Calibration of chlorine model by genetic algorithms in drinking water network of Guanajuato

The concentration of chlorine as a disinfectant supplied into the water distribution network decays over time due to its interaction with microorganisms, metals, and other substances present in the water. Computer programs, like EPANET, are used to simulate hydraulic behavior and water quality in distribution networks; with them, physical and operational modifications can be represented in a numerical model to understand the new behavior without compromising the quality of service. The calibration of a water quality model involves referencing the concentration of a substance at various points in the network and adjusting the decay coefficients in the model until the simulated concentrations match the measurements. In this paper, it is shown the automatization of this process, using the heuristic technic genetic algorithms and the EPANET Toolkit with MATLAB (2016). The calibration process is carried out in a sector of the water distribution network of the city of Guanajuato, obtaining a correlation of 0.816 with an absolute average error of 0.08 mg/l in the modeled concentrations with respect to the measurements and this will allow to know the concentration of chlorine at diverse points of the network, thus ensuring that quality service is provided throughout the supply.