant manager approves the night curve for a four-week trial; afterwards the agent compares effluent values and power use.
Your data, wherever it lives
Your treatment plant's data lives today in the control system and its historian, power values in the meters and their exports, lab values and the operating log in spreadsheets, the discharge permit and blower data sheets as PDFs. demi reads, understands and links them per tank and hour, with your rights. The AI finds the hours with too much air in minutes instead of in a yearly review, and plant management decides on new setpoints.
Where the data lives today
- Machines & sensors
- Databases
- Office files
- PDFs
How demi solves it
Three layers in this order: connect the data first, then the rights, then the AI.
- 1 · ConnectData Layer
Data sources in this case: the control system and its historian with oxygen, ammonium, nitrate, inflow, air flow, valve positions and blower power, plus power meters, lab values and the operating log. demi replaces nothing and puts everything on one timeline per tank and hour, together with the effluent limits from the discharge permit.
- 2 · PermitPermission Layer
Operators and energy management see the plant data; the lab releases lab values. Only the authorised person changes setpoints in the control system; demi writes nothing to the controls. Every proposal and every approval is recorded.
- 3 · ActIntelligence Layer
The agent compares air flow and oxygen with the actual load from ammonium and inflow and finds the hours in which more air is blown in than the biomass needs. It proposes setpoints per time of day or a coupling to the ammonium value, works out the power saving from the meter data and checks every proposal against past effluent values and the permit limits. The numbers come from the measurements; the language model only explains the proposal.
How it runs
- 1
demi reads the trends of oxygen, ammonium, nitrate, inflow, air flow and blower power from the control system.
- 2
The agent lines them up with power meters, lab values and the operating log, per tank and hour.
- 3
It shows when aeration delivers more air than the load needs and how much power that costs.
- 4
From this it proposes new setpoints, checked against past effluent values and the permit limits.
- 5
The plant manager approves the change; the agent then follows effluent and power use.
What changes
Less power for aeration
Setpoints that match the actual load
Effluent values watched with every change
What does it bring you?
Check it with your own numbers
Every starting value is an assumption you can change. The calculation sits right below, step by step.
Your numbers
Result
+81 hfreed per year
Assumptions and calculation
demi analyses oxygen, ammonium and blowers and proposes setpoints; you approve them.
45 min → 31.5 min per case
How we calculate
30 daily reviews × 45 min × 30% × 1.0 (manual) × 12 months = 4,860 min = 81 h
Value per year, before demi costs
Freed per year+81 h
Assumption: labour cost of €49.50 per hour in German manufacturing in 2025 (Destatis), plus workplace, IT and overheads.
= Value per year, before demi costs
+€5,265
gained per year
An estimate, not a promise. The starting values are assumptions for an industrial company with 150 to 500 people.
Combine with more cases and get the packageQuestions about this use case
Does demi intervene in the plant's controls?
No. demi reads the data from the control system and its historian and makes proposals. The authorised person changes setpoints in the control system, as today, and the change is recorded together with the proposal.
How does the agent know how much air the biomass needs?
From the interplay of ammonium, nitrate, oxygen, inflow and air flow over weeks and months. It compares hours with similar load and shows when the same treatment result was reached with less air. It uses lab values to check the online measurements.
What happens in winter, when nitrification slows down?
The agent takes the water temperature into account and proposes setpoints per season. Your plant management sets minimum oxygen values, and it proposes nothing below them. If ammonium rises after a change, it reports this at once and shows the previous setpoint.
Does this also work for an industrial treatment plant?
Yes. Food, chemical and paper plants with their own biological treatment see load peaks from production, for example after cleaning or product changes. demi also links the production plan, so that aeration matches the load that is coming.
