Aplicação da técnica simulated annealing na investigação da ciclagem de nitrogênio na inteface água-sedimento

• Ano de defesa: 2014
• Autor(a) principal: Kalas, Francine de Almeida
• Orientador(a): Rodrigues, Pedro Paulo Gomes Watts
• Banca de defesa: Oliveira, Vicente de Paulo Santos de , Câmara, Leôncio Diógenes Tavares , Moura Neto, Francisco Duarte , Limaverde Filho, Aricelso Maia
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade do Estado do Rio de Janeiro
• Programa de Pós-Graduação: Programa de Pós-Graduação em Modelagem Computacional
• Departamento: Centro de Tecnologia e Ciências::Instituto Politécnico
• País: BR
• Palavras-chave em Português: Simulated annealing (Matemática); Ciclo do nitrogênio - Modelos matemáticos; Otimização combinatória; Difusão - Modelos matemáticos; Porosidade Modelos matemáticos; Sedimentos; Coeficiente de difusão
• Palavras-chave em Inglês: Nitrogen; Diffusion coefficient; Porosity; Sediments; Simulated annealing
• Área do conhecimento CNPq: CNPQ::CIENCIAS EXATAS E DA TERRA::MATEMATICA::MATEMATICA APLICADA
• Link de acesso: http://www.bdtd.uerj.br/handle/1/13680

This work presents the application of an optimization method to estimate parameters that are usually present in the mathematical modeling of chemical species dynamics in the water-sediment interface . Here, the Direct Problem was the simulation of organic and inorganic (ammonium and nitrate) nitrogen species concentrations in an idealized environment, which was fractionated into four layers: a layer of water (1 meter depth) and three layers of sediment (0-1 cm 1-2 cm and 2-10 cm). The Direct Problem was solved by the Runge Kutta method, yielding a 50 days simulation. Thus, the Simulated Annealing (SA) Method was applied to estimate diffusion coefficients and porosity. The strategy efficiency was evaluated by comparing synthetic experimental data with those yielded by the direct problem solution, adopting the parameters estimated by SA Method. The best fitting between experimental and calculated concentrations was achieved when the porosity was the estimated parameter. Regarding the Objective Function minimization, the estimative of this parameter also required lower computational effort. After introducing a random noise to the concentrations of nitrogenous species, SA technique was unable to obtain a satisfactory estimate for the diffusion coefficient, with the exception of sediment layer 0-1 cm. For the other layers, concentrations errors as high as 10% were found (for the ammonia concentration in the water layer, for example). The results showed that the methodology adopted here can be a quite promising tool in the water bodies management, especially in those submitted to low energy, as lakes and coastal lagoons.