Otimização de leis de manobras de comportas de eclusas de navegação com uso de algoritmo genético

• Ano de defesa: 2013
• Autor(a) principal: Denis Furstenau Plec
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Meio Ambiente e Recursos Hídricos; Saneamento; Engenharia hidráulica Modelos matemáticos; Recursos hídricos Desenvolvimento
• Link de acesso: http://hdl.handle.net/1843/BUOS-97HHEJ

The use of mathematical modeling has been proved to be of a great value in many areas of scientific research. Associated with physical modeling, mathematical modeling enables a broad analysis of hydraulic behavior during a lock operation so that a pre-plan or even the alteration of locks that have been already constructed can result in gate operations with a better hydraulic behaviour, without jeopardizing the structures and ships in transit. The methodology of this research is composed by two fronts. The first front is the calibration of a simulation model that derives of the equations of position, velocity and acceleration of the water level inside a chamber of a lock obtained by Laplace transform in the general equation of the outflow in the locks aqueduct. The calibration occurred, in a first moment, comparing the Laplace transform with the application of the method of Runge-Kutta in its forth order and, in a second moment, by choosing the coefficients of load loss of a lock and turned effective through the comparison of the results with the data obtained from the lock hydraulic model constructed in the Laboratório de Hidráulica Experimental e Recursos Hídricos (LAHE/RJ) of Furnas Centrais Elétricas. The second front of this research consists in developing an optimization model for valve operation of a lock through a genetic algorithm. The algorithm was applied and compared to the results obtained for the Promissão lock, situated on the Tietê river, in São Paulo, by usinga model that based on the linear sequential programming or Kelleys cutting plane method. This algorithm is better applied in floodgates that have the system of opening and closing of the gates with a limited number of operation speeds because the genetic algorithm is characterized for limiting the problems variable into a finite number of values. The biggest advantages of the utilization of the algorithm are the easy insertion of additional restrictions and the fact that the they dont need to be linearized. This optimization model was also applied to the physical model available at the LAHE/RJ.