392039 Machine Learning for Water Distribution Networks (S) (SoSe 2026)

This course is about applications of Machine Learning (ML) in Water Distribution Networks (WDNs).
In the first part of the course (approx. 4 weeks), students will learn about the fundamentals of WDNs such as their physical properties (e.g. hydraulic and quality dynamics) and state-of-the-art modeling and simulation approaches. Students will gain hands-on experience in how to create and simulate complex WDN scenarios in Python.
In the second part, (groups of) students will work on a project of their choice where they have to use ML to solve a given task in a WDN. Potential projects include: Event detection in WNDs such as leakage, fault, or contamination detection; Reinforcement Learning for water quality control; forecasting of quantities such as flows and chemical concentrations; sensor placements; etc. Students are expected to write a short summary and present their results.

Please note that "392040 Tutorials for Machine Learning for Water Distribution Networks (Ü)" will be used as an additional consultation hour where students can go and ask the instructor for help and guidance when working on their projects.

Requirements for participation, required level

Prerequisite:
- Foundations of Machine Learning and Deep Learning