An Interoperable Decision-Support Platform for Drinking-Water Treatment. Soft-Sensor Integration, Digital-Asset Reliability, and Techno-Economic Evaluation in the Waternet ToDrinQ Demonstration Case

Developed for the Waternet ToDrinQ Demonstration Case Associated with the Leiduin Drinking-Water Production Chain

Engd Thesis (2026)
Author(s)

A.A. Alexandraki (TU Delft - Civil Engineering & Geosciences, TU Delft - Civil Engineering & Geosciences, TU Delft - Reflection & Lifestyle)

Contributor(s)

L.C. Rietveld – Supervisor (TU Delft - Civil Engineering & Geosciences)

G. Kyritsakas – Supervisor (TU Delft - Civil Engineering & Geosciences)

Alex van der Helm – Supervisor (Waternet)

Research Group
Reflection & Lifestyle
More Info
expand_more
Publication Year
2026
Language
English
Graduation Date
11-09-2026
Awarding Institution
Sanitary Engineering , Water Management , Delft University of Technology
Research Group
Reflection & Lifestyle
Bibliographical Note
Engineering Doctorate dissertation, Delft University of Technology, 2026. Developed within the Horizon Europe ToDrinQ project and demonstrated in collaboration with Waternet.
Page Views
26
Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Abstract

Drinking-water utilities increasingly rely on measured data, predictive models, and digital advisory tools to respond proactively to changing source-water quality and treatment conditions. The engineering challenge is not only to generate predictions, but also to integrate heterogeneous information sources in a way that preserves their meaning, provenance, status, and professional responsibility.

Within the Horizon Europe ToDrinQ project, an interoperable decision-support platform for drinking-water treatment was developed and evaluated using Waternet, the public water utility serving Amsterdam and its surrounding area, as the demonstration case. The scope covered source-water awareness at Nieuwegein and the coagulation–flocculation and ozonation processes associated with the Leiduin treatment system.

Twenty platform requirements were consolidated into four design principles: separation of platform functions, defined interfaces for interoperability, human-in-the-loop decision support, and separation of data acquisition from model execution. The platform integrated measured information, predictive soft sensors, advisory services, operational risk interpretation, and external information exchange. Implemented services addressed source-water turbidity, coagulation performance, ferric-chloride dosing, ozone exposure, bromate formation, and microbial-removal performance.

The platform was evaluated through software verification, communication with Waternet’s Process Information Management System, external interoperability testing, stakeholder review, and requirement verification. Information exchange with the NESSIE partner platform demonstrated that values could be transferred while retaining their units, timestamps, provenance, and service identity. Stakeholder feedback further refined the operator-facing workflow. Measurements, predictions, recommendations, and unavailable information remained distinguishable, while selected predictions were translated into Low, Moderate, High, or Unknown operational states. Invalid or unavailable information resulted in an Unknown state rather than being interpreted as safe, and final operational decisions remained with the responsible professionals.

A reliability-centred method was also developed to manage soft sensors and related software services as digital assets. Application to the ozonation concentration–time soft sensor showed that incomplete or unavailable input information and model drift were key reliability concerns. The results highlighted that digital reliability depends not only on model execution, but also on input validity, visibility of degraded conditions, and lifecycle monitoring.

A techno-economic evaluation compared Business-as-Usual with platform-supported operation. For Waternet continuation, the base case resulted in an estimated seven-year Net Present Value of approximately €7,161. For another utility adopting the existing platform, the base-case NPV was approximately −€42,839, while the optimistic case increased to approximately €132,379 with a payback period below two years. Development of an equivalent platform from scratch remained financially negative in the evaluated cases.

Overall, the research demonstrates how measured, predictive, advisory, risk, reliability, and interoperability services can be integrated within a reusable decision-support environment while maintaining human responsibility for drinking-water treatment decisions.

Files

License info not available
warning

File under embargo until 30-09-2028