Thesis Cálculo de la distancia peligrosa de un relave mediante herramientas de fluidodinámica computacional (CDF)
Loading...
Date
2023
Authors
Journal Title
Journal ISSN
Volume Title
Program
Ingeniería Civil de Minas
Campus
Campus Santiago San Joaquín
Abstract
En el presente estudio se ha modelado el flujo de un relave posterior a ocurrido un colapso de un tranque de relave o rompimiento del muro de este, con el objetivo de predecir la distancia peligrosa. Se realiza el modelamiento computacional de los relaves como un fluido no newtoniano, utilizando herramientas de fluidodinámica computacional (CFD), específicamente el software FLUENT desarrollado por ANSYS, con la finalidad de estudiar su comportamiento.
El flujo es modelado en un régimen laminar, siguiendo el modelo de viscosidad de Herschel-Bulkley y además el sistema a modelar incluye la consideración de una fase granular para la incorporación de las partículas sólidas, que para el caso de estudio es considerada constante de un cinco por ciento de concentración en volumen.
La geometría corresponde a un canal rectangular con una superficie lisa de 6 metros de longitud en donde el fluido puede escurrir libremente, este dominio físico es discretizado en 510,255 celdas de tipo poliédricas con un tamaño de 3 centímetros. En cada una de estas celdas se resuelven las ecuaciones de conservación de masa y momentum en cada paso de tiempo.
Para la generación de un modelo predictivo se realiza una serie de simplificaciones de las ecuaciones que gobiernan el sistema tanto para el fluido como para las partículas, con la finalidad de encontrar los parámetros adimensionales que acompañan a las variables vectoriales. El modelo propuesto relaciona variables de la fase sólida, el fluido y variables geométricas en función de la distancia máxima.
Los resultados no muestran una correlación entre los parámetros adimensionales y la variable de distancia máxima adimensional, se analiza uno a uno para visualizar una dependencia entre estos parámetros. Mediante la agrupación de estos parámetros se realiza una regresión lineal, obteniendo una buena correlación del grupo adimensional con la distancia máxima, determinado así un modelo predictivo de la distancia máxima.
Se concluye que a partir de los resultados obtenidos existe una incidencia notoria de la distribución granulométrica y la tensión de fluencia en la distancia máxima. La validación de los modelos matemáticos, en general, deben ser validados físicamente para confirmar el grado de confiabilidad y así potenciar el uso de las herramientas computacionales en futuros estudios, que pueden ser aplicados en la industria minera.
In the present study, the tailings flow has been modeled after a collapse of a tailings dam or breakage of its wall has occurred, with the aim of predicting the dangerous distance. Computational modeling of the tailings as a non-Newtonian fluid is carried out, using computational fluid dynamics (CFD) tools, specifically the FLUENT software developed by ANSYS, in order to study its behavior. The flow is modeled in a laminar regime, following the Herschel-Bulkley viscosity model and in addition the system to be modeled includes the consideration of a granular phase for the incorporation of solid particles, which for the case of study is considered a constant of a five percent concentration by volume. The geometry corresponds to a rectangular channel with a smooth surface 6 meters long where the fluid can flow freely, this physical domain is discretized into 510,255 polyhedral type cells with a size of 3 centimeters. In each of these cells the mass and momentum conservation equations are solved at each time step. To generate a predictive model, a series of simplifications of the equations that govern the system are carried out for both the fluid and the particles, in order to find the dimensionless parameters that accompany the vector variables. The proposed model relates variables of the solid phase, the fluid and geometric variables as a function of the maximum distance. The results do not show a correlation between the dimensionless parameters and the dimensionless maximum distance variable; they are analyzed one by one to visualize a dependence between these parameters. By grouping these parameters, a linear regression is performed, obtaining a good correlation of the dimensionless group with the maximum distance, thus determining a predictive model of the maximum distance. It is concluded that from the results obtained there is a notable incidence of the granulometric distribution and the yield stress at the maximum distance. The validation of mathematical models in general must be physically validated to confirm the degree of reliability and thus enhance the use of computational tools in future studies, which can be applied in the mining industry.
In the present study, the tailings flow has been modeled after a collapse of a tailings dam or breakage of its wall has occurred, with the aim of predicting the dangerous distance. Computational modeling of the tailings as a non-Newtonian fluid is carried out, using computational fluid dynamics (CFD) tools, specifically the FLUENT software developed by ANSYS, in order to study its behavior. The flow is modeled in a laminar regime, following the Herschel-Bulkley viscosity model and in addition the system to be modeled includes the consideration of a granular phase for the incorporation of solid particles, which for the case of study is considered a constant of a five percent concentration by volume. The geometry corresponds to a rectangular channel with a smooth surface 6 meters long where the fluid can flow freely, this physical domain is discretized into 510,255 polyhedral type cells with a size of 3 centimeters. In each of these cells the mass and momentum conservation equations are solved at each time step. To generate a predictive model, a series of simplifications of the equations that govern the system are carried out for both the fluid and the particles, in order to find the dimensionless parameters that accompany the vector variables. The proposed model relates variables of the solid phase, the fluid and geometric variables as a function of the maximum distance. The results do not show a correlation between the dimensionless parameters and the dimensionless maximum distance variable; they are analyzed one by one to visualize a dependence between these parameters. By grouping these parameters, a linear regression is performed, obtaining a good correlation of the dimensionless group with the maximum distance, thus determining a predictive model of the maximum distance. It is concluded that from the results obtained there is a notable incidence of the granulometric distribution and the yield stress at the maximum distance. The validation of mathematical models in general must be physically validated to confirm the degree of reliability and thus enhance the use of computational tools in future studies, which can be applied in the mining industry.
Description
Keywords
Fluidodinámica computacional, Modelo predictivo, Distancia peligrosa
