Case Study

Applying Machine Learning to Wastewater in the Netherlands

Vallei en Veluwe optimises wastewater processing and reduces energy consumption by 15% with Twinn Aqua Suite.


Related Topics

Wastewater Treatment

Energy Efficiency

Machine Learning

Responsible for 16 wastewater treatment plants, Dutch water authority Vallei en Veluwe processes 340 million litres. The organisation’s goal is to draw all its energy from renewable sources by 2050. As such, it’s constantly searching for innovative treatment solutions, including strategies to improve surface water quality and extract more energy and raw materials from sludge.


Vallei en Veluwe was searching for ways to optimise the entire wastewater treatment process with the overarching aim of meeting its sustainability objectives. It also needed to reduce energy and chemical usage, as well as improve nutrient removal processes. In addition, local control systems and operations differed for each plant and were reactive rather than predictive. As a result, manual intervention was often required, which was an inefficient use of staff time.


Vallei en Veluwe implemented two, Twinn Aqua Suite solutions. The flow prediction solution uses real-time data and rainfall forecasts to predict flows in the sewage network and optimise discharge to the treatment plant with predictive control. This auto-pilot reduces peak flows and bypass of post-treatment, resulting in calmer networks and better effluent quality.

The organisation also deployed the process optimisation solution, which learns wastewater inflow patterns and predicts daily inflow, including detection of rain-weather flow to anticipate changing loads. It applies machine learning for feedforward control, finetuned with feedback control. Then, it adjusts aerators, pumps and valves to optimise nutrient removal and reduce energy and chemical use as far as possible.


  • 50% bypass reduction of post-treatment filters at the 4 treatment plants where Twinn flow prediction capabilities are used to control wastewater discharge from the sewage network.
  • 15% reduction in energy consumption across the 8 locations where process optimisation capabilities have been introduced.
  • Process operators have found the robust auto-pilot controller easy to use, and wastewater treatment system performance has been enhanced as a result.
  • Staff have more stable, predictive control, as well as an instant overview of the actual situation at the treatment plants.

With increasingly stringent effluent quality regulations, efficient systems are becoming more and more important. Twinn Aqua Suite provides great insights and smart control, which means better management of load peaks and usage of our installed treatment steps.

– Frank van de Grootevheen,
Process Engineer, Vallei en Veluwe
Featured SWAN Members: