Efeito de diferentes estratégias de discretização hidráulica e hidrológica na modelagem de bacias urbanas

• Ano de defesa: 2022
• Autor(a) principal: Persch, Cristiano Gabriel
• Orientador(a): Piccilli, Daniel Gustavo Allasia
• Banca de defesa: Pinto, Leandro Conceição, Garcia, Joaquin Ignacio Bonnecarrere, Bravo, Juan Martin, Vasconcelos Neto, José Goes
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Santa Maria; Centro de Tecnologia
• Programa de Pós-Graduação: Programa de Pós-Graduação em Engenharia Civil
• Departamento: Engenharia Civil
• País: Brasil
• Palavras-chave em Português: Modelagem; Discretização; Soluções baseadas na natureza
• Palavras-chave em Inglês: Modeling; Discretization; Nature-based solutions
• Área do conhecimento CNPq: CNPQ::ENGENHARIAS::ENGENHARIA CIVIL
• Link de acesso: http://repositorio.ufsm.br/handle/1/25622

The design and application of Nature-Based Solutions (NBSs) systems in the control of surface runoff is commonly coupled with hydraulic-hydrological simulation to define the best strategies, potential effects, and benefits on the surface. Since the urban environment is defined by high heterogeneity and the hydrological processes are marked by strong non-linearity, it is necessary to understand how spatial discretization influences the modeling process, seeking to find a balance between the level of discretization and the efforts undertaken, mainly in urban and periurban regions, where this characteristic is even more highlighted. A prominent tool to perform this task is the SWMM (Storm Water Management Model), a model widely used in the world with considerable success in the modeling process. SWMM solves the hydraulic and hydrological component of a basin during the simulation process, and it is suitable for most applications by adequately simulating the flow processes in the systems. However, it was observed in the literature that the model does not represent dynamic changes well, particularly transitions to pressurized flow conditions, which is characteristic of extreme events. Since the model does not use spatial discretization within conduits as a solution of the system, it can lead to high values of continuity error, presence of peaks or numerical oscillations. In the same way, hydrological discretization also influences systems, and the role of the modeler is to adopt solutions that minimize adverse effects of non-discretization. The level of discretization in the hydrological components has impact on the representation of urban surfaces, and on the performance evaluation and projects of NBSs systems. Thus, in this research, the effect of different strategies of hydraulic and hydrological discretization in the simulation of a typical urban basin and its effect on the application of NBSs-type structures were evaluated. Regarding the hydraulic component of the model, different strategies of discretization of the hydraulic network were implemented in SWMM and their effect on the improvement in the simulations, continuity errors and computational time were evaluated. The results indicated that the traditional discretization (DTrad) did not accurately represent the observed data. On the other hand, the artificial discretization (ASD) showed a significant improvement from the representation of conduits with lengths smaller than 20m (NSE>0.7). However, longer computational times were necessary as smaller discretization were adopted. ASD discretization also significantly reduced continuity errors compared to DTrad. Regarding the hydrological component, four levels of detail of the surface of the study area (1; 5; 12 and 21 basins) were implemented in the SWMM (with and without NBSs) evaluating the model's sensitivity related to the flow dynamics in terms of maximum flows and the sensitivity of the parameters on the flow. The results indicated that higher resolutions (12 and 21 sub-basins) led to higher peak flows for both monitored and design events for scenarios with and without NBSs. As the drainage system is represented in more detail, higher peak runoff flows were found. Regarding the parameters, the greater influence of the CN variables and basin width stood out, especially in the less refined representations (1 and 5 basins), reinforcing that less detailed discretization are more susceptible to uncertainties compared to the others, which may potentialize errors in eventual evaluations of the effects and dimensioning.