The revised Urban Wastewater Treatment Directive expands secondary treatment obligations to agglomerations of at least 1,000 PE by 2035 and introduces a pathway towards energy-neutral wastewater treatment for larger plants. This creates a regulatory and cost pressure on thousands of small and medium municipal wastewater treatment plants that are often manually operated, poorly instrumented, and without a permanent on-site operator.

Aeration is the main controllable energy consumer, typically representing 60–70% of plant electricity use, but current AI/advanced control retrofit solutions are usually engineered plant by plant and assume online probes, historical data and operator support. This does not scale to heterogeneous fleets of small SBR, package, ditch or MBBR plants that lack NH₄/NO₃ sensors, calibrated process models and sufficient data history.

Utilities therefore need a low-cost retrofit approach that can reduce aeration energy while maintaining effluent compliance, support remote operation of many plants as one fleet, and produce auditable UWWTD-oriented reporting without adding costly instrumentation to every site.

AQUASCADA will combine low-cost edge hardware, fleet-level remote supervision and advanced control software to optimise aeration across heterogeneous wastewater treatment plants. Instead of relying on expensive online nutrient probes or individually calibrated activated sludge models, the system will infer NH₄-N and NO₃-N through label-free soft-sensing based on affordable operational signals and limited grab samples.

The control approach will be model-free and self-calibrating, with safe operating envelopes that constrain autonomous actions and automatically switch back to baseline control in case of anomalies. Cross-plant transfer and cold-start learning will allow knowledge from existing plants to be reused when onboarding new sites, reducing the need for months of per-plant engineering.

The same operational data will feed auditable UWWTD compliance and energy reporting. The intended innovation is not generic AI aeration control, which already exists for instrumented plants, but scalable, safe fleet autonomy for small, under-instrumented municipal plants where current retrofit solutions are too site-specific and costly.

The consortium will define pilot baselines, data requirements and safety constraints for small and medium wastewater treatment plants, focusing on aeration energy, effluent compliance and available low-cost signals such as blower power, flow, temperature and conductivity. Pilot operation will be narrowly focused on biological reactor aeration and clearly separated from unrelated telemetry, odour or sewer-network applications.

The technical work will develop and integrate label-free NH₄-N/NO₃-N soft-sensing, self-supervised and physics-informed learning from low-cost operational data, model-free aeration control, cross-plant transfer mechanisms and edge-based fail-safe execution. The system will include anomaly detection, automatic fallback to baseline control and operator override to ensure that energy optimisation does not compromise permit compliance.

The partners will validate the retrofit kit on pilot plants against measured fixed-DO or timer-based baselines, targeting reduced aeration energy per PE, compliant effluent quality, fast cold-start deployment on new plants and remote supervision of a fleet by a limited number of operators. They will also implement compliance-by-design reporting workflows aligned with UWWTD requirements and national transpositions, and benchmark the results against existing retrofit aeration control solutions.