Application of reduced order models in industrial applications for aviation
This work focuses on the development and application of a Reduced-Order Model (ROM) based on Proper Orthogonal Decomposition (POD) for aerodynamic analysis in the aviation industry. Reynolds-Averaged Navier–Stokes (RANS) simulations, which are the most commonly used approach in industrial aerodynami...
| Autor: | |
|---|---|
| Tipo de recurso: | tesis de maestría |
| Fecha de publicación: | 2025 |
| País: | España |
| Institución: | Universitat Politècnica de Catalunya (UPC) |
| Repositorio: | UPCommons. Portal del coneixement obert de la UPC |
| Idioma: | inglés |
| OAI Identifier: | oai:upcommons.upc.edu:2117/455194 |
| Acceso en línea: | https://hdl.handle.net/2117/455194 |
| Access Level: | acceso embargado |
| Palabra clave: | Aerodynamics Computational fluid dynamics Reduced Order Model Aerodinàmica Dinàmica de fluids computacional Àrees temàtiques de la UPC::Aeronàutica i espai::Aerodinàmica |
| Sumario: | This work focuses on the development and application of a Reduced-Order Model (ROM) based on Proper Orthogonal Decomposition (POD) for aerodynamic analysis in the aviation industry. Reynolds-Averaged Navier–Stokes (RANS) simulations, which are the most commonly used approach in industrial aerodynamic studies, remain computationally expensive and often impractical for iterative tasks such as design optimization or parametric studies. ROMs offer a promising alternative by approximating the behavior of complex systems using a reduced set of governing modes, significantly lowering computational cost while retaining essential flow physics. The Common Research Model in its High-Lift configuration (CRM-HL) was selected as the test case, providing a representative benchmark for commercial aircraft during take-off and landing phases. To ensure the reliability of the ROM, the CODA solver was first validated against available reference data, demonstrating its capability to reproduce aerodynamic fields accurately. A comprehensive database was then generated at a freestream Mach number of 0.2 and a Reynolds number of 5.49 × 106 , consisting of pressure coefficient (Cp) and skin-friction coefficient (Cf ) distributions over the CRM’s wing for 21 angles of attack ranging from 0º to 20º. A full ROM was built using training data from angles of attack 0º, 5º, 10º, 15º, 17º, 19º, and 20º. For intermediate angles, predictions were obtained by interpolating the temporal modes matrix, allowing accurate estimation of aerodynamic fields across the full range. These predicted Cp and Cf distributions were subsequently integrated over the wing surface to compute the lift and drag coefficients, Cℓ and Cd. The results demonstrate that the POD-based ROM accurately reproduces the dominant flow features while requiring only a fraction of the computational resources compared to full RANS simulations. It is estimated that, for the case studied, the ROM could save approximately 840 hours of simulation time and prevent the emission of around 2 tons of CO2, highlighting both economic and environmental benefits. Overall, this study shows that ROMs can provide a reliable, efficient, and sustainable alternative to high-fidelity CFD simulations, facilitating rapid design exploration and optimization while supporting informed engineering decisions in industrial aerodynamic workflows. |
|---|