Modelagem numérica computacional aplicada à calibração através de novas equações para o uso do método iterativo do gradiente hidráulico/concentração alternativo (MIGHA) em redes de distribuição de água

Abstract

This work aims at the development of new equations for the use of the Iterative of the Alternative Hydraulic Gradient (MIGHA), in water distribution networks, based on the proportionality between the calibrated elements and the hydraulic gradient, the calibration of the Hazen-Williams coefficient C, the calibration of the friction factor of Darcy- Weisbach followed by the calculation of the absolute roughness, the calibration of the flow and nodal consumptions, both for the loss of energy along the flow calculated by equation of Hazen-Williams as well as the universal formula, or calculation of losses along the network and, for end, the calibration of the total chlorine decay constant along the tubing and posterior calculation of the chlorine decay coefficient on the pipe walls. One software, called UFC10, was developed allowing the calibration of the parameters of a network through the methodologies proposed in this work using as simulator hydraulic and water quality, Epanet. Several calibrations were performed, of which 348 C of Hazen-Williams, 348 of the friction factor, 348 of the flow with the equation of Hazen-Williams and 348 of the flow with the universal formula, to calculate the consumption of the nodes, all calibrations divided into 4 distribution networks. Also, 82 flow calibrations were performed for analysis of the physical losses of water in the network and, finally, 194 calibrations of the decay constant total chlorine. For all cases, the influence of the number of observed data was tested, the type of data observed, pressure or flow, the location of the observed data and the values parameters to be found. For this last test, a computational subroutine was developed with the purpose of seeking the best initial value. It was found that when the initial parameter chosen is close to the real one, the better the results found, however, values ​​with a certain distance also generate good results, but, Incoherent values ​​generate results that are different from reality. The computational subroutine developed did not reach the best result in any case, but it was a good tool. The results also showed that the use of flows as input data did not is suitable in the MIGHA calibration process of the C or the friction factor, where but it works well for flow calibration. Regarding the arrangement of nodes with data should be centralized and, if possible, distant from one another. To be compared to another method, MIGHA showed good results, but not the best ones.