Calibração de rugosidades em redes de distribuição de água através de uma adaptação do método iterativo do gradiente hidráulico alternativo (MIGHA)

• Ano de defesa: 2020
• Autor(a) principal: Sousa, David Lopes de
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Palavras-chave em Português: Recursos hídricos; Calibração; Absolute roughness; Standardization routine; Objective functions
• Link de acesso: http://www.repositorio.ufc.br/handle/riufc/50048

The calibration of water distribution networks is an important tool for obtaining computational models that better represent the installed network, directly contributing supply systems management. The present work proposed to test a new routine that adapts the Alternative Hydraulic Gradient Method (MIGHA), used in calibration of the absolute roughness and the Hazen-Williams C coefficient, making it more the results obtained for these parameters in pipes that have the same material. To test the new model, software was developed that made it possible to 192 calibrations were performed on six networks, where three of them were calibrated absolute roughness and in the other three the Hazen-Williams C coefficients. In these tests, the results obtained by admitting the use and non-use of the new proposed standardization, also verifying the influence of the amount of data observed and its impact on solutions. In addition, different objective functions were tested, with in order to verify which is the most suitable for the use of the method. In the tests that involved the absolute roughness calibration, the results obtained suggest that the new routine improves significantly the roughness values, requiring little data observed to return great solutions. The pressures calculated in virtually all networks calibrated had small errors, less than 0.50 m in all tests, proving that MIGHA is extremely efficient in reducing errors in this variable. The results obtained for flows in these networks were satisfactory. The tests that involved the calibration of the Hazen-Williams C coefficient suggested that the new routine needs many observed data to calibrate networks with many materials. Despite this, the new model presented better results in most of the calibrations performed, indicating that the new standardization routine is beneficial to the method. The different objective functions tested significantly influenced the results.