Desenvolvimento de metodologia para geração de malha estruturada em tubo de vórtice

Abstract

The primary objective of this thesis was to develop and validate a robust methodology for generating structured computational meshes applied to the simulation of a Ranque-Hilsch Vortex Tube (RHVT). Accurate numerical prediction of the hot and cold flow separation in this device is historically hindered by model sensitivity to spatial discretization quality. To overcome the limitations of unstructured meshes, a three-dimensional model was developed using Computational Fluid Dynamics (CFD) in OpenFOAM. The central innovation lay in the implementation of a parametric mesh generation approach ("topology-first") using the classy_blocks Python library. This strategy allowed for the construction of a purely hexahedral mesh with precise anisotropic refinement control and guaranteed topological connectivity. Turbulent flow simulation was performed using the Large Eddy Simulation (LES) approach with the rhoPimpleFoam solver. Results demonstrated the effectiveness of the proposed methodology, yielding exceptional quality metrics with maximum non-orthogonality lower than 30° and eliminating critical numerical instabilities. Functional validation confirmed the model's ability to stably reproduce energy separation, generating consistent temperature gradients. In summary, this study provides a reliable and parameterized computational tool, establishing a solid foundation for future RHVT optimization.