Stochastic approaches for sensors placement against intentional contaminations in water distribution systems

Water distribution systems are vulnerable to intentional contamination: in this paper, optimising the placement of a set of monitoring stations to promptly detect this type of attack is considered. Due to the uncertainty about the place and the time of the beginning of the attack, and the inherent variability of the hydraulic conditions throughout the water distribution network, the impact probability distribution (IPD) has been defined in order to take into account the random variability of the conventional measures of impact. Two different approaches for the optimal placement of the monitoring stations are compared: the first minimises the expected value of a conventional damage measure, while the second minimises a given percentile of the IPD. The two approaches are applied to a real-world case study, showing their feasibility.