Trier’s wastewater plant uses a model of its biological treatment process to select aeration settings. Supplier Xylem dates commissioning of the described system to November 2017, after training and trial operation.[1]
The oxygen requirement keeps changing
Biological treatment needs oxygen to support the removal of organic pollution and ammonia. The US Environmental Protection Agency explains that demand changes during the day and with the composition of arriving wastewater. Rainwater can dilute the incoming pollution, while a discharge rich in ammonia can increase the work required. Water volume alone therefore cannot describe the aeration task.[2]
Supplying excess oxygen consumes unnecessary energy and can disrupt treatment; too little can impair the biological process. Conventional automatic controls already use oxygen probes to adjust blowers and valves. The target is enough oxygen for the changing treatment load, with equipment responding as conditions change.[2]
Forecasting the next settings
Xylem’s neural networks use measurements from the existing plant-control system to model pollutant removal. The resulting digital model compares possible operating scenarios and determines aeration setpoints for six parallel biological tanks. It also supports decisions about chemicals used to remove phosphorus.[1]
What the aeration case reported
| Measure | Reported result |
|---|---|
| Aeration electricity | More than 20% reduction; approximately 200,000 kWh saved annuallySource 1 |
| Assessment | Early-2018 comparison with prior operation; effluent monitoring limits maintainedSource 1 |
A fair comparison includes the work done
Consider two hypothetical operating days. Removing 1,000 kilograms of a measured pollution load with 800 kWh takes 0.80 kWh per kilogram. Removing 1,200 kilograms with 900 kWh takes 0.75. Electricity use has risen, yet energy per kilogram removed has fallen by 6.25%. These invented figures illustrate why the treatment task belongs beside the energy reading.
The treatment plant also produces energy
Trier’s own June 2024 account describes a wider sequence of improvements. Equipment and process upgrades during 2013–2016 had already lowered annual plant electricity demand from about four million to 3.2 million kWh. That reduction preceded the AI biological-treatment system described in the account.[3]
The utility also installed generation and storage: sewage-gas combined heat-and-power units, rooftop solar, hydropower, hot-water buffers and gas storage. Its 2024 report says AI balanced electricity generation and consumption in real time, and the site supplied the neighbouring energy and technology park as well as the treatment plant.[3]
These are complementary engineering tasks. Better treatment control reduces the energy needed for a given job; generation and storage determine how the remaining demand is supplied. Together they make the plant’s operating choices more useful than an isolated electricity-saving percentage.
Sources & further reading
- Stadtwerke Trier: optimization of wastewater treatment plant processes thanks to AI Xylem · Accessed 12 Oct 2026
- Evaluation of Energy Conservation Measures for Wastewater Treatment Facilities US Environmental Protection Agency · Accessed 12 Oct 2026
- Rathaus Zeitung, 11 June 2024: Klimaschutz aktuell, page 5 Stadtwerke Trier, in the City of Trier newspaper · 11 Jun 2024 · Accessed 12 Oct 2026