Calibração de modelos de drenagem urbana utilizando algoritmos evolucionários multiobjetivo

• Autor(a) principal: CARVALHO, Maíra de
• Data de Publicação: 2011
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da UFG
• Texto Completo: http://repositorio.bc.ufg.br/tede/handle/tde/613
• topic: modelos hidrológicos; otimização multiobjetivo; algoritmos; evolucionários; drenagem urbana; hydrological models; multiobjective optimization; evolutionary algorithms; urban drainage; CNPQ::ENGENHARIAS::ENGENHARIA SANITARIA::RECURSOS HIDRICOS

CARVALHO, M. Calibration models of urban drainage using multiobjective evolutionary algorithms. 2011. Dissertation (Masters of Environmental Engineering) - Civil Engineering College, Post-Graduation Stricto Sensu Program in Environmental Engineering - Federal University of Goiás, Goiânia, Goiás, Brazil, 2011.. This study proposed to develop and implement a calibration routine hydrological models applied to urban drainage using multiobjective optimization techniques. To make this work possible model was adopted Storm Water Management Model (SWMM) and the computational algorithms developed in MATLAB environment using an evolutionary algorithm. The method was applied to two different levels of detail in representing the Arroyo Cancels basin, located in the urban area of Santa Maria-RS, submitted to the hydrological processes involved in the process of rainfall-runoff transformation in the search for optimal values of hydrological parameters the basin. Objective functions were defined and applied simultaneously in the calibration parameters. Worked with the simulation of events of low and high intensity settings for two discretization of the watershed, and other simple and subdivided into 18 sub-basins. The sensitivity analysis performed made it possible to check that the parameters that most influenced the basin were simple: Percentage of impervious area and outlet width. Regarding the results for the various watershed discretization can be seen that in most cases when working with a more detailed watershed they were better, except for some isolated events. Overall the model showed better results when high-intensity simulated events for the best compromise solutions, thus showing the importance of using a multiobjective model.