Objectives

The main scientific and technical objectives are to:

1 -Protect a drinking water distribution network and its data against security attacks:  cyber-attacks, physical attacks, and intentional contamination.
2 - For the use cases from the literature and then for the Eurométropole Strasbourg case studies be able to :
    a -- Develop an attack detection and characterisation system,
    b -- Improve classification and reduce the false positive rate,
    c -- Develop a method for improving resilience through the design of WDNs.

Figure Concept idea of CoRREau. The two physical infrastructures (WDN & IS) are shown with their respective cyber physical systems and supervision systems to form the so-called real world (see the WDN, OT & IT supervision at the bottom layer). In blue and yellow at the Models layer, the two digital models (DMs) are built, the first for the WDN (in WP3), and the second for the IS (in WP4), from their physical models and from available data. It is both data-driven (see arrows a and c) from collected data, and model-driven (arrow a’ and c’) from an existing off-line high-fidelity model for the WDN, and from business rules and expert knowledge for the IS. The WDN DM is a DT because the arrow b represents feedback to propose countermeasures (levels to be reached, valves to be closed, etc.).  The IS DM is a DS because the arrow d symbolises the knowledge enrichment needed for its training/building. In green, at the top of the Models layer, we can see the unification of IT, OT and WDN data in the unified data lake, and the mutual contributions of the DS and DT to the anomaly detection and overall resilience of both infrastructures. The objective of WP5 is to build a unified digital model capable of automatically planning and implementing remediations or countermeasures following the detection of targeted aqua-cyber-attacks on IS and/or WDN infrastructures. The DT’s implementation of an attack remediation plan (arrow f) assumes the existence, within the unified DMs, of a unified attack detection model on which is based a decision support system capable of processing data (arrow e) coming from the Real world. The production of this unified model and the decision support system are challenges addressed by WP5.