Plataforma integrada para optimización y diseño automático de rodetes de turbinas Francis
Autores
Trillos Sierra, Jose David
Director
Tipo de contenido
Trabajo de grado - Maestría
Idioma del documento
EspañolFecha de publicación
2018
Título de la revista
ISSN de la revista
Título del volumen
Documentos PDF
Resumen
Este trabajo presenta el desarrollo de una plataforma integral que permite el diseño de rodetes de turbinas Francis, basados en los parámetros de sitio (altura, caudal y velocidad de rotación), mediante la optimización de dos funciones objetivo: la eficiencia hidráulica y la resistencia a la erosión. Como variables de diseño se han seleccionado el ángulo relativo de flujo a la salida β_2 y el coeficiente de presión a la entrada ψ_1. Se muestra el flujo de diseño de la herramienta, los métodos de análisis por mecánica computacional de fluidos (CFD), la formulación del problema multiobjetivo, los algoritmos metaheurísticos empleados para este proceso de optimización y la ayuda de estrategias de inteligencia artificial - mas específicamente, el uso de redes neuronales artificiales -, para la evaluación de los objetivos. Finalmente, se muestra la integración del proceso de diseño en una única plataforma de software: FRAN-DES, y se presentan dos casos de prueba que permiten observar la validez de los resultados de la misma.
Abstract: This dissertation presents the development of an integral platform for design of Francis turbine runners, based on the location parameters of the turbine (Head, flow volume and rotational speed), by the means of the optimization of two objective functions: Efficiency and Erosion resistance. As design variables, two have been selected: the relative angle between the flow and the runnet at the outlet β_2 , and the pressure coefficient at the inlet ψ_1 . Here is presented the workflow of the platform, analysis methods via computational fluid mechanics (CFD), the formulation of the multi-objective problem, the metaheuristic algorithms employed to solve this problem and the help to the evaluations of such algorithms throug the use of artificial intelligenge strategies - specifically, the use of artificial neural networks (ANN) -. Finally, the incorporation of this design process in a single software platform named FRAN-DES is presented, and two test cases are studied, which allow to validate the results presented by this tool.
Abstract: This dissertation presents the development of an integral platform for design of Francis turbine runners, based on the location parameters of the turbine (Head, flow volume and rotational speed), by the means of the optimization of two objective functions: Efficiency and Erosion resistance. As design variables, two have been selected: the relative angle between the flow and the runnet at the outlet β_2 , and the pressure coefficient at the inlet ψ_1 . Here is presented the workflow of the platform, analysis methods via computational fluid mechanics (CFD), the formulation of the multi-objective problem, the metaheuristic algorithms employed to solve this problem and the help to the evaluations of such algorithms throug the use of artificial intelligenge strategies - specifically, the use of artificial neural networks (ANN) -. Finally, the incorporation of this design process in a single software platform named FRAN-DES is presented, and two test cases are studied, which allow to validate the results presented by this tool.