Comparativa de modelación CFD: casos tipo industria minera en OpenFOAM vs ANSYS Fluent.
No Thumbnail Available
Date
2026
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Universidad de Concepción
Abstract
La Dinámica de Fluidos Computacional (CFD) es una herramienta indispensable en la industria minera para optimizar el diseño de infraestructuras. Sin embargo, la dependencia de softwares comerciales como ANSYS Fluent conlleva elevados gastos operativos en licencias comerciales, posicionando a OpenFOAM como una alternativa de código abierto atractiva por su escalabilidad sin costos de licenciamiento. Pese a esto, aún existe incertidumbre sobre la precisión de OpenFOAM frente a estándares comerciales al simular fluidos complejos. Por lo tanto, el objetivo de este estudio fue comparar la fiabilidad técnica y el rendimiento computacional de ambas plataformas mediante la evaluación de dos escenarios comunes en la minería: un modelo en régimen permanente de transferencia de calor en tuberías y un modelo transitorio multifásico para el transporte de relaves.
Para comparar estrictamente los algoritmos de resolución, se emplearon mallas poliédricas idénticas generadas en ANSYS Meshing. Las discrepancias en los perfiles térmicos, campos de velocidad y tiempos de cómputo se cuantificaron utilizando el Error Cuadrático Medio Normalizado (NRMSE). En el régimen estacionario térmico, OpenFOAM demostró una equivalencia técnica sobresaliente, calculando el equilibrio de temperaturas con desviaciones inferiores a 0,3 °C y procesando las ecuaciones en un 65% menos de tiempo que Fluent. Por el contrario, en el modelo transitorio multifásico (método VOF), las diferencias en la captura de la interfaz generaron discrepancias numéricas en OpenFOAM, induciendo errores cinemáticos del 27,82 % en la aproximación bidimensional, los cuales solo pudieron reducirse a un 9.5% en un dominio tridimensional tras una exhaustiva sintonización de esquemas limitadores de gradiente.
Los hallazgos demostraron que la viabilidad técnica de sustituir el estándar comercial por código abierto depende directamente del régimen de flujo modelado. Se concluye que OpenFOAM es una herramienta financieramente óptima y de alta precisión para estudios termodinámicos estacionarios, pero su inestabilidad transitoria frente a fuertes gradientes de presión hidrostática limita su aplicación directa en procesos de transporte de pulpas. Superar esta limitación en ingeniería de detalle exigirá en el futuro una validación empírica en laboratorio y el desarrollo de funciones de amortiguación interna en los solucionadores de código abierto.
Computational Fluid Dynamics (CFD) is an indispensable tool in the mining industry for optimizing infrastructure design. However, the reliance on commercial software such as ANSYS Fluent entails high operating expenses for commercial licenses, positioning OpenFOAM as an attractive open-source alternative due to its scalability without licensing costs. Despite this, there is still uncertainty regarding the accuracy of OpenFOAM compared to commercial standards when simulating complex fluids. Therefore, the objective of this study was to compare the technical reliability and computational performance of both platforms by evaluating two common scenarios in mining: a steady-state model of heat transfer in pipes and a transient multiphase model for tailings transport. To strictly compare the solving algorithms, identical polyhedral meshes generated in ANSYS Meshing were used. The discrepancies in the thermal profiles, velocity fields, and computation times were quantified using the Normalized Root Mean Square Error (NRMSE). In the thermal steady state, OpenFOAM demonstrated outstanding technical equivalence, calculating the temperature equilibrium with deviations of less than 0.3 °C and processing the equations in 65% less time than Fluent. Conversely, in the transient multiphase model (VOF method), differences in interface capturing generated numerical discrepancies in OpenFOAM, inducing kinematic errors of 27.82% in the two-dimensional approximation, which could only be reduced to 9.5% in a three-dimensional domain after exhaustive tuning of gradient limiter schemes. The findings demonstrated that the technical feasibility of replacing the commercial standard with open-source software depends directly on the modeled flow regime. It is concluded that OpenFOAM is a financially optimal and highly accurate tool for steady-state thermodynamic studies, but its transient instability in the face of strong hydrostatic pressure gradients limits its direct application in slurry transport processes. Overcoming this limitation in detailed engineering will require future empirical laboratory validation and the development of internal damping functions in open-source solvers.
Computational Fluid Dynamics (CFD) is an indispensable tool in the mining industry for optimizing infrastructure design. However, the reliance on commercial software such as ANSYS Fluent entails high operating expenses for commercial licenses, positioning OpenFOAM as an attractive open-source alternative due to its scalability without licensing costs. Despite this, there is still uncertainty regarding the accuracy of OpenFOAM compared to commercial standards when simulating complex fluids. Therefore, the objective of this study was to compare the technical reliability and computational performance of both platforms by evaluating two common scenarios in mining: a steady-state model of heat transfer in pipes and a transient multiphase model for tailings transport. To strictly compare the solving algorithms, identical polyhedral meshes generated in ANSYS Meshing were used. The discrepancies in the thermal profiles, velocity fields, and computation times were quantified using the Normalized Root Mean Square Error (NRMSE). In the thermal steady state, OpenFOAM demonstrated outstanding technical equivalence, calculating the temperature equilibrium with deviations of less than 0.3 °C and processing the equations in 65% less time than Fluent. Conversely, in the transient multiphase model (VOF method), differences in interface capturing generated numerical discrepancies in OpenFOAM, inducing kinematic errors of 27.82% in the two-dimensional approximation, which could only be reduced to 9.5% in a three-dimensional domain after exhaustive tuning of gradient limiter schemes. The findings demonstrated that the technical feasibility of replacing the commercial standard with open-source software depends directly on the modeled flow regime. It is concluded that OpenFOAM is a financially optimal and highly accurate tool for steady-state thermodynamic studies, but its transient instability in the face of strong hydrostatic pressure gradients limits its direct application in slurry transport processes. Overcoming this limitation in detailed engineering will require future empirical laboratory validation and the development of internal damping functions in open-source solvers.
Description
Tesis presentada para optar al título de Ingeniero/a Civil Mecánico/a.
Keywords
Dinámica de fluidos computacional, Industria minera, Software computacional