Development of Reduced Order Modeling Approaches for Nonlinear Aeroelastic Analyses

• Grant number: 25/05625-6
• Support Opportunities: Scholarships abroad - Research Internship - Doctorate
• Start date: September 01, 2025
• End date: August 31, 2026
• Field of knowledge: Engineering - Aerospace Engineering - Aerodynamics
• Principal Investigator: João Luiz Filgueiras de Azevedo
• Grantee: Ana Cristina Neves Carloni
• Supervisor: Carlos Eduardo Stolf Cesnik
• Host Institution: Pró-Reitoria de Pós-Graduação e Pesquisa. Instituto Tecnológico de Aeronáutica (ITA). Ministério da Defesa (Brasil). São José dos Campos , SP, Brazil
• Institution abroad: University of Michigan, United States
• Associated to the scholarship: 24/01647-2 - Development of Reduced Order Modeling Approaches for Nonlinear Aeroelastic Analyses, BP.DR

Abstract

The traditional approach for obtaining high-fidelity aeroelastic solutions involves directly coupling structural-dynamic models with Computational Fluid Dynamics (CFD) solvers, exploiting their physics-informed formulations. This method is particularly valuable for flutter and limit cycle oscillation (LCO) analyses in transonic flows, where accurate modeling of aerodynamic nonlinearities is crucial. These nonlinearities are mainly caused by shock wave motions, which introduce phase lag effects that significantly affect aeroelastic stability. However, performing numerous unsteady CFD simulations to cover the entire flight envelope results in prohibitively high computational costs, limiting the scalability of aeroelastic analyses for complex engineering configurations. To address this challenge, we propose the development of aerodynamic Reduced-Order Models (ROMs) tailored to capture nonlinear transonic aeroelastic phenomena. By using CFD-based techniques to construct the aerodynamic operator within the ROM framework, one can preserve the essential nonlinear effects and enhance computational efficiency. This research aims to thoroughly investigate existing methods for developing CFD-based ROMs, focusing on enabling fully nonlinear aeroelastic analyses. Potential techniques for nonlinear system identification will be explored, including the Koopman theory, Sparse Identification of Nonlinear Dynamics (SINDy), and neural networks. The selected methodologies will be tested on representative non-classical flutter cases, such as LCO, to assess their predictive capabilities. The accuracy of the proposed ROMs will be evaluated by comparing their predictions to high-fidelity CFD data. This will ensure that the ROMs can reliably identify critical stability boundaries, such as flutter and LCO onset. In addition to accuracy, the research will also prioritize computational efficiency, evaluating both training and execution times to assess the feasibility of integrating these models into aeroelastic design workflows. By combining high-fidelity CFD with reduced-order modeling, the research project aims to establish an efficient and reliable framework for predicting fully nonlinear aeroelastic behavior, ultimately facilitating the advancement of flexible aircraft analysis and design.