Aeroelastic Optimization for Dynamic Model Adjustment Based on Ground- / Flight Test Data and Increased Stability Margin

Kurzfassung

In the context of aeroelastic stability assessment, the development of a reliable Finite Element Model (FEM) is paramount for ensuring aircraft safety and performance. This paper presents a comprehensive approach to dynamic model adjustment and aeroelastic optimization, integrating ground vibration test and flight vibration test data within a unified optimization framework. The proposed methodology, implemented via the NASTRAN optimization module, enables simultaneous calibration of numerical models to both ground and flight test data, overcoming previous limitations that restricted adjustments to GVT. The framework now supports the inclusion of effects from FVT (aerodynamics), and allows the adaptation of distributed design fields of wing, fuselage and empennage for more accurate representation of the aeroelastic behaviour. The Dynamic Model Adjustment Optimization (DynAOPT) process is demonstrated on a civil transport test aircraft, where transfer functions, particularly rudder-to-cockpit responses, were used as validation targets. Optimization was performed across multiple frequency bands, including frequencies up to 10 Hz, approximately. The results show significant improvement in correlation between test data and simulation, especially in amplitude and frequency matching, with the optimized model successfully eliminating unphysical peaks observed in the baseline model. Notably, the use of move limits on design variables was shown to influence both convergence behaviour and the distribution of stiffness and damping adjustments, highlighting the importance of constraint definition. Furthermore, the paper addresses the challenge of increasing aeroelastic stability margins through structural optimization, using the same optimization tool environment. A case study on a transport aircraft wing model demonstrates the successful increase of flutter speed by 5% through targeted modifications of wing skin thickness, with minimal mass increase. The optimization was conducted on an isolated wing model to simplify the analysis, with results successfully transferred to the full aircraft model, yielding a 4.62% increase in flutter velocity. This confirms the efficiency and scalability of the approach.