Securing urban water infrastructures with digital twins

Abstract:

The advent of Industry 4.0 is transforming the urban water sector: physical infrastructures are now coupled with programmable logic controllers, SCADA systems, sensors, and other field devices that improve efficiency and reliability. To harness such advancements, we need to deploy novel digital solutions that can better operate our systems and inform decisions by fast assimilation and processing of large datasets. Following the developments in other critical sectors, water utilities are now increasing their efforts to build digital twins, that is, virtual entities replicating at varying levels of detail the processes and assets within the infrastructure. Digital twins can be devised to reproduce with higher-fidelity specific operations and phenomena to better suit targeted applications, such as predictive maintenance or real-time control. Here, we present the first digital twin for studying the behavior of water distribution systems under cyber-attacks, an increasing threat due to the progressive digitalization of water infrastructures. Our approach builds on a two-way interaction between a process-based numerical model (EPANET/WNTR) and a network emulation platform (MiniCPS), which is able to accurately reproduce the digital processes in the cyber-layer of the infrastructure. By emulating both cyber and physical layers, the digital twin can replicate the effects of the attacks on system hydraulics as well as the anomalous traffic and operations engendered in the cyber-network. We implement the digital twin on a real-world water distribution system and disrupt its operations with a number of cyber-physical attacks—the physical consequences are broad, ranging from limited water supply to the malfunctioning of pumps and valves. Importantly, we show that combining data retrieved from both physical processes and communication and control network allows us to detect and characterize the attacks’ nature. This is an important step towards securing modern water networks.