Optimal Sensor Locations in Water Distribution Networks

Water distribution networks are inherently vulnerable to accidental or intentional water contaminations. There is a research aiming to prevent such threats in water networks. One of the problems arising is to identify sensor locations that minimize the impacts like contaminated volume consumed, population exposure and detection time of entrance of contamination. In this paper, we present a model of placing sensors in water network. we focus on the sensor configuration that minimizes the expected population exposed and time to detect. We formulate these problems as binary nonlinear objectives, and propose a solution based on simulated annealing. We simulated the contamination transport in EPANET in order to derive flow and velocity information. A novel approach also is introduced in order to handle different flow patterns occurring during the day.

Keywords:

sensor placement , simulated annealing , expected population , optimization